Image AFTD Auto-Formalizing Theoretical Domains

AFTD An open research programme

Auto-Formalizing Theoretical Domains

Open proof to the community.
Open verdict by the community.

A machine that reads a curriculum of the theoretical sciences, states their theorems in Lean 4, proves them against Mathlib, and publishes whatever Lean accepts — immediately, to everyone, with no paper and no author line. Nothing is here because a language model said it was true.

SelfHaltingProblemUndecidable.lean

The diagonal halting problem is not computable.

/-- The diagonal halting problem is undecidable (not computable). -/
theorem self_halting_problem_undecidable : ¬ComputablePred (fun c : Nat.Partrec.Code => (Nat.Partrec.Code.eval c (Encodable.encode c)).Dom) := by
  intro h
  obtain ⟨_, hc⟩ := h
  let f : Nat.Partrec.Code → ℕ →. ℕ := fun c _ =>
    cond (decide (Nat.Partrec.Code.eval c (Encodable.encode c)).Dom) Part.none (Part.some 0)
  have hf : Partrec₂ f :=
    Partrec.cond (hc.comp Computable.fst) Partrec.none (Computable.const 0).partrec
  obtain ⟨c, e⟩ := Nat.Partrec.Code.fixed_point₂ hf
  have e_app := congr_fun e (Encodable.encode c)
  dsimp [f] at e_app
  by_cases H : (Nat.Partrec.Code.eval c (Encodable.encode c)).Dom
  · have h_none := e_app ▸ H
    simp [H] at h_none
  · apply H
    rw [e_app]
    simp [H]
Lean acceptedaxioms: Classical.choice · Quot.sound · propextOpen →

News

Launches and newly proved results

Daily reading

Daily reading: 2 research-level, 91 proved, 80 open

Research-level results today: 2. (medium tier) Establishes weighted half-neighborhood coverage for oriented graphs whose every induced non-neighborhood has an out-kernel, including graphs outside the tournament–acyclic class treated in arXiv:2610.23487; (low tier) refutes OP-1087 (Partition of Odd-Girth 5 Subcubic Graphs with Component Order at Most Three). Each research-level result is one Research Blog post; theorems that bear on the same open problem or paper share it.

Details

The Proposer read 34 new paper(s) and listed 39 open problem(s); 53 new question(s) of AFTD's own were adopted onto the list; 19 list entries were stated in Lean, as were 21 results the papers prove; the Proposer posed 86 more from what the knowledgebase already holds (173 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 91 theorem(s), 43 of them helper lemmas, which are not listed below; 80 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. digraph_nonneighborhood_kernels_weighted_half_cover (arXiv:2610.23487): In a finite loopless oriented graph in which the induced non-neighborhood of every vertex has an out-kernel, every nonnegative integer weighting has a quasi-kernel whose closed out-neighborhood contains at least half the total weight. — (medium tier) Establishes weighted half-neighborhood coverage for oriented graphs whose every induced non-neighborhood has an out-kernel, including graphs outside the tournament–acyclic class treated in arXiv:2610.23487

Research-level. subcubic_odd_girth_five_partition_component_le_three_negation (arXiv:2611.05178): It is not the case that every finite simple graph with maximum degree at most 3, clique-free of order 3, and not bipartite admits a partition of its vertex set into an independent set L and a complement R = Lᶜ such that every connected component of the… — (low tier) refutes OP-1087 (Partition of Odd-Girth 5 Subcubic Graphs with Component Order at Most Three)

Partial result on OP-1095. ex_post_one_mms_chore_lottery_single_agent (arXiv:2611.04884): For a single agent (n = 1) and any m indivisible chores with nonnegative additive costs c, there exists an allocation lottery that is simultaneously ex-ante proportional and ex-post 1-MMS.

Partial result on OP-1088. subcubic_odd_girth_seven_partition_of_card_le_five (arXiv:2611.05178): Every finite simple graph G on at most 5 vertices admits a partition into an independent set L and a subset R = Lᶜ such that every connected component of the induced subgraph G[R] has at most 5 vertices.

Partial result on OP-926. simple_graph_forty_nine_regular_five_cycle_one_undominated (arXiv:2401.00392): There is a triangle-free 9-regular simple graph on 40 vertices and a five-vertex set whose induced graph has degree two at every vertex, such that exactly one vertex has no neighbor in that set.

Partial result on OP-926. simple_graph_circulant_thirty_five_five_coloring_no_bichromatic_eight (arXiv:2401.00392): The 35-vertex circulant graph with steps ±1, ±7, ±11, ±16 has a proper five-coloring in which every independent set of eight vertices uses at least three colors.

Partial result on OP-926. simple_graph_regular_forty_induced_thirty_five_total_dominating (arXiv:2401.00392): In a triangle-free 9-regular graph on 40 vertices, the complement of an induced 8-regular 35-vertex subgraph is a total dominating five-set.

Partial result on OP-1087. subcubic_partition_component_le_three_of_isBipartite (arXiv:2611.05178): Every bipartite finite simple graph admits a partition of its vertex set into an independent set L and a complement R = Lᶜ such that every connected component of the induced subgraph G[R] has order at most 3.

Partial result on OP-1087. subcubic_partition_component_le_three_of_card_le_three (arXiv:2611.05178): Every finite simple graph on at most 3 vertices admits a partition of its vertex set into an independent set L and a complement R = Lᶜ such that every connected component of the induced subgraph G[R] has order at most 3.

Partial result on OP-926. simple_graph_twenty_one_three_independent_seven_parts_alpha_seven (arXiv:2401.00392): There is a triangle-free simple graph on 21 vertices, of maximum degree at most eight, in which each of the three consecutive seven-vertex classes is independent and there is no independent set of eight vertices.

Formalized. boxProd_completeGraph_all_powers_well_covered (arXiv:2611.03956 Theorem 2.11 (1)): Let G be a finite connected nontrivial graph, and let t be an integer with t > |V(G)|. If every positive power G^d (d ≥ 1) is well-covered, then every positive power of the Cartesian product G □ K_t is well-covered.

Formalized. exists_connected_all_powers_well_covered_diam_ge (arXiv:2611.03956 Corollary 2.13): For every positive integer D, there exists a finite connected simple graph G with diameter at least D such that every positive power G^d (d ≥ 1) is well-covered.

Formalized. binary_code_odd_forbidden_dist_max_size (arXiv:2611.04523, Proposition 3.1): Let n ≥ 1 and 1 ≤ d ≤ n be integers such that d is odd. Then every subset F : Finset (Fin n → Fin 2) such that no two elements of F have Hamming distance d satisfies F.card ≤ 2 ^ (n - 1), and there exists a subset F : Finset (Fin n → Fin 2) with no two…

Formalized. mordell_ten_integer_norm_sum (arXiv:2611.03088 Section 5.1): The sum of squared norms of the ten integer points in the witness configuration W is 9999608174.

Formalized. mordell_ten_integer_witness_ineq (arXiv:2611.03088 eq. (8)): For the ten integer points W = ![(11183, 0), (-11183, 0), (32765, 14669), (32765, -14669), (-32765, 14669), (-32765, -14669), (12323, 31572), (12323, -31572), (-12323, 31572), (-12323, -31572)] in Fin 10 → ℤ × ℤ and q = 9999608174, 9^36 times the product…

Formalized. mordell_ten_suboptimal (arXiv:2611.03088 Proposition 5.1): In Mordell's problem for ten points in the complex plane, the supremum sSup of { d : ℝ | ∃ z : Fin 10 → ℂ, (∑ i : Fin 10, ‖z i‖^2 = 10) ∧ d = ∏ i : Fin 10, ∏ j : Fin 10 with i < j, ‖z i - z j‖^2 } strictly exceeds (10 : ℝ)^45 / (9 : ℝ)^36.

Formalized. graph_edit_cost_permutation_decomposition_le_two_mul (arXiv:2611.05048 Theorem 6): Let $Q$ be an $n \times n$ doubly stochastic matrix with a convex decomposition $Q = \sum_{i \in \iota} w_i P_{\pi_i}$, where $w_i \ge 0$, $\sum_{i} w_i = 1$, and each $\pi_i$ is a permutation of $\{0, \dots, n-1\}$. Then for any binary matrices $A, B \in…

Formalized. anti_ramsey_k5_lower_bound_nineteen (arXiv:2611.03878 Theorem 3.2): There exists a symmetric 3-edge-coloring of the complete graph K_19 on vertices ZMod 19 such that every 5-element vertex subset contains at least one edge of each of the three colors, implying that the three-colored anti-Ramsey number satisfies xi(K_5) >= 19.

Corrected. not_boolean_subspace_degree_reduction_criterion: The statement boolean_subspace_degree_reduction_criterion is false: The statement is false for d = 0. When d = 0: Let n = 1, p : (Fin 1 → ZMod 2) → ZMod 2 be the constant 1 function, and U = ⊥. 1. booleanDegree p = 0, because the ANF coefficient for S = ∅…

Open. subcubic_odd_girth_seven_partition (arXiv:2611.05178): Every finite simple graph G with maximum degree at most 3 and odd girth at least 7 admits a partition of its vertex set into an independent set L and a subset R = Lᶜ such that every connected component of the induced subgraph G[R] has at most 5 vertices.

Open. balanced_tsallis_seven_eighths_sharp_dictator_stability_transition (knowledgebase): For every correlation 0 < rho < 1, natural number n >= 1, and balanced Boolean indicator f : {0,1}^n -> {0,1}, the order-7/8 noise stability excess E[(T_rho f)^(7/8)] - h(rho) is at least min(C_3(rho), C_G(rho)) times the dictator distance delta(f).

Open. subcubic_triangle_free_partition_component_le_six (arXiv:2611.05178): Every finite simple graph with maximum degree at most 3 that is clique-free of order 3 admits a partition of its vertex set into an independent set L and a complement R = Lᶜ such that every connected component of the induced subgraph G[R] has order at most 6.

Open. triangle_free_subcubic_card_le_four_partition (arXiv:2611.05178): Every finite simple graph on at most 4 vertices with maximum degree at most 3 that is triangle-free admits a partition into an independent set L and complement R = Lᶜ such that G[R] is bipartite and for every connected component C of G[R] and every vertex v…

Open. op_1058_cartesian_product_factor_has_wcp (arXiv:2611.03956): For any finite connected simple graphs G and H, if their Cartesian product G □ H has the well-covered powers property, then at least one of G and H has the well-covered powers property.

Open. op_1058_cartesian_square_factor_has_wcp (arXiv:2611.03956): For any finite connected simple graph G, if its Cartesian square G □ G has the well-covered powers property, then G has the well-covered powers property.

Verified in Lean

Daily reading

Daily reading: 0 research-level, 92 proved, 48 open

No research-level result today. Partial results advanced several graph-theoretic open problems: for OP-864, the Shannon OR-capacity of categorical products was bounded by the minimum factor capacity and evaluated when that minimum is 1 or 2; OP-337 gained private-leaf reductions and a seven-vertex tournament obstruction; and neighborhood matchings for triangle-free graphs were established toward OP-926. Published theorems formalized in full include quantitative_combinatorial_nullstellensatz, two_color_generalized_schur_number_prop31, gadget cut values in max3cut_gadget_cut_values, and a plank system sign vector theorem in plank_system_sign_vector_exists. New papers addressed fair allocation, graph coloring, additive energy, and coding theory, while 116 open-problem entries were added, including a question proposed here in OP-989. Forty-eight statements remain open, including Rényi entropy extremal bounds in OP-989 and subadditivity of regularized Jensen-Shannon mutual information in OP-998.

Details

The Proposer read 44 new paper(s) and listed 62 open problem(s); 54 new question(s) of AFTD's own were adopted onto the list; 13 list entries were stated in Lean, as were 15 results the papers prove; the Proposer posed 62 more from what the knowledgebase already holds (116 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 92 theorem(s), 40 of them helper lemmas, which are not listed below; 48 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Partial result on OP-864. shannon_or_capacity_categorical_prod_le_min (arXiv:2610.41977): For any finite nonempty simple graphs G and H, the Shannon OR-capacity of their categorical product G × H is at most the minimum of the Shannon OR-capacities of G and H.

Partial result on OP-864. shannon_or_capacity_categorical_prod_of_min_eq_one (arXiv:2610.41977): For any finite nonempty simple graphs G and H, if the minimum of their Shannon OR-capacities is 1, then the Shannon OR-capacity of their categorical product G × H is 1.

Partial result on OP-864. shannon_or_capacity_categorical_prod_of_min_eq_two (arXiv:2610.41977): For any finite nonempty simple graphs G and H, if the minimum of their Shannon OR-capacities is 2, then the Shannon OR-capacity of their categorical product G × H is 2.

Partial result on OP-337. digraph_two_private_leaves_avoidance_reduces_large (arXiv:2610.23487): Attach two distinct private outgoing leaves to each vertex of a finite digraph. Source-freeness is preserved and the leaves form the unique maximum independent set. Any quasi-kernel meeting at most half of the second leaves projects to a quasi-kernel of the…

Partial result on OP-337. digraph_quasikernel_independent_cover_avoids_outneighbors (arXiv:2610.23487): In a finite loopless digraph, every independent set is contained in the closed out-neighborhood of a quasi-kernel that contains none of its open out-neighbors.

Partial result on OP-337. digraph_seven_tournament_two_neighborhood_cover_counterexample (arXiv:2610.23487): The displayed source-free seven-vertex tournament has all singleton quasi-kernels, every quasi-kernel is a singleton, no pair of their closed out-neighborhoods covers all vertices, and each vertex belongs to four of those neighborhoods.

Partial result on OP-337. finite_three_weighted_sets_half_cover_implies_pair_cover (arXiv:2610.23487): If a positive total of three nonnegative integer weights gives every point at least half the weight of the sets containing it, two of those sets cover every point.

Partial result on OP-337. digraph_seven_cover_counterexample_rigidity (arXiv:2610.23487): Every independent four-vertex set in the displayed seven-vertex graph equals {1,2,3,6}, and every independent set whose closed out-neighborhood covers {1,2,3,6} also equals it.

Formalized. quantitative_combinatorial_nullstellensatz (arXiv:2610.41820 Lemma 2.1): For a multivariate polynomial f of total degree at most ∑ t_i over a field F, the coefficient of the monomial ∏ x_i^{t_i} equals the sum over points a in the grid ∏ S_i of f(a) divided by ∏_i ∏_{b ∈ S_i \ {a_i}} (a_i - b), where each S_i has cardinality t_i…

Formalized. two_color_generalized_schur_number_prop31 (arXiv:2611.01714 Proposition 3.1): For every integer k ≥ 3, every 2-coloring of Fin (k^2 + k - 1) admits a monochromatic solution to (x.val + 1) + (k - 1) * (y.val + 1) = z.val + 1, and the 2-coloring of Fin (k^2 + k - 2) defined by color 0 on [1, k-1] ∪ [k^2, k^2+k-2] and color 1 on [k,…

Formalized. max3cut_gadget_cut_values (arXiv:2611.01402 Lemma 3.1 -- Improved NP-Hardness of Approximation for Max-3-Cut and for Max-k-Cut with Small k): For every primary coloring x : Fin 3 → ZMod 3, the maximum cut value of the E-gadget is max3cut_v_e x = if max3cut_s x = 0 then 18 else 16, and the maximum cut weight of the N-gadget is max3cut_v_n x = if max3cut_s x ≠ 0 then 11 else 9.

Formalized. plank_system_sign_vector_exists (arXiv:2611.00379, Lemma 3 (A sharp plank theorem for cube annuli)): Let $V$ be a real inner product space, let $\iota$ be a finite type, let $n : \iota \to V$ be unit vectors, let $w : \iota \to \mathbb{R}$ be positive widths ($w_i > 0$ for all $i$), and let $s : \iota \to \mathbb{R}$ be shifts. Then there exists…

Formalized. hyperbinary_eval_cube_root_of_unity (scout/2611-01243-halving-partitions-and-hyperbinary-expansions.md:1200): For every non-negative integer m, evaluating the hyperbinary length polynomial H_m(z) at the primitive cube root of unity ω = (-1 + √(-3))/2 gives an element of {0, ±1, ±ω, ±ω²}.

Formalized. fano_free_local_forbidden (arXiv:2610.01084 Lemma 3.1 inequality (3.1)): In any 3-uniform Fano-free hypergraph H : Finset (Finset V), if {x, y, z} is an edge of H, then for all vertices a, b, c, d in V, the six triples {x, a, b}, {x, c, d}, {y, a, c}, {y, b, d}, {z, a, d}, and {z, b, c} cannot all belong to H.

Formalized. direct_sum_cover_maximal_independent_truncated_cayley (arXiv:2611.05321 Lemma 2.2): Let G be an additive abelian group, and let A ⊆ D ⊆ G and Z ⊆ G be subsets with 0 ∈ Z. Suppose that the sum A + Z is direct and D ⊆ A + Z. Then the preimage of A in D is a maximal independent set in the truncated Cayley graph on D with connection set Z - Z.

Corrected. not_strong_elimination_iff_compatible_and_union_min: The statement strong_elimination_iff_compatible_and_union_min is false: The statement is false as stated because of the empty set: If ∅ ∈ C, then is_compatible_collection C is trivially satisfied by picking C3 = ∅ for any distinct C1, C2 ∈ C and x ∈ C1 ∩…

Open. two_deletion_posterior_renyi_sub_one_alternating_max (knowledgebase): For every real order α with 0 < α < 1, every length n ≥ 2, and every word w : Fin n → Bool, the Rényi entropy renyi_entropy α of two_deletion_posterior n w is at most that of alternating_word n false, with equality if and only if is_alternating_word w holds.

Open. two_deletion_posterior_renyi_order_crossings_exist (knowledgebase): For every real order α > 1, there exist a natural number n ≥ 2 and a word w : Fin n → Bool such that the Rényi entropy renyi_entropy α of two_deletion_posterior n (alternating_word n false) is strictly less than that of two_deletion_posterior n w.

Open. op_989_two_deletion_posterior_renyi_alternating_sharp_order_transition (knowledgebase): For every real order α with 0 < α < 1, every n ≥ 2, and every word w : Fin n → Bool, the Rényi entropy renyi_entropy α of two_deletion_posterior n w is at most that of alternating_word n false, with equality if and only if is_alternating_word w; whereas for…

Open. jensen_shannon_regularized_subadditive_full_support_half (knowledgebase): For regularization parameter c = 1/2, the regularized Jensen-Shannon mutual f-information is subadditive on all full-support Markov chains Y <- X -> Z on finite alphabets.

Open. op_998_jensen_shannon_sharp_kl_regularization (knowledgebase): There exists a sharp regularization threshold c* in (0, 1/2] such that for every non-negative c, the regularized Jensen-Shannon mutual information is subadditive under all finite Markov forks if and only if c* <= c.

Open. simple_graph_circulant_thirty_five_three_link_obstruction (arXiv:2401.00392): In the 35-vertex circulant graph with step set ±1, ±7, ±11, ±16, no three pairwise-disjoint independent sets A, B, C can have every independent eight-set meet both A union B and A union C.

Verified in Lean

Daily reading

Daily reading: 2 research-level, 220 proved, 41 open

Research-level results today: 2. (medium tier) Exhibits an unweighted three-color Max-k-Cut game with a cycle of inclusion-minimal profitable deviations, ruling out a strong potential even when deviations are restricted to minimal coalitions; (low tier) An explicit source-free oriented graph of maximum outdegree four gives f(4) at least 5/4 for the uniform source-neighborhood inequality in arXiv:2601.11847, Problem 5.2, while itself having a small quasi-kernel. Each research-level result is one Research Blog post; theorems that bear on the same open problem or paper share it.

Details

The Proposer read 45 new paper(s) and listed 56 open problem(s); 46 new question(s) of AFTD's own were adopted onto the list; 15 list entries were stated in Lean, as were 4 results the papers prove; the Proposer posed 16 more from what the knowledgebase already holds (198 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 220 theorem(s), 181 of them helper lemmas, which are not listed below; 41 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. maxcut_unweighted_minimal_deviation_cycle_exists: On the specified unweighted 71-vertex graph there is an infinite three-color strong-deviation sequence returning after 33 steps, and every step is inclusion-minimal: no proper subcoalition of its changing players has a profitable deviation from that step’s… — (medium tier) Exhibits an unweighted three-color Max-k-Cut game with a cycle of inclusion-minimal profitable deviations, ruling out a strong potential even when deviations are restricted to minimal coalitions

Research-level. digraph_four_outdegree_five_four_counterexample: There exists a loopless source-free oriented digraph on 24 vertices with maximum outdegree at most four such that for every vertex v, if S(v) is the set of sources after deleting the closed out-neighborhood of v and E(v) is the union of the out-neighborhood… — (low tier) An explicit source-free oriented graph of maximum outdegree four gives f(4) at least 5/4 for the uniform source-neighborhood inequality in arXiv:2601.11847, Problem 5.2, while itself having a small quasi-kernel

Partial result on OP-697. open_problem_qudit_pure_state_invariants (arXiv:2610.36076): For any natural number d ≥ 3 and any family T of d^2 - 1 matrices satisfying is_orthonormal_hermitian_basis d T, the constant function (fun _ => 1) is a quadratic pure-state invariant for T, and for all matrices rho satisfying is_pure_state rho,…

Partial result on OP-826. op826_envelope_at_two (knowledgebase): For value ratio R = 2, the discrete envelope min (1 / floor(sqrt 2)) (ceil(sqrt 2) / 2) equals 1.

Partial result on OP-826. op826_envelope_perfect_square (knowledgebase): For any positive integer k, when the value ratio is a perfect square R = k^2, the conjectured envelope min (1 / floor(sqrt(k^2))) (ceil(sqrt(k^2)) / k^2) equals 1 / k.

Partial result on OP-826. op826_online_single_high_efx_ratio_le_one (knowledgebase): For every real value ratio R, the online single-high EFX competitive ratio for three agents is at most 1.

Partial result on OP-826. op826_online_single_high_efx_upper_bound_two (knowledgebase): For three agents and value ratio R = 2, the online single-high EFX competitive ratio is at most the conjectured envelope min (1 / floor(sqrt 2)) (ceil(sqrt 2) / 2).

Partial result on OP-828. chain_trees_floor_claw_packing_gap_ge (knowledgebase): For every integer k ≥ 1, the normalized gap (3k - ⌊5k/2⌋) / (16k) between the integer tripartite information bounds on the 16k-vertex chain trees is at least 1/32.

Partial result on OP-828. chain_trees_extensive_mmi_gap_of_exact_claw_packings (knowledgebase): For any integer k ≥ 1, if max_claw_packing (chain_tree_1 k) = 5 * k / 2 in ℕ (using integer division) and max_claw_packing (chain_tree_2 k) = 3 * k in ℕ, then ((max_claw_packing (chain_tree_2 k) : ℝ) - (max_claw_packing (chain_tree_1 k) : ℝ)) / (16 * k : ℝ)…

Partial result on OP-828. chain_tree_vertex_count (knowledgebase): For every integer k with 1 ≤ k, the Cartesian product type Fin k × Fin 16 has cardinality Fintype.card (Fin k × Fin 16) = 16 * k.

Thread RT-1. cycle_power_dyadic_outer_pair_reverse: For exponent m at least three, consider a Boolean integer-indexed dyadic class model with exact dyadic ownership at every rank, coverage of every ordinary distance below 2^m, equal boundary bits at offsets 1 and 2^m-2, and only endpoints selected in every…

Thread RT-1. cycle_power_dyadic_domain_wall_boundary_pairing: The existing radius-eight domain-wall Boolean class model has equal bits at offsets 1 and 6 at every integer root.

Thread RT-1. cycle_power_dyadic_count_free_boundary_endpoint_counterexample: There exists a radius-eight Boolean integer class model with tight classes, exact dyadic ownership, full ordinary distance coverage, and equal boundary bits at offsets 1 and 6 at every root, but whose class at root zero selects offset 3 while omitting offset…

Formalized. exact_cutoff_filtering (arXiv:2610.42105 Theorem 4.3): In a finite marriage market with strict complete preference profile Q, for every agent (woman w with stable partner t in M, or man m with stable partner t in W), the set of stable matchings of the profile obtained by t-cutoff tanking by that agent equals the…

Formalized. cube_root_identity_distinct (arXiv:2610.41820 Lemma 2.4, Eq. (15)): If ω is a primitive cube root of unity in ℂ and x, y ∈ {1, ω, ω^2} are distinct, then (x - y)^2 = -3xy.

Formalized. cube_root_four_cycle_identity (arXiv:2610.41820 Lemma 2.4, Eq. (16)): If ω is a primitive cube root of unity in ℂ and a, b, c, d ∈ {1, ω, ω^2} satisfy a ≠ b, b ≠ c, c ≠ d, and d ≠ a, then (a - b)(b - c)(c - d) = 3bc(a - d).

Formalized. circulant_1_3_three_colorable (arXiv:2610.41820 Section 2.3): For every integer n ≥ 8, the circulant graph C_n(1,3) has a proper vertex 3-colouring, so N_3(C_n(1,3)) > 0.

Formalized. truncation_local_states_strong_coloring_iff (arXiv:2610.41950 (Strong edge-colourings of truncations of cubic graphs), Lemma 2.1 (the first, signature form; the equivalent form (2.3)…): Lemma 2.1 of arXiv:2610.41950. Let G be a loopless cubic multigraph, given by ends : E → Sym2 V (no Fintype or Finite instance on V or E is assumed: the cubic condition already makes the set of edges at each vertex finite, and the lemma is local) with no…

Proved. maxcut_fair9_nash_fairness_obstruction: On the specified nine-vertex three-color graph there is a singleton-stable profile with every utility at least four, no strong equilibrium has every utility at least four, and a strong equilibrium exists.

Open. op_828_tree_graph_renyi_spectral_collision_extensive_mmi_gap (knowledgebase): Open Problem OP-828: for every real number ε > 0 and every natural number N₀, there exist a natural number N ≥ N₀, a finite type V with decidable equality and Fintype.card V = N, and two simple trees T and T' on V with decidable adjacency relations having…

Open. two_chore_types_pmms_fractional_pareto_sharp_four_thirds (knowledgebase): The sharp universal constant for pairwise maximin share fairness under fractional Pareto optimality for two chore types is 4/3: two_type_chores_have_alpha_pmms_fpo (4 / 3) holds, and for every real number alpha < 4/3, two_type_chores_have_alpha_pmms_fpo…

Open. ladder_graph_state_tripartite_info_lower_bound (knowledgebase): For every integer n ≥ 2, there exists a four-partition A, B, C, D of the vertices of the 2-by-n ladder graph achieving tripartite information tree_graph_state_tripartite_info (SimpleGraph.pathGraph n □ SimpleGraph.completeGraph (Fin 2)) A B C D = ((n - 2) /…

Open. ladder_two_graph_state_max_tripartite_info_eq_zero (knowledgebase): For the 2-rung ladder graph SimpleGraph.pathGraph 2 □ SimpleGraph.completeGraph (Fin 2), the tripartite information of every four-partition is at most 0, and there exists a four-partition achieving tripartite information equal to 0.

Open. ladder_five_graph_state_tripartite_info_ge_one (knowledgebase): For the 5-rung ladder graph SimpleGraph.pathGraph 5 □ SimpleGraph.completeGraph (Fin 2), there exists a four-partition A, B, C, D of its vertices satisfying is_four_partition A B C D such that 1 ≤ tree_graph_state_tripartite_info (SimpleGraph.pathGraph 5 □…

Open. ladder_graph_state_exact_max_tripartite_information (knowledgebase): For every integer n ≥ 2, the maximum tripartite information of the 2-by-n ladder graph SimpleGraph.pathGraph n □ SimpleGraph.completeGraph (Fin 2) over all four-partitions A, B, C, D satisfying is_four_partition A B C D is exactly ((n - 2) / 3 : ℤ).

Verified in Lean

Daily reading

Daily reading: 1 research-level, 115 proved, 66 open

Research-level results today: 1. (low tier) Exhibits a cycle of profitable coalitional deviations in an unweighted three-color game, ruling out a strong potential for all such deviations; the cited cycle used weighted edges. This does not assert that the large coalition is minimal or resolve strong-equilibrium existence. Each research-level result is one Research Blog post; theorems that bear on the same open problem or paper share it.

Details

The Proposer read 27 new paper(s) and listed 28 open problem(s); 48 new question(s) of AFTD's own were adopted onto the list; 18 list entries were stated in Lean, as were 3 results the papers prove; the Proposer posed 12 more from what the knowledgebase already holds (150 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 115 theorem(s), 82 of them helper lemmas, which are not listed below; 66 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. maxcut_unweighted_strong_deviation_cycle_exists: There is an unweighted simple graph on 46 vertices and an infinite sequence of three-color profiles that returns to its starting profile after five steps, such that every consecutive pair is a profitable deviation by exactly the players who change colors. — (low tier) Exhibits a cycle of profitable coalitional deviations in an unweighted three-color game, ruling out a strong potential for all such deviations; the cited cycle used weighted edges. This does not assert that the large coalition is minimal or resolve strong-equilibrium existence

Partial result on OP-670. parallel_uniform_merino_welsh_strict_balancing_three_two (knowledgebase): For any function q : Fin 3 → ℕ with 0 < q i for all i : Fin 3, and any indices i, j : Fin 3 with q i + 2 ≤ q j, parallel_mw_ratio 2 (parallel_mw_transfer q i j) < parallel_mw_ratio 2 q.

Partial result on OP-809. fixed_offset_hash_worst_mean_mixed_radix_prime (knowledgebase): For any prime p, any bijective offset sequence r : Fin p → ZMod p, and any load t < p, if linear probing achieves expected probe cost at least that of r, then there exists an ordering P of the prime factors of p such that the expected probe cost under r is…

Partial result on OP-736. ghz_standard_frame_lambda_le (arXiv:2610.35838): For the generalized GHZ state with parameter alpha and standard identity frames e = 1 and f = 1, the two-copy closed-network quantity Lambda(tau_{mu nu}, tau_{mu' nu'}) is at most 1 + sin^2(2*alpha) for all setting indices mu, nu, mu', nu' in {0, 1, 2, 3}.

Partial result on OP-736. ghz_closed_network_lambda_le_two (arXiv:2610.35838): For every parameter alpha in R, all orthonormal frames e, f in O(3), and all setting indices mu, nu, mu', nu' in {0, 1, 2, 3}, the two-copy closed-network Bell measurement quantity Lambda(tau_{mu nu}, tau_{mu' nu'}) for the generalized GHZ state is at most 2.

Partial result on OP-736. ghz_network_lambda_at_pi_div_four_le (arXiv:2610.35838): For the maximally entangled state parameter alpha = pi/4, all orthonormal frames e, f in O(3), and all setting indices mu, nu, mu', nu' in {0, 1, 2, 3}, the two-copy closed-network quantity Lambda(tau_{mu nu}, tau_{mu' nu'}) is at most 1 + sin^2(2 * (pi / 4)).

Partial result on OP-794. tree_graph_state_tripartite_info_trivial_zero (knowledgebase): For any finite graph G on type V with decidable equality and decidable adjacency, tree_graph_state_tripartite_info G Finset.univ ∅ ∅ ∅ equals 0.

Partial result on OP-801. prime_fixed_offset_linear_probe_stochastic_maximum_three (knowledgebase): For a 3-slot hash table (p = 3), every bijective probe sequence r : Fin 3 → ZMod 3 with r 0 = 0, every load t < 3, and every integer k ≥ 0, the probability hash_probe_tail_prob 3 r t k of requiring strictly more than k probes is at most that under linear…

Partial result on OP-801. prime_fixed_offset_mean_le_of_tail_le (knowledgebase): For any prime table size p, offset sequence r : Fin p → ZMod p, and load t, if every probe-cost tail probability under r is bounded by the corresponding tail probability under linear probing, then the expected probe cost under r is at most the expected probe…

Thread RT-1. cycle_power_dyadic_boundary_pairing_alternates: In an exact dyadic Boolean class model with exponent at least two, equality of the two boundary bits in every class forces the first inner bit to alternate between successive classes.

Thread RT-1. cycle_power_dyadic_outer_pair_forward: In a fully covered exact dyadic Boolean class model with exponent at least three and matching boundary bits in every class, if a class selects offsets 1 and 3 it also selects offset 2.

Formalized. arrangement3_e4_zero_iff_is_magic (arXiv:2610.38386, Proposition 5.1): An arrangement of order 3 satisfies the four-term line-sum energy zero condition $E_4(S) = 0$ if and only if it is a magic square.

Formalized. tournament_exists_covering_realization_of_realizes (arXiv:2610.40521 Lemma 6 -- Directed Cut-Query Complexity of Finding a King): Let d be a degree vector on Fin n realized by a tournament T, and let v, w be vertices in Fin n satisfying d v < d w. Then there exists a tournament T' on Fin n realizing d in which w covers v.

Formalized. graph_at_most_one_between (arXiv:2610.39966 Lemma 3): In a connected graph G, if y lies between x and z and x lies between y and z, then x = y; hence among three distinct vertices, at most one lies between the other two.

Corrected. not_hash_probe_tail_prob_zero_of_lt: The statement hash_probe_tail_prob_zero_of_lt is false: Take n = 1, r = (fun _ => 0 : Fin 1 → ZMod 1), t = 1, k = 2. Hypotheses hold: htn : 1 ≤ 1 and hkt : 1 < 2. But hash_probe_tail_count 1 r 1 2 = 1 (kernel-checked by unfolding the definition), so…

Open. tree_graph_state_max_tripartite_info_eq_claw_packing (knowledgebase): For every finite tree T with decidable equality and decidable adjacency, the tripartite information tree_graph_state_tripartite_info T A B C D of every four-partition A, B, C, D : Finset V (satisfying is_four_partition A B C D) is at most max_claw_packing T,…

Open. claw_free_tree_graph_state_mmi (knowledgebase): For every finite tree T with decidable equality, decidable adjacency, and max_claw_packing T = 0, every four-partition A, B, C, D : Finset V satisfying is_four_partition A B C D has tree_graph_state_tripartite_info T A B C D ≤ 0.

Open. tree_with_claw_tripartite_info_pos (knowledgebase): For every finite tree T with decidable equality, decidable adjacency, and max_claw_packing T ≥ 1, there exists a four-partition A, B, C, D : Finset V satisfying is_four_partition A B C D such that 1 ≤ tree_graph_state_tripartite_info T A B C D.

Open. plurality_veto_core_exact_regularization_threshold (knowledgebase): For all integers m ≥ 3 and n ≥ max(8, m), the exact regularization threshold for plurality veto core containment is plurality_regularization_threshold n m = 1 / ((m - 2)n - 2m): for all 0 < ε < plurality_regularization_threshold n m, the regularized…

Open. prime_fixed_offset_linear_probe_stochastic_maximum (knowledgebase): For every prime p, every bijective offset sequence r : Fin p → ZMod p with r 0 = 0, every load t < p, and every integer k ≥ 0, the probability hash_probe_tail_prob p r t k that an additional uniform insertion requires strictly more than k probes under r is…

Open. energy_filter_rational_resonance_theta_crossover (knowledgebase): For any real number tau > 0, coprime positive integers p, q, and real detuning c, let L be the 3x3 real matrix with diagonal 2 and adjacent off-diagonal entries -1, let D = p^2 + q^2, let u = pi^2 * q^2 / (2 * tau^2) + pi^2 * c^2 * q^4 / (4 * D), let S_2 =…

Verified in Lean
Earlier news (58)

Weekly report

Week of 26 Oct – 1 Nov 2026: Social Optima Vulnerable to Coalitions, Latin Tableaux Formula, and Spectral Bounds Refuted

A question in game theory was settled when researchers disproved the conjecture that socially optimal network colorings are always immune to group defections, exhibiting a 182-vertex counterexample. In combinatorics, an exact counting formula for staircase Latin tableaux was proved, and perfect Mondrian partitions were ruled out for all two-prime tile areas. In quantum theory, a proposed universal spectral bound was refuted in all dimensions three and above. Further results proved wild-number formulas and new fairness guarantees.

Details

Disproved. Socially optimal cut colorings can fail to be strong equilibria

Partial. Conjectured wild number formula proved for all large complete graphs

Proved. Exact counting formula proved for Latin tableaux with double-three tails

Proved. No integer square admits a perfect rectangle tiling of two-prime area

Disproved. Universal spectral bound fails for positive trace-preserving maps in dimension three

Lean accepted 1510 new theorems this week, drawn from 97 papers and 23 topics. Of these, 53 are new here, 0 correct a published claim, 150 formalize a published result and 1307 are helper lemmas. They make 14 research-level results, in 16 theorems -- they lead the report. 1276 statements posed this week are not yet proved. The reading covered 109 new papers. This report covers 16 results in 6 fields.

Daily reading

Daily reading: 2 research-level, 167 proved, 118 open

Research-level results today: 2 (3 theorems). (medium tier) Answers the question left open in arXiv:2610.37187 (Sec. 9) and raised in arXiv:2506.02145: positive trace-preserving maps in dimension three can violate the universal spectral bound Tr Φ ≤ d m(Φ) + d² − d that holds for all 2-positive ones; an explicit decomposable map with trace 11/2 and eigenvalue −1/2 exceeds the bound 9/2; (medium tier) Answers the question of arXiv:2610.37187 (Sec. 9, raised in arXiv:2506.02145) in every dimension: for each d ≥ 3 an explicit decomposable positive trace-preserving map violates the bound Tr Φ ≤ d m(Φ) + d² − d, which holds for all 2-positive trace-preserving maps, by 2/(d − 1); so 2-positivity cannot be weakened to positivity in that theorem; (low tier) Extends common least-good insertion to two universally small goods with arbitrary within-pair rankings, preserving every old bundle and weakly improving all own utilities; iteration gives complete EFX for any number of agents and goods ordered in common pairs. Each research-level result is one Research Blog post; theorems that bear on the same open problem or paper share it.

Details

The Proposer read 0 new paper(s) and listed 0 open problem(s); 16 new question(s) of AFTD's own were adopted onto the list; 9 list entries were stated in Lean, as were 10 results the papers prove; the Proposer posed 37 more from what the knowledgebase already holds (270 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 167 theorem(s), 135 of them helper lemmas, which are not listed below; 118 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. not_universal_spectral_bound_positive_trace_preserving (arXiv:2610.37187): Not every positive trace-preserving linear map on d × d complex matrices with d ≥ 3 satisfies Tr(Φ) ≤ d m(Φ) + d² − d: in dimension 3 the positive trace-preserving map X ↦ K X K + (3/4) P Xᵀ P, K = diag(1/2, 1/2, 1), P = diag(1, 1, 0), has operator trace… — (medium tier) Answers the question left open in arXiv:2610.37187 (Sec. 9) and raised in arXiv:2506.02145: positive trace-preserving maps in dimension three can violate the universal spectral bound Tr Φ ≤ d m(Φ) + d² − d that holds for all 2-positive ones; an explicit decomposable map with trace 11/2 and eigenvalue −1/2 exceeds the bound 9/2

Research-level. exists_positive_tp_violating_universal_spectral_bound (arXiv:2610.37187): For every d ≥ 3 there is a positive trace-preserving linear map Φ on d × d complex matrices with Tr(Φ) > d m(Φ) + d² − d. With a = (d−2)/(d−1), K = diag(a, a, 1, …, 1) and P the diagonal projection onto the first two coordinates, Φ(X) = K X K + (1 − a²) P Xᵀ… — (medium tier) Answers the question of arXiv:2610.37187 (Sec. 9, raised in arXiv:2506.02145) in every dimension: for each d ≥ 3 an explicit decomposable positive trace-preserving map violates the bound Tr Φ ≤ d m(Φ) + d² − d, which holds for all 2-positive trace-preserving maps, by 2/(d − 1); so 2-positivity cannot be weakened to positivity in that theorem

Research-level. efx_exists_of_two_universal_small_goods_ir (arXiv:2208.08782): Let m and n be natural numbers with n positive. There are m+2 goods with nonnegative real additive values for n agents. Each agent values each of the last two goods no more than every one of the first m goods, with no required order between the last two.… — (low tier) Extends common least-good insertion to two universally small goods with arbitrary within-pair rankings, preserving every old bundle and weakly improving all own utilities; iteration gives complete EFX for any number of agents and goods ordered in common pairs

Partial result on OP-651. cyclic_cubic_t_count_isotropy_lower_bound (knowledgebase): For every natural number n ≥ 6 and every finite set S : Finset (Finset (Fin n)), if is_parity_repr S (cyclic_cubic_syndrome n), then 2 * n + 1 ≤ S.card.

Partial result on OP-666. three_token_k2n_domination_number_four (knowledgebase): The domination number of the 3-token graph F_3(K_{2,4}) of the complete bipartite graph K_{2,4} is 4.

Partial result on OP-736. ghz_network_lambda_standard_frame (arXiv:2610.35838): For the generalized GHZ state with parameter alpha, the identity frames e = I and f = I at indices (3, 0) and (1, 1) satisfy Lambda(tau_{30}, tau_{11}) = 1 + sin^2(2*alpha).

Partial result on OP-736. ghz_two_copy_closed_lower_bound (arXiv:2610.35838): For every parameter alpha, there exist orthonormal frames e and f in O(3) and setting indices mu, nu, mu', nu' in {0, 1, 2, 3} such that the two-copy closed-network quantity Lambda(tau_{mu nu}, tau_{mu' nu'}) equals 1 + sin^2(2*alpha).

Partial result on OP-735. dichotomic_sum_diff_sq_sum (arXiv:2610.35838): For any two dichotomic projective observables B_0, B_1 on a Hilbert space, (B_0 + B_1)^2 + (B_0 - B_1)^2 = 4 I.

Partial result on OP-716. fourier_threshold_nu_odd_ge_arithmetic_bound (arXiv:2610.38447): For every integer m >= 1, the arithmetic bound lambda_{2m+1} is a lower bound on the Fourier threshold nu(2m+1).

Partial result on OP-716. fourier_threshold_nu_one (arXiv:2610.38447): The Fourier threshold nu(1) for functions from ZMod 2 to ZMod 3 with unique zero at 0 is equal to 2.

Partial result on OP-716. fourier_threshold_nu_two (arXiv:2610.38447): The Fourier threshold nu(2) for functions from (ZMod 2)^2 to ZMod 3 with unique zero at 0 is equal to 3.

Formalized. observable_upper_bound_base_polytope (arXiv:2610.35433 Corollary 1 proof): For any vector a whose values are sorted non-decreasingly by permutation π, and any point z in the base polytope of C, the inner product ∑ i, a i * z i is bounded by a(π(0)) + ∑_{k=0}^{n-2} (a(π(k+1)) - a(π(k))) * C(S_{k+1}).

Formalized. anticommuting_hermitian_expectation_sq_le_one (arXiv:2610.35838 Appendix E, Fact 1): If continuous linear operators P and Q on a complex Hilbert space are self-adjoint, have operator norm at most one, and anticommute (P * Q + Q * P = 0), then for any unit vector psi, the sum of the squares of the real expectation values satisfies (re <psi, P…

Formalized. concave_power_sum_le (arXiv:2610.35838 Appendix E, Fact 2): For non-negative real numbers p and q such that p^2 + q^2 <= 1, and any real r with 0 < r <= 1, the sum of powers satisfies p^r + q^r <= 2^(1 - r / 2).

Formalized. open_network_kernels_anticommute (arXiv:2610.35838 Appendix E, Eq. (61)): In any ring, dichotomic observables A_0, A_1 and B_0, B_1 squaring to 1 that commute with each other yield anticommuting kernel operators: (A_0 * B_0 - A_1 * B_1) * (A_0 * B_1 + A_1 * B_0) + (A_0 * B_1 + A_1 * B_0) * (A_0 * B_0 - A_1 * B_1) = 0.

Formalized. parity_effect_quantum_network_bounds (arXiv:2610.35838 Theorem 4): For every natural number n with 1 <= n, and all real numbers p, q with 0 <= p, 0 <= q, p <= 1, and q <= 1, if ¬ Even n implies p^2 + q^2 <= 1, then Real.sqrt (p ^ (2 / (n : ℝ)) + q ^ (2 / (n : ℝ))) ≤ closed_network_bound n and p ^ (1 / (n : ℝ)) + q ^ (1 / (n…

Formalized. mixture_representation_monotone_distribution (arXiv:2610.01563 Lemma 3): Let $P : \mathbb{N}^+ \to \mathbb{R}$ be a nonincreasing probability distribution on the positive integers, so that $P(k) \ge 0$ for all $k$, $P$ is antitone, and $\sum_{k=1}^\infty P(k) = 1$. Then the mixture weights $w_k = k(P(k) - P(k+1))$ satisfy: $w_k…

Formalized. dual_melas_non_degenerate_orbit_shift_equiv_iff (arXiv:2610.38017 Lemma 5): Let k ≥ 3, let F be a finite field of order 2^k, and let α ∈ F be primitive of order 2^k - 1. For binary polynomials f and g such that f is non-degenerate (neither α nor α⁻¹ is a root of f), f and g are shift-equivalent modulo the product of the minimal…

Formalized. grid_arrangement_column_deviation_subset_sum_bound (arXiv:2610.38386 Lemma 3.8): Let $n \ge 3$, $N = n^2$, and let $S : \mathrm{Fin}\ n \times \mathrm{Fin}\ n \to \mathbb{Z}$ be an injective arrangement with entries in $\{1, \dots, N\}$. For the column-deviation vector $\gamma_j = \sum_{i=0}^{n-1} S(i,j) - n(N+1)/2$ and any subset of…

Corrected. not_simple_graph_diam_le_two_mul_div_add_one_of_dist_two_transport_bound: The statement simple_graph_diam_le_two_mul_div_add_one_of_dist_two_transport_bound is false: The statement is false due to truncated natural number subtraction k - c : ℕ in the hypothesis hbound. Specifically, if c > 2 * k (e.g. k = 2, c = 5),…

Corrected. not_antipodal_hellinger_majority_three_ratio: The statement antipodal_hellinger_majority_three_ratio is false: The theorem statement claims that dictator_dist 3 majority_three = 1 / 4. However, dictator_dist n f is defined as `⨅ (g : (Fin n → Bool) → ℝ) (_ : is_signed_dictator n g), (1 / (2 : ℝ) ^ n)…

Open. open_problem_736 (arXiv:2610.35838 Sec. III.4): For every parameter alpha in (0, pi/2), all orthonormal frames e, f in O(3), and all setting indices mu, nu, mu', nu' in {0, 1, 2, 3}, the two-copy closed-network Bell measurement quantity Lambda(tau_{mu nu}, tau_{mu' nu'}) for the generalized GHZ state…

Open. magic_rectangle_algebraic_sos_identity: Let A be an ℝ-algebra and let D, C, B : Fin 3 → A satisfy D j ^ 2 = C j ^ 2 = B j ^ 2 = 1 for all j, D 0 * D 1 * D 2 = 1, C 0 * C 1 * C 2 = -1, pairwise commutativity within D and within C, and every D and C commuting with every B. Setting X j = D j + C j, Y…

Open. magic_rectangle_sos_certificate (arXiv:2610.12001 Theorem 3.1): Let H be a complex Hilbert space and let D, C, B : Fin 3 → (H →L[ℂ] H) be self-adjoint continuous linear operators satisfying D j ^ 2 = 1, C j ^ 2 = 1, and B j ^ 2 = 1 for all j, D 0 * D 1 * D 2 = 1, C 0 * C 1 * C 2 = -1, D j * D k = D k * D j, C j * C k = C…

Open. universal_spectral_bound_positive_trace_preserving (arXiv:2610.37187): For every integer d ≥ 3, every positive trace-preserving linear map Φ on d × d complex matrices satisfies the universal spectral bound Tr(Φ) ≤ d * m(Φ) + d^2 - d, where Tr(Φ) is the real part of the operator trace of Φ and m(Φ) is the minimum real part of an…

Open. pocket_depth_bound_from_omitted_max_types (arXiv:2610.37940 Theorem 16): Let B be a finite set of nonempty permutations of natural numbers from 1 to their length. If B contains no pattern whose maximum entry is at the leftmost position, or contains no pattern whose maximum entry is at the rightmost position, then every B-avoiding…

Open. op716_fourier_threshold_odd_eq_arithmetic_bound (arXiv:2610.38447): Open Problem OP-716: for every integer m >= 1, the Fourier threshold nu(2m+1) of functions from (ZMod 2)^(2m+1) to ZMod 3 with unique zero at 0 equals the arithmetic bound lambda_{2m+1}.

Verified in Lean

Monthly report

October 2026: Optimal Graph Partitions, Quantum Parity Bounds, and Tableau Enumeration Settled

This month settled open questions across discrete mathematics. Optimal graph cuts were shown to fail coalitional stability. Exact parity bounds for line cubic circuits were proved. A Latin tableau enumeration formula was established, and squares resist two-prime tile partitions.

Details

Disproved. Unweighted Max-3-Cut games can fail to admit any optimal strong equilibrium

Proved. Line cubic quantum syndrome has exact size 2n + 1 parity representation

Proved. Latin tableaux with double-three tail numbered by closed formula

Partial. Squares have no perfect partitions with two-prime tile areas

Lean accepted 3307 new theorems this month so far, drawn from 205 papers and 34 topics. Of these, 198 are new here, 8 correct a published claim, 484 formalize a published result and 2617 are helper lemmas. They make 17 research-level results, in 20 theorems -- they lead the report. 1808 statements posed this month are not yet proved. The reading covered 330 new papers. This report covers 8 results in 6 fields.

Daily reading

Daily reading: 2 research-level, 79 proved, 26 open

Research-level results today: 2. (medium tier) Proves the failure half of the golden-ratio conjecture OP-726 for every exponent, not just at sample points: for every beta in (0, (sqrt 5 - 1)/2) there is a strictly positive binary Markov chain Y <- X -> Z on which mutual power-divergence information is strictly superadditive. The construction is a rare input symbol of probability p and a channel entry of order p, and the slack has derivative -(1-beta)a^2 - (1-(1-a)^beta)^2 + 2 kappa^beta((a+kappa)^(1-beta) - (kappa+a-a^2)^(1-beta)) at p = 0, positive for suitable a, kappa exactly when beta^2 + beta < 1. The source paper (arXiv:2610.21231) gives only sufficient conditions for subadditivity, and the knowledgebase had refutations only at beta = 1/4 and 1/2 plus numerical evidence for the threshold; (low tier) settles OP-744 (Existence of a non-transitive cycle of length at least 12 for 11-sided dice). Each research-level result is one Research Blog post; theorems that bear on the same open problem or paper share it.

Details

The Proposer read 0 new paper(s) and listed 0 open problem(s); 0 new question(s) of AFTD's own were adopted onto the list; 10 list entries were stated in Lean, as were 8 results the papers prove; the Proposer posed 36 more from what the knowledgebase already holds (83 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 79 theorem(s), 30 of them helper lemmas, which are not listed below; 26 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. power_divergence_exists_binary_counterexample_of_lt_golden_ratio (knowledgebase): For every exponent beta in (0, (sqrt 5 - 1)/2), there exist strictly positive binary probability mass functions pX, PY, PZ forming a Markov chain Y <- X -> Z such that mutual power-divergence information subadditivity strictly fails. — (medium tier) Proves the failure half of the golden-ratio conjecture OP-726 for every exponent, not just at sample points: for every beta in (0, (sqrt 5 - 1)/2) there is a strictly positive binary Markov chain Y <- X -> Z on which mutual power-divergence information is strictly superadditive. The construction is a rare input symbol of probability p and a channel entry of order p, and the slack has derivative -(1-beta)a^2 - (1-(1-a)^beta)^2 + 2 kappa^beta((a+kappa)^(1-beta) - (kappa+a-a^2)^(1-beta)) at p = 0, positive for suitable a, kappa exactly when beta^2 + beta < 1. The source paper (arXiv:2610.21231) gives only sufficient conditions for subadditivity, and the knowledgebase had refutations only at beta = 1/4 and 1/2 plus numerical evidence for the threshold

Research-level. op744_dice11_has_directed_cycle_ge_twelve (arXiv:2610.40321): There exists an integer k ≥ 12 such that has_directed_cycle (dice_graph 11).Adj k holds. — (low tier) settles OP-744 (Existence of a non-transitive cycle of length at least 12 for 11-sided dice)

Partial result on OP-161. cycle_seven_all_powers_well_covered (arXiv:2610.09300): For every positive integer d, the d-th power of the 7-cycle graph C_7 is well-covered.

Partial result on OP-537. exists_planar_all_powers_well_covered_diam_eq_three (knowledgebase): There exists a finite connected simple planar graph G with at least two vertices whose diameter is exactly 3 and such that every positive integer power G^d (for d ≥ 1) is well-covered.

Partial result on OP-538. bipartite_all_powers_well_covered_diameter_le_two (knowledgebase): Every finite connected bipartite simple graph G with at least two vertices whose positive integer powers G^d (for d ≥ 1) are all well-covered has diameter at most 2.

Partial result on OP-593. cycle_not_proportionally_two_choosable (knowledgebase): For every integer n ≥ 3, the simple n-cycle C_n is not proportionally 2-choosable: it is not the case that for every finite color palette α and every assignment L of 2 colors to each vertex, there exists a proper coloring f of C_n with f(v) ∈ L(v) for all v…

Partial result on OP-726. mutual_power_divergence_not_subadditive_of_lt_golden_ratio (knowledgebase): For every exponent beta in (0, (sqrt 5 - 1)/2), the mutual power-divergence information with generator f_beta fails to be subadditive under Markov chains Y <- X -> Z.

Partial result on OP-729. op_729_case_n_two (knowledgebase): For every Boolean function f : (Fin 2 → Bool) → Bool on 2 variables, optimalTorusError f 0 = predictedTorusError 2 f and HasRationalTorusApproximation f 0 (predictedTorusError 2 f).

Partial result on OP-719. yang_mills_residual_root_m3 (knowledgebase): The rational residual polynomial yang_mills_residual_polynomial 3 evaluated at the rational number 1/21 equals 0.

Partial result on OP-786. min_avg_pairwise_hamming_dist_density_one_sixteenth_n_eq_four (arXiv:2610.40486): For $n = 4$, every subset $F \subseteq (\mathrm{Fin}\ 4 \to \mathrm{Bool})$ of cardinality $2^{4-4} = 1$ attains average pairwise Hamming distance $(4-4)/2 = 0$ and is a $0$-dimensional subcube is_subcube (4 - 4) F.

Thread RT-1. cycle_power_dyadic_class_weight_congr: The dyadic class weight cycle_power_dyadic_class_weight m b depends only on the values b i for i < 2^m: if b i = b' i for every i < 2^m, the two weights are equal.

Thread RT-1. cycle_power_dyadic_no_island_of_alternating_paired: Uniform induction step for the no-island class shape. Let m ≥ 1 and suppose every tight, exactly dyadically covered, fully ordinarily covered integer Boolean class model at exponent m with exactly m+1 active classes at each vertex has no island (a < b < c <…

Thread RT-1. simple_graph_equivalence_covering_number_cycle_power_two_pow_even_not_dvd_of_alternating_paired: Conditional resolution of OP-205: if, for every m ≥ 5, every tight, exactly dyadically covered, fully ordinarily covered integer Boolean class model at exponent m with exactly m+1 active classes per vertex has the alternating-paired structure (x (p+1) 1 =…

Thread RT-1. cycle_power_dyadic_tight_class_inner_imp_one: For m ≥ 2 and a tight dyadic class b : ℕ → Bool at exponent m (class weight equal to class size) with b 1 = b (2^m - 2) and b (2i) = b (2i+1) whenever 1 ≤ i and i + 1 < 2^(m-1): every inner bit forces bit 1, i.e. if 0 < j < 2^m - 1 and b j = true then b 1 =…

Thread RT-1. cycle_power_dyadic_alternating_paired_of_pairing: For m ≥ 2, a tight, exactly dyadically covered integer Boolean class model x : ℤ → ℕ → Bool at exponent m with x p 1 = x p (2^m - 2) for every p and the pairings x p (2i) = x p (2i+1) for 1 ≤ i, i + 1 < 2^(m-1), has the alternating-paired structure: x (p+1)…

Formalized. subdivided_k33_line_graph_isk4_diamond_free_four_chromatic (arXiv:2610.32404 Proposition 11.2, with H as constructed in Section 11.): Let H be the graph obtained from K3,3 by subdividing one edge exactly once, and let G be its line graph. Then G has no induced subdivision of K4, has no induced diamond, and has chromatic number exactly four.

Formalized. cylinder_construction_distance_set (arXiv:2610.34758 Lemma 2.3 -- Geometric progressions in distance sets in odd dimensions over prime fields): Let $p$ be an odd prime (with $p$ prime and $p$ odd) and let $m \ge 0$ be a natural number. There exist an $m$-dimensional submodule $W \le (\mathbb{Z}/p\mathbb{Z})^{2m+1}$ that is totally isotropic with respect to $Q_{2m+1}$ and a vector $e \in…

Formalized. finite_bi_algebra_is_semigroup_with_left_units (arXiv:2610.37108 Proposition 3.1): Every finite BI-algebra is a semigroup with B and I as left units: let (A, ·) be a finite magma equipped with elements B, I in A such that B is a B-element (((Bx)y)z = x(yz) for all x, y, z in A) and I is an I-element (Ix = x for all x in A). Then the binary…

Formalized. tournament_not_king_of_covered (arXiv:2610.40521 Observation 4): If a vertex w covers another vertex v in a tournament T (that is, w -> v and w -> x for every out-neighbor x of v), then v is not a king in T.

Formalized. tournament_is_king_of_max_outdegree (arXiv:2610.40521 Equation 2): In every tournament on n vertices, every vertex of maximum outdegree is a king: every other vertex is reachable from it by a directed path of length at most two.

Formalized. tournament_exists_covering_realization (arXiv:2610.40521 Lemma 6): Let d be a realizable degree vector of an n-vertex tournament, and let v, w be vertices satisfying d v < d w. Then there exists a tournament T realizing d in which w covers v (that is, w -> v and w -> x for every out-neighbor x of v).

Formalized. tournament_universal_kings_iff_max_outdegree (arXiv:2610.40521 Corollary 7): For every realizable degree vector d of an n-vertex tournament and every vertex v, v is a king in every tournament realizing d if and only if v has maximum outdegree according to d (that is, d u ≤ d v for all vertices u).

Formalized. tournament_directed_cut_eq_of_outdegrees_eq (arXiv:2610.40521 Lemma 2: Directed Cut-Query Complexity of Finding a King): Let T₁ and T₂ be tournaments on vertex set Fin n such that every vertex v has the same outdegree in T₁ and T₂. Then for every subset S ⊆ Fin n, the number of edges directed from S to its complement in T₁ equals that in T₂.

Open. yang_mills_additional_root_distinct_explicit (knowledgebase): For every natural number m ≥ 12, with d = 2m + 5, b_coeff = -2*(d^3 - 16*d^2 + 89*d - 146)/((d - 2)*(d - 1)^2), and c_coeff = ((d - 10)*(d - 5)/((d - 2)*(d - 1)))^2, let Q be the monic quadratic polynomial X^2 + b_coeff * X + c_coeff; if a real number z…

Open. min_avg_pairwise_hamming_dist_density_one_sixteenth (arXiv:2610.40486): Let $n \ge 4$ be a natural number. For any subset $F \subseteq (\mathrm{Fin}\ n \to \mathrm{Bool})$ of cardinality $2^{n-4}$, the average pairwise Hamming distance avg_pairwise_hamming_dist F is at least $(n-4)/2$, and equality holds if and only if $F$ is…

Open. subcube_avg_pairwise_hamming_dist (arXiv:2610.40486): For any $n \ge 4$, every $(n-4)$-dimensional subcube $F \subseteq (\mathrm{Fin}\ n \to \mathrm{Bool})$ of cardinality $2^{n-4}$ has average pairwise Hamming distance equal to $(n-4)/2$.

Open. op_776_explicit_polynomial_hitting_small_char (arXiv:2610.37677): There exist constants C, k : ℕ such that for all natural numbers m ≥ 1, S ≥ 1, and every prime p ≤ ((m+2)S choose 2), for every field F of characteristic p, there exist a matrix order R ≤ C * (m + S)^k and a tuple of matrices T : Fin m → Matrix (Fin R) (Fin…

Open. op_776_char_two_hitting_point (arXiv:2610.37677): There exist constants C, k : ℕ such that for all natural numbers m ≥ 1, S ≥ 1, and every field F of characteristic 2, there exist a matrix order R ≤ C * (m + S)^k and a tuple of matrices T : Fin m → Matrix (Fin R) (Fin R) F such that T is an ncABP hitting…

Open. graph_interval_index_path_graph (arXiv:2610.39966 Theorem 5): For every n ≥ 1, the interval index of the path graph P_n equals (n^3 - n) / 6.

Verified in Lean

Daily reading

Daily reading: 2 research-level, 189 proved, 43 open

Research-level results today: 2. (medium tier) Settles the k = 15 case of the residual even case of arXiv:2610.15635: eq(C_n^15) = 6 for every even n ≥ 3825 with 16 ∤ n, where the paper only has eq(C_n^15) ∈ {5, 6}; the first case beyond k = 3 where the exact value is now known; (medium tier) Settles the k = 7 case of the residual even case of arXiv:2610.15635: eq(C_n^7) = 5 for every n ≥ 98 with 8 ∤ n, where the paper only has eq(C_n^7) ∈ {4, 5} for even n ≥ 441. Each research-level result is one Research Blog post; theorems that bear on the same open problem or paper share it.

Details

The Proposer read 56 new paper(s) and listed 71 open problem(s); 59 new question(s) of AFTD's own were adopted onto the list; 18 list entries were stated in Lean, as were 11 results the papers prove; the Proposer posed 38 more from what the knowledgebase already holds (205 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 189 theorem(s), 139 of them helper lemmas, which are not listed below; 43 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. simple_graph_equivalence_covering_number_cycle_power_fifteen_of_not_dvd (arXiv:2610.15635): For every even n ≥ 3825 = 15²·17 not divisible by 16, the equivalence covering number of the 15th power of the n-cycle is 6 = 2 + ⌈log₂ 16⌉. This settles the case k = 15 (m = 4) of the open problem: the paper's Theorems 1.3-1.4 leave eq(C_n^15) ∈ {5, 6} open… — (medium tier) Settles the k = 15 case of the residual even case of arXiv:2610.15635: eq(C_n^15) = 6 for every even n ≥ 3825 with 16 ∤ n, where the paper only has eq(C_n^15) ∈ {5, 6}; the first case beyond k = 3 where the exact value is now known

Research-level. simple_graph_equivalence_covering_number_cycle_power_seven_of_not_dvd (arXiv:2610.15635): For every n ≥ 98 not divisible by 8, the equivalence covering number of the 7th power of the n-cycle is 5 = 2 + ⌈log₂ 8⌉. This settles the case k = 7 (m = 3) of the open problem: for even n ≥ 441 not divisible by 8 the paper's Theorems 1.3-1.4 only give… — (medium tier) Settles the k = 7 case of the residual even case of arXiv:2610.15635: eq(C_n^7) = 5 for every n ≥ 98 with 8 ∤ n, where the paper only has eq(C_n^7) ∈ {4, 5} for even n ≥ 441

Partial result on OP-601. four_cycle_graph_code_density_tendsto_one_along_powers_of_two (knowledgebase): As the natural number r tends to infinity, the sequence ((2 : ℝ) ^ r) * kst_code_max_density 2 2 (2 ^ r) tends to 1.

Partial result on OP-641. edge_geodesic_transversal_number_corona_k1_card_two (arXiv:2610.36821): For every finite type V with Fintype.card V = 2 and every connected simple graph G on V, the edge geodesic transversal number of the corona product corona_k1 G satisfies edge_geodesic_transversal_number (corona_k1 G) = 1.

Partial result on OP-665. thue_morse_word_length (knowledgebase): For every natural number n, the length of the n-th Thue-Morse word t_n is 2^n.

Partial result on OP-665. thue_morse_random_attractor_prob_zero (knowledgebase): For every natural number n, thue_morse_random_attractor_prob n 0 = 0.

Partial result on OP-605. cycle_graph_six_book_ramsey_witness (arXiv:2610.36522): In the 6-cycle graph on Fin 6, every pair of adjacent vertices has at most 0 common neighbors, and every pair of distinct non-adjacent vertices has at most 1 common neighbor in the complement graph.

Partial result on OP-605. book_ramsey_b1_b2_graph_exists (arXiv:2610.36522): There exists a simple graph G on 6 vertices with decidable adjacency such that every pair of adjacent vertices has at most 0 common neighbors in G, and every pair of distinct non-adjacent vertices has at most 1 common neighbor in the complement graph Gᶜ.

Partial result on OP-561. online_ef1_approximation_ratio_nonneg (arXiv:2610.34625): For any number of agents n, the online EF1 approximation ratio under exact frequency predictions is nonnegative.

Partial result on OP-639. op_639_exists_intersecting_four_uniform_covering_four_card_ge_forty_two (arXiv:2610.37315): There exists an intersecting 4-uniform hypergraph on the natural numbers with covering number 4 having at least 42 edges.

Thread RT-1. cycle_power_fifteen_tight_type_sequence_shape_rules_core_v30_41: Window refutation for k = 15: in the layer-free tight-class model on the integer line (x p j: the class starting at p contains p + j; inner count 3 per vertex, distances 14, 12, 8 covered exactly once, the other distances covered, classes closed under the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_shape_rules_core_v40_51: Window refutation for k = 15: in the layer-free tight-class model on the integer line (x p j: the class starting at p contains p + j; inner count 3 per vertex, distances 14, 12, 8 covered exactly once, the other distances covered, classes closed under the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_shape_rules_core_v50_61: Window refutation for k = 15: in the layer-free tight-class model on the integer line (x p j: the class starting at p contains p + j; inner count 3 per vertex, distances 14, 12, 8 covered exactly once, the other distances covered, classes closed under the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_shape_rules_core_v60_71: Window refutation for k = 15: in the layer-free tight-class model on the integer line (x p j: the class starting at p contains p + j; inner count 3 per vertex, distances 14, 12, 8 covered exactly once, the other distances covered, classes closed under the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_shape_rules_core_v81_90: Window refutation for k = 15: in the layer-free tight-class model on the integer line (x p j: the class starting at p contains p + j; inner count 3 per vertex, distances 14, 12, 8 covered exactly once, the other distances covered, classes closed under the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_shape_rules_core_v91_100: Window refutation for k = 15: in the layer-free tight-class model on the integer line (x p j: the class starting at p contains p + j; inner count 3 per vertex, distances 14, 12, 8 covered exactly once, the other distances covered, classes closed under the…

Formalized. seidel_matrix_codegree_eq (arXiv:2610.36522 Lemma 7): For any simple graph G on a finite vertex set V with Seidel matrix S and any pair of adjacent vertices u and v in G, four times the number of common neighbors of u and v equals |V| - 4 - ∑_k S_{uk} - ∑_k S_{vk} + (S^2)_{uv}.

Formalized. seidel_matrix_compl_codegree_eq (arXiv:2610.36522 Lemma 7): For any simple graph G on a finite vertex set V with Seidel matrix S and any pair of adjacent vertices u and v in the complement Gᶜ, four times the number of common neighbors of u and v in Gᶜ equals |V| - 4 + ∑_k S_{uk} + ∑_k S_{vk} + (S^2)_{uv}.

Formalized. admissible_card_sq_add_five_le (arXiv:2610.34372, Lemma 2.1): If A is a nonempty admissible subset of {1, ..., N}, then the square of the cardinality of A plus 5 is at most 6 times N.

Formalized. admissible_card_le_nat_sqrt (arXiv:2610.34372, Lemma 2.1): If A is a nonempty admissible subset of {1, ..., N}, then the cardinality of A is at most the integer square root of 6 times N.

Formalized. recurrence_b_transition_bijective_and_orbit_periodic_and_unique_fixed_point (arXiv:2610.34300 Proposition 7): For any integer k ≥ 2, the transition map of recurrence (B) on (ℤ/3ℤ)^k is bijective, every orbit is purely periodic (for every state w, there exists m > 0 such that the m-th iterate of the map on w is w), and the zero state is the unique fixed point of the…

Formalized. eefx_imp_approx_efk (arXiv:2610.34625 Lemma 3.2 -- Online Fair Division with Frequency Predictions): Let n ≥ 3 and let σ be an allocation of m goods among n agents with nonnegative additive valuations u such that every agent's bundle is nonempty. If σ is an EEFX allocation, then σ is a 1/(n-1)-EF(n-2) allocation.

Formalized. petersen_colorable_of_three_edge_colorable (arXiv:2610.35794 (Computational complexity of the H-coloring problem in cubic graphs), Sec. 3 (H-coloring for 3-edge-colorable H reduces…): If es is an edge list with edge_list_connected_cubic n es that has a proper 3-edge-coloring c : Fin es.length → Fin 3 (any two distinct edges i ≠ j sharing an endpoint get different colors), then it has a Petersen coloring: edge_list_h_colorable…

Formalized. kst_code_linear_construction_finite_field (arXiv:2610.35136 Lemma 4.2): Let n, s, t, r ≥ 1 be integers, and let w_1, ..., w_n in (ZMod 2)^r be vectors such that the sum of w_i over S is nonzero for every subset S of [n] with |S| in {s, t}. Then there exists a binary linear K_{s,t}-code C in the set of simple graphs on Fin n of…

Corrected. not_not_isContained_book_graph_of_commonNeighbors_le: The statement not_isContained_book_graph_of_commonNeighbors_le is false: The statement is FALSE for k = 0. Under Lean's Nat subtraction, 0 - 1 = 0. When k = 0, the hypothesis h requires that every pair of adjacent vertices u, v in G has at most 0 - 1 =…

Corrected. not_card_fiber_one_eq_one: The statement card_fiber_one_eq_one is false: The statement is false. The hypotheses require |A| = |P| + 1, f(A) ⊆ {c0} ∪ P, and that every c ∈ P has at least one preimage in A. However, there is no hypothesis requiring c0 to have any preimage in A.…

Open. radius_independence_connectivity_hamiltonian_interpolation (knowledgebase): Every finite simple graph G on at least three vertices with vertex connectivity kappa(G) >= 2 and independence number alpha(G) <= kappa(G) + rad(G) - 2 contains a Hamilton cycle.

Open. radius_independence_connectivity_hamiltonian_radius_two (knowledgebase): Every finite simple graph G on at least three vertices with vertex connectivity kappa(G) >= 2, radius 2, and independence number at most kappa(G) contains a Hamilton cycle.

Open. thue_morse_random_attractor_lattice_gumbel (knowledgebase): The sequence in n of the supremum over all k in ℕ of |thue_morse_random_attractor_prob n k - Real.exp (-8 * (n : ℝ) * (5 / 6 : ℝ) ^ k)| tends to 0 as n tends to infinity (in Filter.atTop and nhds 0).

Open. op_605_ramsey_b63_b64_graph_exists (arXiv:2610.36522): There exists a simple graph G on 254 vertices with decidable adjacency such that every pair of adjacent vertices has at most 62 common neighbors in G, and every pair of distinct non-adjacent vertices has at most 63 common neighbors in the complement graph Gᶜ.

Open. amply_regular_icosahedron_cartesian_power (arXiv:2610.36620 Theorem 3.6 -- Diameter Bounds and Finiteness of Graphs from Local Intersection Data): For every integer m ≥ 2, the m-fold Cartesian power of the icosahedron graph is an amply regular graph with parameters (v, k, λ, μ) = (12^m, 5m, 2, 2) and diameter 3m, satisfying μ > 1 and diam(G) > 2k/(λ+2), while having order strictly greater than 12.

Open. seidel_matrix_book_ramsey_lower_bound (arXiv:2610.36522 Corollary 8): Let n ≥ 2 be an integer, and let G be a simple graph on a finite vertex set V of cardinality 4n - 2 with Seidel matrix S. If every row sum of S is 1 and (S^2)_{ij} ≤ 0 for all distinct vertices i ≠ j, then every pair of adjacent vertices in G has at most n -…

Verified in Lean

Daily reading

Daily reading: 0 research-level, 194 proved, 31 open

No research-level result today. For generalized k-server on k uniform m-point metrics, the list entry OP-613 gained an expected cost bound of (k + 1) ln(2k) times optimal from arXiv:2610.37192, as well as competitive ratio 1 for k = 1. In graph coverings and social choice, thread RT-1 and the list entry OP-205 saw reductions forcing upper-vertical defects from prefix defects alongside an average-barrier countermodel, while the list entry OP-19 gained voter-partition conditions for half-majority challengers. Fully formalized theorems established that finite BMI-algebras are idempotent semigroups (arXiv:2610.37108), intersecting k-uniform hypergraphs with covering number k have cardinality at most k^k (arXiv:2610.37315), and (n - 1)-term subsequence sums of maximal zero-sum-free multisets fill the group (arXiv:2610.35819). No new papers were read today, seventeen questions proposed here were listed as OP-647 through OP-663, and uniform_kserver_randomized_competitive remains open.

Details

The Proposer read 0 new paper(s) and listed 0 open problem(s); 17 new question(s) of AFTD's own were adopted onto the list; 13 list entries were stated in Lean, as were 10 results the papers prove; the Proposer posed 48 more from what the knowledgebase already holds (168 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 194 theorem(s), 142 of them helper lemmas, which are not listed below; 31 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Partial result on OP-613. gen_k_server_uniform_log_competitive (arXiv:2610.37192): For every k ≥ 1, every m and every start configuration a : Fin k → Fin m of generalized k-server on k uniform metrics with m points each (unit distances), there are transition probabilities P with gen_k_server_uniform_is_rand_alg P such that for every finite…

Partial result on OP-613. gen_k_server_uniform_one_server_competitive_one (arXiv:2610.37192): With one server (k = 1, configurations Fin 1 → Fin m) and any start configuration a, there are transition probabilities P with gen_k_server_uniform_is_rand_alg P whose expected cost on every request list ρ is at most gen_k_server_uniform_opt a ρ.

Partial result on OP-205. cycle_power_sixteen_block_vertical_average_barrier (arXiv:2610.15635): Eight tight radius16 dyadic profiles with positive integer weights totaling60 satisfy optimal mean size5, all exact dyadic pair averages, every ordinary pair-coverage average, and monotone outer eighths; the weighted upper-vertical defect at offsets1/3…

Partial result on OP-205. cycle_power_exact_rank_boundary_successor (arXiv:2610.15635): For an exact dyadic-rank matching with both parent implications, if a root selects the last lower offset2q-1 and the edge one unit longer than that rank is covered, its owner has offset a<q and root at least p+q. Unless a=q-1, it also owns the next…

Partial result on OP-205. cycle_power_rooted_reflect_exact (arXiv:2610.15635): Reflection of rooted profiles preserves exact pair ownership at every rank length L with0<L<=D; the physical starting point changes from v to -v-(D-L).

Partial result on OP-205. cycle_power_rooted_reflect_cover (arXiv:2610.15635): Reflection preserves coverage of every ordinary distance strictly between0 and the arc length.

Partial result on OP-205. cycle_power_rooted_reflect_parent (arXiv:2610.15635): Reflection preserves both parent implications at every even rank2q, interchanging the lower and upper branches.

Partial result on OP-205. cycle_power_sixteen_block_geometry_reflect (arXiv:2610.15635): The full radius16h Geometry hypotheses, including all ordinary distances and exact12h/14h/15h ownership, are preserved by simultaneous root and offset reflection.

Thread RT-1. cycle_power_fifteen_tight_type_sequence_eight_iff_nine_core_a01111001: One branch of the window refutation behind cycle_power_fifteen_tight_type_sequence_eight_iff_nine_core: under the layer-free tight-cover constraints for k = 15 and 7 fixed bits of the class-type sequence near position 0, the constraints are contradictory.…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_one_alternates_core_alt00000100: One branch of the window refutation behind cycle_power_fifteen_tight_type_sequence_one_alternates_core: under the layer-free tight-cover constraints for k = 15 and 95 fixed bits of the class-type sequence near position 0, the constraints are contradictory.

Thread RT-1. cycle_power_sixteen_block_no_vertical_overlap: For any positive h and Boolean rooted model of radius16h satisfying the explicit Geometry hypotheses (selected endpoints, three parent matchings, exact12h/14h/15h ownership, monotone outer eighths, full ordinary coverage), suppose globally every selected C_t…

Thread RT-1. cycle_power_sixteen_block_child_forces_full_prefix: Under the explicit Geometry hypotheses and the global no-vertical-defect condition C_t implies A_t, every root with any selected C_t has its entire first2h prefix selected.

Thread RT-1. cycle_power_sixteen_block_overlap_reaches_upper_vertical: Under the explicit Geometry hypotheses, every overlap A_t,C_t selected and A_s missing with t<s<2h forces some upper-half vertical defect: at another integer root r some h<=a<2h selects C_a but misses A_a.

Thread RT-1. cycle_power_dyadic_sixteen_block_geometry: Every tight dyadic rooted model of radius2^(k+4), with all-rank exactness, full ordinary coverage and monotone outer eighths, satisfies the explicit sixteen-block Geometry conditions at h=2^k. This is a faithful bridge with no graph-covering conclusion.

Formalized. finite_bmi_algebra_is_idempotent_semigroup_with_left_units (arXiv:2610.37108 Theorem 3.3): Every finite BMI-algebra is an idempotent semigroup with B, M, and I as left units: let $(A, \cdot)$ be a finite magma equipped with elements $B, M, I \in A$ such that $B$ is a $B$-element ($((Bx)y)z = x(yz)$ for all $x, y, z \in A$), $M$ is an $M$-element…

Formalized. edge_cut_transversal_theorem (arXiv:2610.36821 Theorem 5): Let G be a connected simple graph. Let S be a finite set of edges of G of cardinality k that is an edge cut induced by a vertex subset X (meaning there exists X : Set V such that for all e, e ∈ S ↔ e ∈ G.edgeSet ∧ ∃ u ∈ X, ∃ v ∉ X, e = s(u, v)), such that…

Formalized. edge_geodesic_transversal_bound_of_edge_cut (arXiv:2610.36821 Lemma 3): Let G be a connected simple graph. Let S be a finite set of edges of G that is an edge cut induced by a vertex subset X (meaning there exists X : Set V such that for all e, e ∈ S ↔ e ∈ G.edgeSet ∧ ∃ u ∈ X, ∃ v ∉ X, e = s(u, v)). If for all vertices u, v and…

Formalized. intersecting_uniform_hypergraph_card_le_pow (arXiv:2610.37315 Theorem 1.1 -- The maximum size of intersecting k-graphs with covering number k): Let α be a type with decidable equality and let k ≥ 1 be a natural number. Let ℱ be a finite family of finite subsets of α such that every F ∈ ℱ has cardinality k. Suppose that for all F1, F2 ∈ ℱ, the intersection F1 ∩ F2 is nonempty, and that for every…

Formalized. is_maximal_zero_sum_free_subsequence_sums_pred_eq_univ (arXiv:2610.35819 Section 8): Let G be a finite additive commutative group with decidable equality, let n ≥ 1 be a natural number, and let S be a multiset over G satisfying is_maximal_zero_sum_free n S. Then the finite set of (n - 1)-term subsequence sums of S, subsequence_sums (n - 1)…

Formalized. subsequence_sums_maximal_zero_sum_free_term_mul_pred_eq (arXiv:2610.35819 Proposition 8.2): Let G be a finite additive commutative group with decidable equality and exponent n (AddMonoid.exponent G = n). Let S be a multiset over G satisfying is_maximal_zero_sum_free n S. If some element t of G has multiplicity n - 1 in S (S.count t = n - 1), then…

Formalized. regular_graph_lap_mul_pow_scalar_sub_all_ones_trace (arXiv:2610.36127 Lemma 34): For each d-regular graph G on n vertices and each pair of positive integers i and j, the trace of L(G)^i ((d+1)I_n - J_n)^j equals (d+1)^j ℓ_i(G), where J_n is the all-ones matrix and ℓ_i(G) is the Laplacian trace sequence.

Corrected. not_edge_geodesic_transversal_bound_of_edge_disjoint: The statement edge_geodesic_transversal_bound_of_edge_disjoint is false: The statement is false as formalized because it omits a finiteness hypothesis on G (the paper's Lemma 4 is about finite graphs; edge_geodesic_transversal_number is an infimum over…

Corrected. not_div_padicValNat_dvd_div_padicValNat: The statement div_padicValNat_dvd_div_padicValNat is false: The statement is false for non-prime p. Take d = 6, n = 48, p = 4. Then 0 < 6 and 6 ∣ 48. Since 4 ∤ 6, padicValNat 4 6 = 0, so d / 4^{padicValNat 4 6} = 6. Since 4^2 = 16 divides 48 but 4^3 = 64…

Open. uniform_kserver_randomized_competitive (arXiv:2610.37192 Theorem 1.1 -- Tight Bounds for Randomized Generalized k-Server on Uniform Metrics): Let k be a positive integer, let M be a family of uniform metric spaces indexed by Fin k with decidable equality, and let a be an initial configuration. For every finite request sequence ρ, the expected cost uniform_kserver_expected_cost k a ρ of the…

Open. amply_regular_diameter_bound_conjecture (arXiv:2610.36620): Conjecture 3.7 of arXiv:2610.36620: every amply regular graph G on a finite vertex set with parameters (k, ℓ, μ) and μ > 2 has diameter at most 2 * k / (ℓ + 2).

Open. amply_regular_diameter_bound_qpk (arXiv:2610.36620): Theorem 1.3 of arXiv:2610.36620: every amply regular graph G on a finite vertex set with parameters (k, ℓ, μ) satisfying μ > 1 and μ ≥ ℓ has diameter at most 2 * k / μ + 1.

Open. binary_nonzero_common_difference_octuple_count: In any finite commutative exponent-two group with at least8 elements, the number of injective nonzero ordered8-tuples satisfying the stated zero-sum/common-nonzero-difference constraint equals the stated product of group-cardinality factors.

Open. digraph_five_strong_semicomplete_multipartite_two_linked (arXiv:2610.36895): Problem 5.2 (OP-621): Every 5-strong semicomplete multipartite digraph is 2-linked.

Open. digraph_five_strong_co_bipartite_unilaterally_two_linked (arXiv:2610.36895): Every 5-strong co-bipartite digraph is unilaterally 2-linked.

Verified in Lean

Daily reading

Daily reading: 3 research-level, 233 proved, 38 open

Research-level results today: 3. (medium tier) settles OP-512 (Exact enumeration of Latin tableaux with a double-three staircase tail); (medium tier) Proves the conjectured equality for every n>=30000001 and every minimum proper edge coloring, establishing sharpness at all sufficiently large orders rather than only the paper's n<=12 range, the isolated n=279 case, or one congruence class; (medium tier) Answers the paper's Question 2 negatively for k = 3: a 182-vertex graph (the paper's 91-vertex G_3 with every vertex doubled into independent twins, plus pin edges from each core twin to two anchors) has an optimal Max-3-Cut coloring that is unique up to color permutation and not a strong equilibrium, so no optimal coloring is a strong equilibrium. The paper only shows that some optimal coloring can fail. Each research-level result is one Research Blog post; theorems that bear on the same open problem or paper share it.

Details

The Proposer read 20 new paper(s) and listed 31 open problem(s); 23 new question(s) of AFTD's own were adopted onto the list; 16 list entries were stated in Lean, as were 9 results the papers prove; the Proposer posed 38 more from what the knowledgebase already holds (130 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 233 theorem(s), 58 of them helper lemmas, which are not listed below; 38 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. latin_double_three_tail_count_formula: The open formula in arXiv:2610.29291 Conjecture 6.3: the number of Latin tableaux of shape (k,k-1,...,4,3,3) is 9 times 4 to the power k-3 for every k at least 5. — (medium tier) settles OP-512 (Exact enumeration of Latin tableaux with a double-three staircase tail)

Research-level. wild_number_complete_graph_large_eq (arXiv:2610.22142): For every integer n>=30000001, every natural number k equal to the chromatic number of the line graph of K_n, and every proper edge coloring c of K_n with color set Fin k, the wild number of c equals floor(4*ceil(n/2)/3). Thus the minimum-proper-coloring… — (medium tier) Proves the conjectured equality for every n>=30000001 and every minimum proper edge coloring, establishing sharpness at all sufficiently large orders rather than only the paper's n<=12 range, the isolated n=279 case, or one congruence class

Research-level. maxcut_op89_optimal_strong_equilibrium_negation (arXiv:2610.04948): It is not true that for every finite type V, simple graph G on V with decidable adjacency, and k ≥ 3, the unweighted Max-k-Cut game on G has an optimal coloring that is a strong equilibrium (the negation of maxcut_op89_optimal_strong_equilibrium_exists); the… — (medium tier) Answers the paper's Question 2 negatively for k = 3: a 182-vertex graph (the paper's 91-vertex G_3 with every vertex doubled into independent twins, plus pin edges from each core twin to two anchors) has an optimal Max-3-Cut coloring that is unique up to color permutation and not a strong equilibrium, so no optimal coloring is a strong equilibrium. The paper only shows that some optimal coloring can fail

Partial result on OP-539. third_principal_specialization_card_eq (knowledgebase): The third principal specialization of any finite tree on at least one vertex uniquely determines its number of vertices.

Partial result on OP-545. tournament_near_regular_card_degrees (knowledgebase): In any near-regular tournament on 2m vertices with m ≥ 1, exactly m vertices have out-degree m - 1 and exactly m vertices have out-degree m.

Partial result on OP-517. op_517_zero_excess_of_eq_pair (arXiv:2610.03808): For any instance inst of WiringDiagramPair, if the two signotopes inst.u and inst.v are equal, then inst has braid-move excess 0 (wiring_diagram_zero_excess inst): the distance is 0 and the number of differing triples is 0.

Partial result on OP-517. op_517_zero_excess_of_le_two (arXiv:2610.03808): For any instance inst of WiringDiagramPair on m pseudolines with m ≤ 2, there are no ordered triples in Fin m, so inst.u = inst.v holds vacuously and therefore inst has braid-move excess 0 (wiring_diagram_zero_excess inst).

Partial result on OP-593. cycle_three_proportionally_three_choosable (knowledgebase): The triangle cycle C_3 is proportionally 3-choosable: for every finite color palette α and every assignment L of 3 colors to each vertex, there exists a proper coloring f of C_3 with f(v) ∈ L(v) for all v such that for every color c : α, the number of…

Partial result on OP-595. strong_odd_five_coloring_of_list_colorable (knowledgebase): If a finite simple graph is strongly odd 5-list-colorable, then it admits an ordinary strong odd 5-coloring.

Partial result on OP-596. maxcut_optimal_no_clique_deviation (knowledgebase): In the unweighted Max-k-Cut game on a finite simple graph, no coalition that induces a clique has a strong deviation from a welfare-optimal coloring.

Partial result on OP-596. chordal_maxcut_optimal_strong_equilibrium_of_card_le (knowledgebase): For every finite simple chordal graph G on at most 2k + 1 vertices and every integer k ≥ 3, the unweighted Max-k-Cut game admits an optimal coloring that is a strong equilibrium.

Thread RT-1. simple_graph_cycle_power_dyadic_cover_class_model: Every optimal dyadic cycle-power equivalence cover admits a faithful periodic integer class model with the original rooted classes and real layer colours, tight class weights, constant active counts, exact dyadic coverage, all ordinary coverage, and disjoint…

Thread RT-1. simple_graph_cycle_power_dyadic_cover_fibre_endpoint_unique: In any clique labeling of the cycle power C_n^{2^m-1} with n ≥ 3(2^m-1)+1, each clique contains at most one pair at distance 2^m-1.

Thread RT-1. simple_graph_cycle_power_dyadic_cover_layer_root_unique: In an optimal (m+1)-layer cover of C_n^{2^m-1}, for each layer s and each vertex v, there is a unique offset a ≤ 2^m-1 such that v-a is the root spanning the class of v in layer s.

Thread RT-1. simple_graph_cycle_power_dyadic_cover_edge_span: If a monochromatic clique contains v and v+d with 1 ≤ d < 2^m, its root v-a (with a ≤ 2^m-1) satisfies a < 2^m - d, so the entire edge {v, v+d} is contained in the span of v-a.

Thread RT-1. cycle_power_dyadic_class_weight_eq_of_sum_eq: If the total sum of dyadic class weights over all residues equals the total sum of active counts, then pointwise equality holds for every residue.

Thread RT-1. simple_graph_cycle_power_dyadic_class_model_disjoint_colours: Two active rooted classes at a common vertex with distinct offsets have distinct layer colours.

Formalized. counterexample_partition_diameter_ratio_three_101 (arXiv:2610.28828 -- A counterexample to Gale's conjecture in $\mathbb{R}^3$, Theorem 2): There exists a finite subset V of 3-dimensional Euclidean space of cardinality 101 and strictly positive diameter such that for every 4-partition of V (given by a map c : V → Fin 4), at least one part has diameter strictly greater than 0.88928 times the…

Formalized. six_circle_two_symbols_labeled_impossible (arXiv:2610.33032, Theorem 8): In the anonymous-slip puzzle for six sages seated in a circle with six labeled seats, there is no deterministic protocol c : (Fin 6 × Fin 6 × Fin 6 × Fin 6) → Fin 2 using two symbols {0, 1} such that every sage can reconstruct the full seating permutation…

Formalized. kst_code_density_odd_odd (arXiv:2610.35136, Sharp Bounds for Graph-Codes Avoiding Complete Bipartite Differences, Proposition 3.2): Let s and t be odd natural numbers, and let n be a natural number such that s + t ≤ n. Then the maximum density of a K_{s,t}-code in simple graphs on vertex set [n], kst_code_max_density s t n, is equal to 1/2.

Formalized. partition_family_A_57_evaluation (arXiv:2610.01747 Proposition 6.1): Evaluating the partition cardinality formula with the partition numbers p(54)=386155, p(53)=329931, p(52)=281589, p(51)=239943, p(50)=204226, p(49)=173525, and p(48)=147273 gives 527254, which strictly exceeds 526823 = p(56).

Formalized. unrestricted_partition_family_four_intersecting_and_card (arXiv:2610.01747 Proposition 6.1): For every natural number n ≥ 1, the family partition_A n of partitions of n is 1-intersecting (is_t_intersecting_partition_family 1 (partition_A n)), and its cardinality coerced to ℤ satisfies ((partition_A n).card : ℤ) = partition_number (n - 3 : ℤ) +…

Formalized. difference_squares_avoiding_gp3_bounds (arXiv:2610.34758 Lemma 3.1): Let $p \ge 7$ be a prime and let $A \subset \mathbb{F}_p$ be nonempty. If the set of squared differences $(A - A)^2$ contains no nondegenerate three-term geometric progression, then $|A|^3 \le p^2$, $|A - A| \ge |A|^{5/4}$, and for every $t \in…

Formalized. bmi_is_left_unit_b (arXiv:2610.37108 Proposition 3.1 -- Finite BM-algebras and fixed-point combinators): Let $(A, \cdot)$ be a finite magma equipped with a $B$-element $B$ and an $I$-element $I$. Then $B$ is a left unit: $Bx = x$ for all $x \in A$.

Formalized. bmi_assoc_of_is_b_of_left_unit (arXiv:2610.37108 Proposition 3.1 -- Finite BM-algebras and fixed-point combinators): Let $(A, \cdot)$ be a magma with a $B$-element $B$. If $B$ is a left unit ($Bx = x$ for all $x$), then the operation is associative: $(xy)z = x(yz)$ for all $x, y, z \in A$.

Proved. maxcut_op89_residual_exists_negation: The residual case maxcut_op89_residual_exists of OP-89 is false: it is not true that every finite simple graph G with k ≥ 3, at least 2k+2 vertices, not k-colorable and with a vertex of degree at least 2k, has an optimal Max-k-Cut coloring that is a strong…

Open. cycle_proportional_choice_number_eq_three (knowledgebase): For every integer n ≥ 3, the simple n-cycle C_n is proportionally 3-choosable: for every finite color palette α and every assignment L of 3 colors to each vertex, there exists a proper coloring f of C_n with f(v) ∈ L(v) for all v such that for every color c…

Open. cycle_not_proportionally_two_choosable (knowledgebase): For every integer n ≥ 3, the simple n-cycle C_n is not proportionally 2-choosable: it is not the case that for every finite color palette α and every assignment L of 2 colors to each vertex, there exists a proper coloring f of C_n with f(v) ∈ L(v) for all v…

Open. unicyclic_spectral_sustainability_maximum_degree (knowledgebase): For every integer n >= 9, the maximum degree among spectrally sustainable connected unicyclic simple graphs of order n is equal to ceil(n/2), represented as (n + 1) / 2: every such graph satisfies Delta(G) <= ceil(n/2) and the bound is attained.

Open. sparse_strong_odd_five_list_coloring (knowledgebase): Every finite simple graph G with maximum average degree strictly less than 7/3 (that is, satisfying has_mad_lt G (7 / 3)) is strongly odd 5-list-colorable.

Open. chordal_maxcut_optimal_strong_equilibrium_exists (knowledgebase): For every finite simple chordal graph G and every integer k ≥ 3, the unweighted Max-k-Cut game on G admits an optimal coloring that is a strong equilibrium.

Open. even_odd_cycle_strong_product_equivalence_cover_eq_five (knowledgebase): For every even integer m ≥ 6 and odd integer n ≥ 5, the equivalence covering number of the strong product of cycles C_m and C_n (the strong product of simple_graph_cycle_power m 1 and simple_graph_cycle_power n 1) is equal to 5.

Verified in Lean

Daily reading

Daily reading: 2 research-level, 296 proved, 77 open

Research-level results today: 2 (3 theorems). (medium tier) Determines n♭_h(2) exactly for all odd h ≥ 5 except h ≡ 59 (mod 60), where arXiv:2610.19470 only shows n♭_h(2) < n_h(2): the gap n_h(2) - n♭_h(2) is 1 iff h ≡ 9 (mod 12), 2 for h ≡ 1, 3, 5, 7 (mod 12) and 3 for h ≡ 11 (mod 12); for h ≡ 59 (mod 60) it is at least 4, and it is at least 6 at h = 419; (medium tier) a new question answered: For h ≡ 59 (mod 60) the gap is at least 4: at range n_h(2) - 3 the Theorem 4.1 inequalities leave only sets with gcd(x, y) > 1 or a divisibility contradiction; (low tier) Excludes perfect partitions of every integer square with common tile area q^s*r*p^a for distinct primes p,q,r with p>r+1, uniformly in a and s. This extends the two-prime common-area obstruction by using the superincreasing signed-height argument. Each research-level result is one Research Blog post; theorems that bear on the same open problem or paper share it.

Details

The Proposer read 22 new paper(s) and listed 28 open problem(s); 28 new question(s) of AFTD's own were adopted onto the list; 17 list entries were stated in Lean, as were 9 results the papers prove; the Proposer posed 39 more from what the knowledgebase already holds (281 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 296 theorem(s), 119 of them helper lemmas, which are not listed below; 77 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. postage_stamp_number_neg_two_of_odd_of_mod_sixty_ne: Exact value of the negative-allowed three-stamp number for odd h ≥ 5 with h ≢ 59 (mod 60): n_h(2) - n♭_h(2) = 1 if h ≡ 9 (mod 12), 2 if h ≡ 1, 3, 5, 7 (mod 12), and 3 if h ≡ 11 (mod 12); here n_h(2) = ⌊(h²+6h+1)/4⌋. — (medium tier) Determines n♭_h(2) exactly for all odd h ≥ 5 except h ≡ 59 (mod 60), where arXiv:2610.19470 only shows n♭_h(2) < n_h(2): the gap n_h(2) - n♭_h(2) is 1 iff h ≡ 9 (mod 12), 2 for h ≡ 1, 3, 5, 7 (mod 12) and 3 for h ≡ 11 (mod 12); for h ≡ 59 (mod 60) it is at least 4, and it is at least 6 at h = 419

Research-level. postage_stamp_number_neg_two_add_four_le_of_mod_sixty_eq: For odd h ≥ 5 with h ≡ 59 (mod 60), n♭_h(2) ≤ n_h(2) - 4, where n_h(2) = ⌊(h²+6h+1)/4⌋. — (medium tier) a new question answered: For h ≡ 59 (mod 60) the gap is at least 4: at range n_h(2) - 3 the Theorem 4.1 inequalities leave only sets with gcd(x, y) > 1 or a divisibility contradiction.

Research-level. mondrian_no_perfect_three_prime_area_one_factor (arXiv:2610.19193): For three distinct primes p,q,r with p>r+1, no perfect admissible integer square partition has common tile area q^s*r*p^a, for arbitrary natural a and s. — (low tier) Excludes perfect partitions of every integer square with common tile area q^s*r*p^a for distinct primes p,q,r with p>r+1, uniformly in a and s. This extends the two-prime common-area obstruction by using the superincreasing signed-height argument

Partial result on OP-529. exists_non_circulant_sharp_cyclic_graph_order_eight (arXiv:2610.31807): There exists a simple graph G on ZMod 8 with at least one edge that admits a sharp cyclic labeling and is not a circulant graph on ZMod 8.

Partial result on OP-529. circulant_graph_admits_sharp_cyclic_labeling (arXiv:2610.31807): For every N ≥ 1 and every subset s of ZMod N, if the circulant graph SimpleGraph.circulantGraph s has at least one edge, then it admits a sharp cyclic labeling on ZMod N.

Partial result on OP-529. sharp_cyclic_graph_is_regular (arXiv:2610.31807): Every finite simple graph G with at least one edge that admits a sharp cyclic labeling on ZMod N is regular of some degree d.

Partial result on OP-518. op_518_positive_fibre_adj_of_adj (arXiv:2610.03808): For every natural number m > 0, and any functions u, v : OrderedTriple (Fin m) → ℤˣ that are rank-3 signotopes whose positive fibres are uniform rank-3 chirotopes on WithTop (Fin m), if u and v are adjacent in the higher Bruhat order bruhat_order_three_graph…

Partial result on OP-517. op_517_zero_excess_iff_exists_walk (arXiv:2610.03808): Two rank-3 signotopes on Fin m have braid-move excess 0 if and only if there exists a walk in the higher Bruhat order B(m, 2) between them whose length equals their Hamming distance.

Partial result on OP-517. op_517_zero_excess_of_eq (arXiv:2610.03808): Every rank-3 signotope on Fin m has braid-move excess 0 with respect to itself: the distance in B(m, 2) from u to u is 0, which equals the number of triples on which u differs from u.

Partial result on OP-163. hunger_transient_time_eq_zero_of_mem_basin (arXiv:2610.08949): For a natural number n with [NeZero n], a rational matrix P : Matrix (Fin n) (Fin n) ℚ, a natural number T, and a rational vector h : Fin n → ℚ, if h belongs to hunger_recurrent_basin P T, then hunger_transient_time P T h = 0.

Partial result on OP-163. hunger_transient_polynomial_bound_of_mem_basin (arXiv:2610.08949): For a natural number n with [NeZero n], a rational matrix P : Matrix (Fin n) (Fin n) ℚ, a natural number T, and a rational vector h : Fin n → ℚ, if h belongs to hunger_recurrent_basin P T, then (hunger_transient_time P T h : ℚ) ≤ (1 : ℚ) * ((n : ℚ) + (T : ℚ)…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_one_iff_fourteen: In the layer-free tight-class model for bandwidth fifteen, the first inner position belongs to a class exactly when the penultimate position does.

Thread RT-1. cycle_power_fifteen_tight_type_sequence_two_iff_three_core_b0: One branch of the window refutation behind cycle_power_fifteen_tight_type_sequence_two_iff_three_core: under the layer-free tight-cover constraints for k = 15 and 8 fixed bits of the class-type sequence near position 0 (a branch of the case split), the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_two_iff_three_core_b1: One branch of the window refutation behind cycle_power_fifteen_tight_type_sequence_two_iff_three_core: under the layer-free tight-cover constraints for k = 15 and 26 fixed bits of the class-type sequence near position 0 (a branch of the case split), the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_two_iff_three_core_a0000: One branch of the window refutation behind cycle_power_fifteen_tight_type_sequence_two_iff_three_core: under the layer-free tight-cover constraints for k = 15 and 13 fixed bits of the class-type sequence near position 0 (a branch of the case split), the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_two_iff_three_core_a0001000: One branch of the window refutation behind cycle_power_fifteen_tight_type_sequence_two_iff_three_core: under the layer-free tight-cover constraints for k = 15 and 79 fixed bits of the class-type sequence near position 0 (a branch of the case split), the…

Thread RT-1. cycle_power_fifteen_tight_type_sequence_two_iff_three_core_a0001001: One branch of the window refutation behind cycle_power_fifteen_tight_type_sequence_two_iff_three_core: under the layer-free tight-cover constraints for k = 15 and 16 fixed bits of the class-type sequence near position 0 (a branch of the case split), the…

Formalized. hamming_equilateral_triangle_local_types (arXiv:2610.32789 "Triangle-Free Sets in Exact-Distance Hamming Graphs", Proposition 2.2): Suppose $x, y, z : \mathrm{Fin}\, n \to \alpha$ satisfy $\mathrm{hammingDist}(x, y) = r$, $\mathrm{hammingDist}(x, z) = r$, and $\mathrm{hammingDist}(y, z) = r$. If $a_x, a_y, a_z$ count the coordinates where $x, y, z$ respectively are exceptional and $d$…

Formalized. unit_hamming_max_triangle_free_eq (arXiv:2610.32789 Theorem 1.3): For every fixed integer q ≥ 2 and every integer n ≥ 1, the maximum size unit_hamming_max_triangle_free n q of a triangle-free vertex subset of (Fin n → Fin q) in the unit-distance Hamming graph equals 2 * q ^ (n - 1).

Formalized. frobenius_extremal_family_frobenius_number (scout/2610-32219-a-sharp-upper-bound-for-the-frobenius-number.md:450): For any integers n ≥ 3 and q ≥ 2, the Frobenius number of the extremal family frobenius_extremal_family n q is equal to ((n - 2) * q + 1) * q - (n - 1) * q - 1.

Formalized. simple_graph_t_chromatic_ps_determines_chromatic_ps (arXiv:2610.31785, Proposition 3.9): Let G and H be finite simple graphs on finite vertex types V and W, respectively, and let n be a positive integer. If the n-th principal specialization of the t-chromatic symmetric function of G equals that of H, then the n-th principal specialization of the…

Formalized. card_two_stack_sortable_perm_ending_in_zero (arXiv:2610.32824 Theorem 1.3): For every non-negative integer n, the number of permutations of {0, 1, ..., n} that end with 0 and are 2-stack-sortable is equal to the n-th Motzkin number.

Formalized. simple_graph_t_chromatic_ps_vertex_count_eq (arXiv:2610.31785 (Principal specializations of the t-chromatic symmetric function), Proposition 3.16(1)): Let n ≥ 2 and let G be a finite simple graph with at least 2 vertices that has a proper coloring with n colors (χ_G(n) > 0). Let d and D be the minimum and maximum degrees of q appearing in the n-th principal specialization ps_n(Z_G(x; t)) (the least and…

Formalized. simple_graph_t_chromatic_ps_q_coeff_symm (arXiv:2610.31785 (Principal specializations of the t-chromatic symmetric function), Proposition 3.16(2)): Let n ≥ 2 and let G be a finite simple graph with at least 2 vertices that has a proper coloring with n colors. Let d and D be the minimum and maximum degrees of q appearing in ps_n(Z_G(x; t)). Then for every j with d ≤ j ≤ D, the coefficient of q^j in…

Formalized. simple_graph_t_chromatic_ps_coeff_t_two_connected_bipartite (arXiv:2610.31785 (Principal specializations of the t-chromatic symmetric function), Proposition 3.16(3); the bipartition is given as a…): Let n ≥ 2 and let G be a connected finite simple graph with at least 2 vertices that has a proper coloring with n colors, and let b : V → Fin 2 be a proper 2-coloring of G (so G is bipartite). Let c1 and c2 be the smaller and the larger of the sizes of the…

Corrected. not_p7_claw_free_dominating_pair_of_no_hole: The statement p7_claw_free_dominating_pair_of_no_hole is false: The statement is FALSE. A concrete counterexample is the net graph (a central triangle with a pendant vertex attached to each vertex of the triangle) on 6 vertices: - Vertex set: {0, 1, 2, 3,…

Corrected. not_rank_score_veto_core_side_count_bound: The statement rank_score_veto_core_side_count_bound is false: The conclusion has its two cardinalities swapped with respect to the paper's lemma (and with respect to the informal statement). In Lean, candidate_rank ord c = #{c' : c < c'} (verified: rank…

Open. op_518_positive_fibre_isometry (arXiv:2610.03808): For every natural number m > 0, and any functions u, v : OrderedTriple (Fin m) → ℤˣ that are rank-3 signotopes whose positive fibres are uniform rank-3 chirotopes on WithTop (Fin m), the mutation walk distance between their positive fibres in…

Open. simple_graph_cycle_power_dyadic_cover_class_model: Every optimal dyadic cycle-power equivalence cover admits a faithful periodic integer class model with the original rooted classes and real layer colours, tight class weights, constant active counts, exact dyadic coverage, all ordinary coverage, and disjoint…

Open. cyclic_matrix_fourier_antidiagonal_iff_circulant: A complex matrix M indexed by ZMod N has its two-dimensional cyclic Fourier transform supported on the anti-diagonal m + n = 0 if and only if M is a circulant matrix.

Open. sharp_cyclic_and_compl_iff_circulant (arXiv:2610.31807 Proposition 4.9 -- Additive energy and Fourier complexity of graphs): Let G be a simple graph on ZMod N, and let G^c be its complement graph. Suppose that both G and G^c have at least one edge. The standard cyclic labeling is sharp for both G and G^c if and only if the adjacency matrix of G is circulant.

Open. cycle_power_dyadic_endpoint_overlap: Conjectured uniform endpoint-overlap core: every class other than the ending class that is active at its last vertex intersects the class starting immediately afterward, in a periodic tight dyadic model with constant optimal count and full coverage.

Open. rational_hunger_game_transient_polynomial_bound (arXiv:2610.08949): There exist natural numbers C and d such that for every natural number n with [NeZero n], every rational matrix P : Matrix (Fin n) (Fin n) ℚ in Matrix.rowStochastic ℚ (Fin n) that is irreducible (is_irreducible_matrix P), every stationary distribution π :…

Verified in Lean

Daily reading

Daily reading: 6 research-level, 244 proved, 35 open

Research-level results today: 6. (medium tier) a new question answered: Evaluate every individual availability permanent as 4 on the double-3-tail staircase family, rather than merely expressing the total count as a sum of unknown permanents or proving top-row monotonicity; (medium tier) Excludes every square side q*p^e for distinct primes p,q with no bound on e and no ordering restriction on the primes; arXiv:2007.09639 Theorem 4.1 excludes prime powers and products of at most three primes but does not give this exponent-uniform two-prime family; (medium tier) Excludes perfect partitions of every integer square when the common tile area is q^s*p^t for distinct primes p,q and arbitrary exponents, and hence excludes all square sides supported on two primes. The prime-power and small prime-product exclusions of arXiv:2007.09639 Theorem 4.1 and the density theorem in HAL:hal-05729522v2 do not give this uniform common-area restriction; (low tier) Proves exact EF1 for every number of agents when m <= 2n+1, a positive guarantee beyond the general 1/2-EF1 bound and the three-agent/eight-good counterexample in arXiv:2610.23686; (low tier) refutes OP-302 (Matroid flats from hop-maximal and leap-maximal subgraphs); (low tier) Proves the conjectured value of the wild number, floor(4*ceil(n/2)/3) = 186, for every minimum proper edge colouring of K_279, a case far beyond the range 3 <= n <= 12 where the paper verified it; the upper bound comes from a seal-free triple partition obtained with a local lemma for random permutations and an exact rational certificate. Results that bear on the same open problem or paper share one Research Blog post.

Details

The Proposer read 11 new paper(s) and listed 24 open problem(s); 14 list entries were stated in Lean, as were 12 results the papers prove; the Proposer posed 49 more from what the knowledgebase already holds (200 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 244 theorem(s), 158 of them helper lemmas, which are not listed below; 35 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. latin_double_three_tail_four_extensions: Conjecture: every Latin tableau of shape (k,k-1,...,4,3,3), k at least 5, extends in exactly four ways after adding the next top row of length k+1. — (medium tier) a new question answered: Evaluate every individual availability permanent as 4 on the double-3-tail staircase family, rather than merely expressing the total count as a sum of unknown permanents or proving top-row monotonicity.

Research-level. mondrian_no_perfect_prime_mul_prime_pow (arXiv:2610.19193): For every pair of distinct primes p and q and every natural exponent e, the square of side q times p^e has no perfect admissible integer rectangle partition. — (medium tier) Excludes every square side q*p^e for distinct primes p,q with no bound on e and no ordering restriction on the primes; arXiv:2007.09639 Theorem 4.1 excludes prime powers and products of at most three primes but does not give this exponent-uniform two-prime family

Research-level. mondrian_no_perfect_two_prime_area (arXiv:2610.19193): For any distinct primes p and q and natural exponents s and t, no perfect admissible partition of any integer square can have common tile area q^s*p^t. — (medium tier) Excludes perfect partitions of every integer square when the common tile area is q^s*p^t for distinct primes p,q and arbitrary exponents, and hence excludes all square sides supported on two primes. The prime-power and small prime-product exclusions of arXiv:2007.09639 Theorem 4.1 and the density theorem in HAL:hal-05729522v2 do not give this uniform common-area restriction

Research-level. round_robin_pne_ef1_of_card_le_two_mul_add_one: With n agents and at most 2n + 1 goods, every pure Nash equilibrium of Round-Robin (nonnegative additive valuations) is EF1. For three agents this is tight: arXiv:2610.23686 Proposition 1 has a non-EF1 equilibrium with eight goods. Whether 2n + 2 goods can break EF1 for every n >= 4 is open. — (low tier) Proves exact EF1 for every number of agents when m <= 2n+1, a positive guarantee beyond the general 1/2-EF1 bound and the three-agent/eight-good counterexample in arXiv:2610.23686

Research-level. simple_graph_hop_or_leap_maximal_not_matroid_flats (arXiv:2610.18463): Neither the collection of hop-closed edge subsets nor the collection of leap-closed edge subsets forms the set of flats of a matroid for every finite simple graph G. — (low tier) refutes OP-302 (Matroid flats from hop-maximal and leap-maximal subgraphs)

Research-level. wild_number_complete_graph_279_min_proper_coloring_eq (arXiv:2610.22142): New case of OP-306: every minimum proper edge colouring of K_279 has wild number 186 = floor(4*ceil(279/2)/3). The lower bound is the paper's; the matching upper bound comes from a seal-free partition of the vertices into 93 triples, found with a local lemma… — (low tier) Proves the conjectured value of the wild number, floor(4*ceil(n/2)/3) = 186, for every minimum proper edge colouring of K_279, a case far beyond the range 3 <= n <= 12 where the paper verified it; the upper bound comes from a seal-free triple partition obtained with a local lemma for random permutations and an exact rational certificate

Partial result on OP-86. maxcut_optimal_not_strong_degree_66 (arXiv:2610.04948): There is a finite simple graph of maximum degree at most 66 with an optimal coloring of the Max-3-Cut game that is not a strong equilibrium; hence Δ*(3) ≤ 65. (The graph G_3 of arXiv:2610.04948, Sec. 2, has 91 vertices and maximum degree 66, attained by the…

Partial result on OP-5. gf8_oracle_min_t_count_ge_20 (arXiv:2610.01024): The slice-rank lower bound: the minimal T-count of the GF(8) multiplication oracle is at least 20.

Partial result on OP-5. gf8_oracle_min_t_count_le_29 (arXiv:2610.01024): The minimal T-count of the GF(8) multiplication oracle is at most 29, certified by an explicit 29-parity representation.

Partial result on OP-314. op314_poset_tower_order_iso_converse (arXiv:2610.20802): For any natural number k and any types α and β equipped with partial orders [PartialOrder α] and [PartialOrder β], if there exists an order-isomorphism α ≃o β, then there exists an order-isomorphism poset_tower α k ≃o poset_tower β k between their height-k…

Partial result on OP-280. mondrian_no_perfect_prime_power (arXiv:2610.19193): No square whose integer side length is a prime power has a perfect admissible partition.

Partial result on OP-305. two_facility_universal_expected_max_cost_nonneg (arXiv:2610.18006): For any universally strategyproof two-facility mechanism M with finite nonempty index type K on a finite nonempty agent set ι in a metric space X and any position profile x : ι → X, two_facility_universal_expected_max_cost M x is nonnegative (0 ≤…

Partial result on OP-280. mondrian_large_prime_descent (arXiv:2610.19193): If c is positive and p is a prime greater than c, any perfect partition of the c times p to the e square yields a perfect partition of the c-square.

Partial result on OP-280. mondrian_perfect_partition_large_prime_iff (arXiv:2610.19193): For every positive integer c, prime p greater than c, and natural exponent e, the square of side c times p to the e has a perfect admissible integer-rectangle partition if and only if the c-square has one.

Thread RT-3. round_robin_pne_envy_zero_ef1_of_two_mul_add_one: With n agents and 2n + 1 goods, in every pure Nash equilibrium of Round-Robin (nonnegative additive valuations) every agent i ≠ 0 is EF1 towards agent 0, the only agent with three goods. With round_robin_pne_ef1_pair_of_count_le_two and…

Thread RT-3. picking_sequence_outcome_pairwise_disjoint: Distinct agents never receive the same good under a picking sequence: the bundles are pairwise disjoint.

Thread RT-3. round_robin_two_mul_add_one_deviation_certificate: Round-Robin with n agents and 2n+1 goods, any reports, agent i ≠ 0: for some good g of agent 0's bundle B_0 (its first pick a1), either some deviation r of agent i alone gets i all of B_0 ∖ {g}, or two deviations r1, r2 of agent i give bundles D1, D2 that…

Thread RT-3. round_robin_two_mul_add_one_outcome_structure: Round-Robin with n agents and 2n+1 goods, any reports: there are goods a2, a3 (agent 0's picks at turns n and 2n) and b1, b2 (agent i's two picks) such that agent 0's bundle is {its first-ranked good, a2, a3}, these three are distinct, agent i's bundle is…

Thread RT-3. round_robin_two_mul_add_one_cover_pairs: Round-Robin with n agents and 2n+1 goods, agent i ≠ 0: let a1 be agent 0's top-ranked good, a2 and a3 the goods it picks at turns n and 2n, and let agent i's bundle be {b1, b2}. Then either some report of agent i alone gives it a bundle containing both a2…

Thread RT-3. picking_turn_alloc_subset: One turn never un-allocates a good: the set of goods already allocated before a picking turn is contained in the set of allocated goods after it.

Formalized. perm_offset_avoiding_count_eq_sum_pie (arXiv:2610.23958 Equation (2)): For every natural number n and finite subset S ⊆ ℤ, the integer count (perm_offset_avoiding_count n S : ℤ) of S-avoiding permutations equals the sum over k from 0 to n of (-1 : ℤ) ^ k * ((n - k).factorial : ℤ) * (perm_offset_rook_number n S k : ℤ).

Formalized. positive_binary_identity_bound_of_perm_cover (arXiv:2610.11098 Lemma 3.6 -- An Improved Lower Bound for Separating Words by Permutation Automata): Let k ≥ 1 be an integer, and let C be a finite set of natural numbers covering the element orders of Equiv.Perm (Fin k) (satisfying is_perm_cover k C). Then there exist distinct binary words u, v : List (Fin 2) with equal counts of each letter (u.count 0 =…

Formalized. simple_graph_is_far_colorable_iff (combinatorics/2610-31489-near-optimal-bounds-for-testing-k-colorability.md:152): A finite simple graph G on a type V with decidable equality and decidable adjacency is ε-far from being k-colorable if and only if every vertex coloring c : V → Fin k satisfies ε * (Fintype.card V : ℝ) ^ 2 ≤ ((simple_graph_monochromatic_edges G c).card : ℝ).

Formalized. two_four_code_sync_single_coord (arXiv:2610.30613, Lemma 2.1): Let $\alpha$ be an alphabet with decidable equality and a distinguished element $0$. Let $\mathcal{C} \subseteq (\mathrm{Fin}\, n \to \alpha)$ be a code containing $0$ such that all distinct codewords have Hamming distance in $\{2, 4\}$. If $x \in…

Formalized. matrix_psd_rank_le_rank_of_hadamard_eq (arXiv:2610.04498 Lemma 2.1 -- Rank-sensitive vertex bounds for semidefinite lifts): For finite types m and n, if M is an entrywise nonnegative real matrix (∀ i j, 0 ≤ M i j) and R is a real matrix such that M = Matrix.hadamard R R, then the positive semidefinite rank of M is at most the ordinary rank of R (matrix_psd_rank M ≤ Matrix.rank R).

Formalized. nonclairvoyant_scheduling_offline_opt_superadditive (arXiv:2610.29037 Lemma 4.2 proof step (1)): For any natural number n and job processing vectors a, z : Fin n → ℝ, the offline optimal total completion time satisfies superadditivity: nonclairvoyant_scheduling_offline_opt a + nonclairvoyant_scheduling_offline_opt z ≤…

Formalized. nonclairvoyant_scheduling_round_robin_hybrid_competitive_ratio (arXiv:2610.29037 Lemma 4.2): Let n be a natural number and let x, a, z : Fin n → ℝ with x = a + z, where a and z have non-negative coordinates (∀ i, 0 ≤ a i and ∀ i, 0 ≤ z i). Suppose an algorithm accumulates non-negative cost C_T ≥ 0 by time T and incurs total completion cost alg_cost…

Formalized. is_k_connected_set_empty (arXiv:2610.29979, Section 2): In every matroid $M$, the empty set is $k$-connected for all $k$.

Corrected. not_is_pk_free_claw_free_hole_dominating: The statement is_pk_free_claw_free_hole_dominating is false: The statement is false as written: it allows k = 4 and m = 3 (since k - 1 = 3 ≤ 3). Counterexample: start from K4 on vertices {0,1,2,3} and attach a pendant vertex 4 adjacent only to 3 (i.e. the…

Corrected. not_single_box_valid_is_partially_adaptive: The statement single_box_valid_is_partially_adaptive is false: The statement is false. IsValidPolicy only constrains the continuation cont v for values v ∈ (inst.dist b).support. In contrast, IsPartiallyAdaptive requires aux order (cont v) for…

Open. op314_poset_tower_two_cancellation (arXiv:2610.20802): For any types α and β equipped with [PartialOrder α], [Fintype α], [Nonempty α], [PartialOrder β], [Fintype β], and [Nonempty β], if there exists an order-isomorphism poset_tower α 2 ≃o poset_tower β 2 between their height-2 exponentiation towers, then there…

Open. op314_cancellation_poset_tower_three (arXiv:2610.20802): For any types α and β equipped with [PartialOrder α], [Fintype α], [Nonempty α], [PartialOrder β], [Fintype β], and [Nonempty β], if there exists an order-isomorphism poset_tower α 3 ≃o poset_tower β 3 between their height-3 right-associated exponentiation…

Open. op314_cancellation_poset_tower (arXiv:2610.20802): For each fixed integer k ≥ 3, and any types α and β equipped with [PartialOrder α], [Fintype α], [Nonempty α], [PartialOrder β], [Fintype β], and [Nonempty β], if there exists an order-isomorphism poset_tower α k ≃o poset_tower β k between their height-k…

Open. two_facility_universal_max_cost_ratio_strictly_below_19_5 (arXiv:2610.18006): In the two-facility location problem with candidate locations on general metric spaces, there exists an approximation ratio α : ℝ with α < 19 / 5 such that for every metric space X : Type*, every candidate set C : Finset X with 2 ≤ C.card, every agent type ι…

Open. op_300_leap_group_prism_and_hypercube (arXiv:2610.18463): Open Problem OP-300: For each odd integer n ≥ 3, the leap group of the prism graph P₁ □ C_n is the full symmetric group on its 2n vertices; for each even integer n ≥ 3, the leap group of P₁ □ C_n has order 2 * (n!)²; and for each integer n ≥ 1, the leap…

Open. two_token_kmn_domination_number (arXiv:2610.28936 Theorem 1.1): Let 2 <= m <= n. The domination number of the 2-token graph F_2(K_{m,n}) is m/2 + n - 1 if m is even; (m-1)/2 + n - 1 if m = 3, or if m is odd, m >= 5, and n = m + 1; and (m-1)/2 + n if m is odd, m >= 5, and n >= m + 2 (or n = m).

Verified in Lean

Weekly report

Week of 19–25 Oct 2026: Minimal turn-taking failures, cycle power covering bounds, and negative stamp bases

Key questions were settled across game theory, number theory, and graph algorithms. Strategic turn-taking among three participants was proved fair up to one item for seven items, showing that eight items is the minimal failure point. A problem on covering cycle powers by equivalence graphs was bounded for power seven. New gap and classification results settled the behavior of three-stamp bases with negative denominations, and line cubic syndrome representations were constructed uniformly for all dimensions.

Details

Proved. Round-Robin pure Nash equilibria are EF1 for three agents and seven goods

Proved. Line cubic syndrome has a parity representation of 2n plus 1 supports

New result. Unit gap for negative-allowed three-stamp bases occurs exactly when h is 9 modulo 12

Partial. Seventh cycle power requires at least five equivalence subgraphs when eight does not divide n

Lean accepted 1074 new theorems this week, drawn from 90 papers and 23 topics. Of these, 33 are new here, 0 correct a published claim, 154 formalize a published result and 887 are helper lemmas. 6 are research-level results -- they lead the report. 768 statements posed this week are not yet proved. The reading covered 86 new papers. This report covers 9 results in 4 fields.

Daily reading

Daily reading: 1 research-level, 205 proved, 50 open

Research-level results today: 1. A uniform recursive construction proves that the line cubic has a parity representation of exactly 2n+1 supports for every n≥5, resolving the general case left open beyond the witnesses through n=12 in arXiv:2610.01024. Results that bear on the same open problem or paper share one Research Blog post.

Details

The Proposer read 0 new paper(s) and listed 0 open problem(s); 18 list entries were stated in Lean, as were 11 results the papers prove; the Proposer posed 37 more from what the knowledgebase already holds (227 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 205 theorem(s), 155 of them helper lemmas, which are not listed below; 50 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. open_problem_line_cubic_t_count (arXiv:2610.01024): For every integer n ≥ 5, there exists a parity representation of the line cubic syndrome of cardinality 2n + 1. — A uniform recursive construction proves that the line cubic has a parity representation of exactly 2n+1 supports for every n≥5, resolving the general case left open beyond the witnesses through n=12 in arXiv:2610.01024

Partial result on OP-446. tetrahedron_boundary_is_partitionable (arXiv:2610.27083): The boundary of a tetrahedron, formed by four triangular facets on four vertices, is partitionable into boolean intervals.

Partial result on OP-437. is_straight_line_cut_consistent_of_parallel_cuts (arXiv:2610.27956): Let P be a finite set of points in ℝ × ℝ, φ a non-zero linear functional from ℝ × ℝ to ℝ, t₁ and t₂ real numbers with t₁ < t₂, and α an allocation function from ℝ × ℝ to Fin 3. If for every point p in P, α p = 0 implies φ p ≤ t₁, α p = 1 implies t₁ ≤ φ p and…

Partial result on OP-437. planar_necklace_two_cuts_six_points (arXiv:2610.27956): For any finite set P of 6 points in the plane ℝ × ℝ colored with two colors c : (ℝ × ℝ) → Fin 2 such that exactly 3 points are colored 0 and 3 points are colored 1, there exists a finite set L of at most 2 straight lines and an allocation function α : (ℝ ×…

Partial result on OP-437. planar_necklace_three_agents_cuts_ge_two (arXiv:2610.27956): Let P be a finite set of points in ℝ × ℝ colored with two colors c : (ℝ × ℝ) → Fin 2, L a finite set of straight lines, and α an allocation function from ℝ × ℝ to Fin 3. If each of the three agents receives exactly 1 point of color 0 and 1 point of color 1…

Partial result on OP-306. wild_number_complete_graph_279_min_proper_coloring_eq (arXiv:2610.22142): New case of OP-306: every minimum proper edge colouring of K_279 has wild number 186 = floor(4*ceil(279/2)/3). The lower bound is the paper's; the matching upper bound comes from a seal-free partition of the vertices into 93 triples, found with a local lemma…

Partial result on OP-1. matrix3_trace_split_reconstruction (arXiv:2610.01639): The cyclic trace coefficient identities reconstruct every 27-coordinate vector through the first-factor column ranges. This is the explicit right-inverse identity behind the split flattening.

Partial result on OP-1. matrix3_trace_rank_sum_ge_27 (arXiv:2610.01639): For a finite family of 3x3 matrices over a field satisfying the cyclic trace identities of matrix multiplication, the sum of the matrix ranks of the first factors is at least 27.

Partial result on OP-396. young_lattice_one_fractional_scd (arXiv:2610.24148): For any positive integer n ≥ 1, the Young lattice L(1, n) admits a fractional symmetric chain decomposition.

Thread RT-1. cycle_power_fifteen_endpoint_layer_period_sixteen: For sixteen-position cyclic class types alternating between endpoint-only types {0,15} and types with equal consecutive membership bits, assume every vertex is active in exactly five classes, the active classes have distinct labels in Fin 5, and every class…

Formalized. canonical_permutation_avoidance_of_local_lemma (Using Lovász Local Lemma in the Space of Random Injections, Electronic Journal of Combinatorics 14 (2007), R63, DOI 10.37236/981, Theorem…): For finitely many injective partial assignments on n points, if weights in [0,1) satisfy the asymmetric local-lemma inequalities for the graph connecting assignments with overlapping domains or ranges, some permutation avoids every assignment.

Formalized. simple_graph_rich_witness_graph_coloring_iff (arXiv:2610.16446, Section 2, Lemma 3 and its direct converse, in the simple-graph specialization.): Colorability of the prescribed-witness graph is equivalent to a proper edge-coloring of G in which every prescribed edge of R is rich.

Formalized. pell_number_closed_form (arXiv:2610.08525 Appendix A): For every natural number m, twice the m-th Pell number equals ((1 + Real.sqrt 2) ^ m - (1 - Real.sqrt 2) ^ m) / Real.sqrt 2.

Formalized. brgc_inverse_flip_preserves_cut_max (Gray-ordered Binary Necklaces (2007), Theorem 3.6, in the equivalent convention of maximal forward Gray rank; the rotation form adds…): If an n-bit word has strictly greater inverse Gray rank than every nontrivial cyclic cut, toggling its last bit has at least as great an inverse Gray rank as every nontrivial cyclic cut of the toggled word.

Formalized. brgc_inverse_flip_preserves_rotation_max (Gray-ordered Binary Necklaces (2007), Theorem 3.6, in the equivalent convention of maximal forward Gray rank; the rotation form adds…): Toggling the last bit of an aperiodic nonconstant word that maximizes its inverse Gray rank over all rotations again maximizes that rank, for word length at least two.

Formalized. matrix_eigenvalues₀_le_of_sub_posSemidef (scout/2610-10618-the-hellinger-conjecture-for-antipodal-boolean-functions.md:402): If $A$ and $C$ are real symmetric matrices such that $C - A$ is positive semidefinite, then each ordered eigenvalue of $A$ is at most the corresponding ordered eigenvalue of $C$.

Formalized. matrix_nuclear_norm_eigenvalue_inequality (arXiv:2610.10618 Lemma 2.2 -- The Hellinger Conjecture for Antipodal Boolean Functions): Let $n$ be a finite type with decidable equality. Let $A$ and $B$ be real symmetric matrices indexed by $n \times n$, with eigenvalues $a_j$ and $b_j$ in nonincreasing order. Then the sum over $j \in \mathrm{Fin}(|n|)$ of $|a_j - b_j|$ is at most the nuclear…

Formalized. matrix_nuclear_norm_of_isHermitian (scout/2610-10618-the-hellinger-conjecture-for-antipodal-boolean-functions.md:315): For a real symmetric matrix $M$, its nuclear norm equals the sum of the absolute values of its eigenvalues.

Open. cohen_macaulay_two_complex_partitionable (arXiv:2610.27083): Every finite pure two-dimensional Cohen-Macaulay simplicial complex over a field k is partitionable into boolean intervals indexed by its facets.

Open. arrows_realization_up_to_63 (arXiv:2610.15145 Theorem 7.3): Every natural number g with g ≤ 63 is realizable as the Grundy value of the Game of Arrows on some finite tree. Furthermore, every finite tree with Grundy value at least 2 has at least 17 vertices, every finite tree with Grundy value at least 4 has at least…

Open. open_problem_young_lattice_fractional_scd (arXiv:2610.24148): For all positive integers m, n ≥ 1, the Young lattice L(m, n) of partitions bounded in an m × n box admits a fractional symmetric chain decomposition with respect to its natural centered rank.

Open. arrows_twin_branch_reduction (arXiv:2610.15145 Theorem 5.6): Let c be a vertex of a tree T, and let B₁, B₂ be distinct components of T - c that are isomorphic via an isomorphism φ from T[B₁ + c] to T[B₂ + c] fixing c. Let T' be obtained from T by deleting B₁ ∪ B₂ and attaching a new pendant edge at c. Then the Grundy…

Open. op_380_thirteen_edge_connected_has_modulo_seven_orientation (arXiv:2610.07034): Every 13-edge-connected finite simple graph has a modulo-7 orientation.

Open. kneser_hom_cycle_reduction_to_cycle_with_chords (arXiv:2610.17884 Lemma 4.1): Let k ≥ 2, m ≥ 6, and 3 ≤ a ≤ m - 3. For every graph G, there exists a graph homomorphism from the G-stable Kneser graph of the disjoint sum of G and the (m-1)-cycle to the G-stable Kneser graph of the disjoint sum of G and the m-cycle with two chords {1, a}…

Verified in Lean

Daily reading

Daily reading: 2 research-level, 203 proved, 32 open

Research-level results today: 2. a new question answered: decides when the strict inequality n♭_h(2) < n_h(2) loses only one: exactly when h ≡ 9 (mod 12) (h ≥ 5); a new question answered: At h = 419 the gap is at least 6: the Theorem 4.1 inequalities at range N - 5 leave only six sets a + {0,x,y}, and each has gcd(x,y) > 1 (y = 210 = 2·3·5·7 with x ∈ {3,5,7,9}, or y = 212 with x ∈ {138,140}).. Results that bear on the same open problem or paper share one Research Blog post.

Details

The Proposer read 0 new paper(s) and listed 1 open problem(s); 14 list entries were stated in Lean, as were 8 results the papers prove; the Proposer posed 49 more from what the knowledgebase already holds (193 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 203 theorem(s), 145 of them helper lemmas, which are not listed below; 32 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. postage_stamp_number_neg_two_add_one_eq_iff: For odd h ≥ 5, the negative-allowed three-stamp number misses n_h(2) = ⌊(h²+6h+1)/4⌋ by exactly one, n♭_h(2) = n_h(2) - 1, if and only if h ≡ 9 (mod 12). — a new question answered: decides when the strict inequality n♭_h(2) < n_h(2) loses only one: exactly when h ≡ 9 (mod 12) (h ≥ 5)

Research-level. postage_stamp_number_neg_four_one_nine_two_add_six_le: For h = 419 (h ≡ 59 mod 60), the negative-allowed three-stamp number satisfies n♭_h(2) ≤ n_h(2) - 6, where n_h(2) = ⌊(h²+6h+1)/4⌋ = 44519. This refutes the conjectured value n♭_h(2) = n_h(2) - 4 for h ≡ 59 (mod 60) (postage_stamp_number_neg_two_of_odd). — a new question answered: At h = 419 the gap is at least 6: the Theorem 4.1 inequalities at range N - 5 leave only six sets a + {0,x,y}, and each has gcd(x,y) > 1 (y = 210 = 2·3·5·7 with x ∈ {3,5,7,9}, or y = 212 with x ∈ {138,140}).

Refutes OP-182. op182_deferred_to_deadline_beyond_half_against_stopping (arXiv:2610.13775): Refutation of Open Problem OP-182 against the deferred-until-stopping benchmark: there does not exist any constant c > 1/2 such that for every Pandora's problem instance with deadlines inst : PandoraDeadlineInstance, there exists a valid adaptive policy…

Partial result on OP-502. prefix_reversal_six_five_three_not_top (arXiv:2610.30892): In the symmetric group Equiv.Perm (Fin 6), the subgroup generated by the three prefix reversals of lengths 3, 5, and 6 does not equal the full symmetric group.

Refutes OP-502. prefix_reversal_necessary_conditions_not_sufficient (arXiv:2610.30892): It is not the case that for all natural numbers 2 ≤ k < m < n satisfying 2 * m > n, at least one of k, m, n even, at least one of k, m, n congruent to 2 or 3 mod 4, gcd(gcd(k, m), n) = 1, m + k ≥ n + 1, and gcd(n - m, n - k) = 1, the three prefix reversals…

Partial result on OP-502. prefix_reversal_gcd_diff_necessary (arXiv:2610.30892): If three prefix reversals of lengths k, m, n with 2 ≤ k < m < n generate the symmetric group Equiv.Perm (Fin n), then gcd(n - m, n - k) = 1.

Answers OP-494. op_494_additive_chromatic_bipartite_maxDegree (arXiv:2610.30174): Let G be a finite connected bipartite simple graph with maximum degree at least 2. Then there exists a vertex labeling f : V → ℕ such that 1 ≤ f(v) ≤ G.maxDegree for all v, and adjacent vertices have different neighborhood sums ∑_{w ∈ N(u)} f(w) ≠ ∑_{w ∈…

Partial result on OP-494. additive_coloring_bipartite_maxDegree_add_one (arXiv:2610.30174): Every finite connected bipartite graph with maximum degree at least 2 admits an additive coloring f with 1 ≤ f(v) ≤ Delta(G) + 1 for all vertices v.

Partial result on OP-487. permutahedron_five_not_psd_factorization_six_of_rank_ge_seven (arXiv:2610.04498): If the positive semidefinite rank of the slack matrix of the standard five-element permutahedron is at least 7, then the slack matrix has no positive semidefinite factorization of size 6.

Partial result on OP-487. permutahedron_five_xc_psd_ge_six (arXiv:2610.04498): The positive semidefinite rank of the slack matrix of the standard five-element permutahedron is at least 6.

Thread RT-1. cycle_power_fifteen_class_weight_certificate: For a subset of the sixteen positions of an arc, the weighted number of pairs at distances 8, 12, 14, 15, with weights 1, 1, 1, 2, is at most its cardinality.

Thread RT-1. cycle_power_fifteen_tight_class_ancestors: Equality in the sixteen-position weighted-pair bound forces equal endpoint bits and closure under the fourteen parent edges of its rooted forest.

Thread RT-1. cycle_power_fifteen_tight_class_spanning: Every nonempty subset attaining equality in the sixteen-position weighted-pair bound contains both positions 0 and 15.

Thread RT-1. cycle_power_fifteen_type_layer_periods_dvd: For sixteen-position classes on a cyclic index set, type period 8 and layer-label period 16 force 16 to divide the cycle length, provided distinct overlapping classes have different labels and some class contains a distance-8 pair. Labels may lie in any type.

Thread RT-1. simple_graph_cycle_power_fifteen_adj_iff: For n at least 46, adjacency in the fifteenth power of the cycle is displacement by a nonzero integer of absolute value at most 15.

Formalized. poset_nonempty_order_iso_of_card_order_hom_eq_of_connected (arXiv:2610.20802 (Two short proofs of the A^A theorem), Lemma 2.4.): Let X and Y be finite posets (possibly empty). If for every finite nonempty connected poset Z the number of order-preserving maps Z → X equals the number of order-preserving maps Z → Y, then X and Y are order-isomorphic. (A poset is connected when its…

Formalized. postage_stamp_number_two_eq (arXiv:2610.19470 (Positivity and negativity for h-bases for n), Theorem 1.1 (classical; cited there).): The classical postage stamp number for three stamps: n_h(2) = ⌊(h² + 6h + 1)/4⌋ for every h ≥ 1.

Formalized. postage_stamp_number_neg_two_eq_of_mod_four_eq_two (arXiv:2610.19470 (Positivity and negativity for h-bases for n), Theorem 3.3 (with the value of n_h(2) from Theorem 1.1).): If h ≥ 2 and h ≡ 2 (mod 4), then n♭_h(2) = n_h(2) = (h² + 6h)/4.

Formalized. mutual_f_information_subadditive_full_support_of_weight_bound (arXiv:2610.21231 (On the subadditivity of mutual f-information), Theorem 1. Differs: the conclusion is for full-support distributions (the…): Theorem 1 of arXiv:2610.21231, on full-support distributions. Let f with f(1) = 0 be convex and four times continuously differentiable on (0, ∞), and g(x) = x² f''(x). If g''(x) < 0 for all x > 0 and there are constants η > 0 and ℓ with g(x) + 2(g'(x) -…

Formalized. subadditive_mutual_f_information_family_full_support (arXiv:2610.21231 (On the subadditivity of mutual f-information), Corollary 1. Differs: subadditivity is stated for full-support…): Corollary 1 of arXiv:2610.21231, on full-support distributions: let s be twice continuously differentiable, positive and concave on (0, ∞) with s'' < 0, let K_s(x) = s(x) + 2(s'(x))²/(-s''(x)) be concave on (0, ∞), and let α ≥ K_s(1). Then f(x) = α x log x +…

Formalized. hash_potential_of_length_eq (arXiv:2610.31363 Lemma 4.2): For any sequence A of length t, the potential hash_potential n r S t A equals the number of slots in S occupied by A, hash_occupy_count n r S A.

Formalized. digraph_semisaturation_bounded_of_isolated_edge (arXiv:2610.18600 Lemma 2.1): Let G : Digraph V be a finite oriented digraph on vertex type V with DecidableEq V and DecidableRel G.Adj such that digraph_is_oriented G and digraph_has_isolated_edge G hold. Then the semisaturation function is bounded: there exist C : ℕ and n₀ : ℕ such…

Formalized. permutahedron_slack_matrix_nonneg (arXiv:2610.04498 Sec. 4.2): For every natural number n, nonempty proper subset S of Fin n, and permutation σ of Fin n, the entry of the permutahedron slack matrix at (S, σ) is nonnegative.

Proved. postage_stamp_number_neg_two_of_odd_of_mod_sixty_ne: Exact value of the negative-allowed three-stamp number for odd h ≥ 5 with h ≢ 59 (mod 60): n_h(2) - n♭_h(2) = 1 if h ≡ 9 (mod 12), 2 if h ≡ 1, 3, 5, 7 (mod 12), and 3 if h ≡ 11 (mod 12); here n_h(2) = ⌊(h²+6h+1)/4⌋.

Proved. postage_stamp_number_neg_two_add_four_le_of_mod_sixty_eq: For odd h ≥ 5 with h ≡ 59 (mod 60), n♭_h(2) ≤ n_h(2) - 4, where n_h(2) = ⌊(h²+6h+1)/4⌋.

Corrected. not_prefix_reversal_sum_bound_necessary: The statement prefix_reversal_sum_bound_necessary is false: The statement claims that m + k ≥ n + 1 is a necessary condition for prefix reversals of lengths k, m, n to generate S_n. However, for (k, m, n) = (2, 3, 5), the prefix reversals of lengths 2, 3,…

Corrected. not_fair_binomial_anticoncentration_bound: The statement fair_binomial_anticoncentration_bound is false: The statement is false for small α. Take n = 2, α = 1/10, and X the identity on the binomial space Bin(2, 1/2). The interval |X - 1| ≤ (1/10)√2 contains the integer 1, whose binomial probability…

Open. picking_prefix_state_alloc_mono: The set of allocated goods is monotone in the number of turns: if s <= t then the goods allocated after s turns are allocated after t turns.

Open. hash_step_delta_sum_fully_occupied_le (arXiv:2610.31363 Lemma 4.4): For any slot subset S and index i : Fin t, the sum of potential changes hash_potential_delta n r S t ((List.ofFn H).take i.val) (H i) over all sequences H in hash_fully_occupied_seqs n r S t is at most (hash_count_fully_occupied n r S t : ℝ) - (n : ℝ) *…

Open. permutahedron_five_xc_psd_ge_seven (arXiv:2610.04498): The positive semidefinite rank of the slack matrix of the standard five-element permutahedron is at least 7.

Open. polygon_psd_rank_four_attains_eight (arXiv:2610.04498): There exists an 8 × 8 real matrix S : Matrix (Fin 8) (Fin 8) ℝ such that is_convex_polygon_slack_matrix 8 S and matrix_has_psd_factorization S 4 hold.

Open. linear_three_graph_le_five_edges_is_iterated_tripartite (arXiv:2610.14829 Proposition 2.8): Every linear 3-uniform hypergraph F with at most 5 hyperedges is iterated tripartite.

Open. cycle_rook_poly_sq_sub (arXiv:2610.23958 Lemma 6.1): For every positive integer n, the square of the rook polynomial of an n-cycle satisfies Q_n(x)^2 = Q_{2n}(x) + 2(-x)^n in ℚ[x].

Verified in Lean

Daily reading

Daily reading: 1 research-level, 148 proved, 23 open

Research-level results today: 1. Settles the lower bound in the open case k = 7 (k+1 = 8) of OP-205. For n >= 22 with 8 not dividing n, the edges of C_n^7 cannot be covered by 4 equivalence subgraphs, so eq(C_n^7) >= 5 = 2 + log2(8). With the paper's restriction of eq to {4, 5} for large n, this means the larger value holds for every even n >= 441 not divisible by 8. Only k = 3 was settled before. Results that bear on the same open problem or paper share one Research Blog post.

Details

The Proposer read 12 new paper(s) and listed 23 open problem(s); 0 list entries were stated in Lean (Claude Code 0, the agents 0), as were 26 results the papers prove (Claude Code 4, the agents 22); the Proposer posed 70 more from what the knowledgebase already holds (122 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 148 theorem(s), 110 of them helper lemmas, which are not listed below; 23 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Research-level. simple_graph_not_has_equivalence_cover_cycle_power_seven_of_not_dvd: For n ≥ 22 not divisible by 8, C_n^7 cannot be covered by 4 equivalence subgraphs; this is the lower bound eq(C_n^7) ≥ 5 = m + 2 of the open case k = 7 (m = 3) of OP-205. — Settles the lower bound in the open case k = 7 (k+1 = 8) of OP-205. For n >= 22 with 8 not dividing n, the edges of C_n^7 cannot be covered by 4 equivalence subgraphs, so eq(C_n^7) >= 5 = 2 + log2(8). With the paper's restriction of eq to {4, 5} for large n, this means the larger value holds for every even n >= 441 not divisible by 8. Only k = 3 was settled before

Proved. round_robin_three_agents_seven_goods_pne_ef1: With three agents and seven goods, every pure Nash equilibrium of Round-Robin (nonnegative additive valuations) is EF1; so the eight-good counterexample of arXiv:2610.23686 Proposition 1 is the smallest possible.

Partial result on OP-216. odd_antihole_colorable_of_common_palette (arXiv:2610.14241): Theorem 8 of arXiv:2610.14241 for the odd antihole: for ℓ ≥ 3, if all lists of an (ℓ-1)-assignment of complement(C_{2ℓ-1}) are subsets of a common ℓ-set P and are not all identical, then the antihole is L-colorable.

Thread RT-1. simple_graph_cycle_power_seven_cover_four_tight_dvd: A tight cover of C_n^7 (n ≥ 22) by 4 equivalence subgraphs (every distance-7, 6, 4 pair covered exactly once, every class spanned by a distance-7 pair) exists only when 8 divides n.

Thread RT-1. cycle_power_seven_tight_layer_colouring_dvd: Colouring core of the k = 7 lower bound. Positions of Z/n alternate between N-positions (where N holds) and P-positions, a is 4-periodic on the N-positions, and at every N-position p, with X = σ(p-2) if a(p-2) and σ(p-4) otherwise, the values σ p, σ(p-6), X…

Thread RT-1. cycle_power_seven_tight_type_sequence_types: Layer-free type classification for tight 4-covers of C_n^7. Let x p j say that the class starting at p (the class of the distance-7 pair (p, p+7)) contains p + j. Assume the counting system of cycle_power_seven_tight_type_sequence_one_iff_six together with…

Formalized. simple_graph_equivalence_covering_number_cycle_power_ge (arXiv:2610.15635 (Covering powers of cycles with equivalence subgraphs), Lemma 2.2 (the lower bound of Theorem 1.2).): For k ≥ 1 and n ≥ 2k + 2, eq(C_n^k) ≥ 1 + ⌈log₂(k + 1)⌉.

Formalized. simple_graph_equivalence_covering_number_cycle_power_dvd (arXiv:2610.15635 (Covering powers of cycles with equivalence subgraphs), Theorem 2.6.): For k ≥ 1 and q ≥ 2, the k-th power of the cycle on q(k + 1) vertices has equivalence covering number exactly 1 + ⌈log₂(k + 1)⌉.

Formalized. mixed_cage_order_one_le (arXiv:2610.14110 (New mixed cages, bounds and computational results), Theorem 4.2 (with Definition 4.1).): For z ≥ 1 and g ≥ 3, n(1, z, g) ≤ N(z, g), where N(z, g) = z(g − 1) + 2 for even z and (z + 1)(g − 1) for odd z: the circulant M(z, g) on 2m = N(z, g) vertices with arcs of jumps 1..⌈z/2⌉ and m+1..m+⌊z/2⌋ and the antipodal edges is a (1, z, g)-mixed graph.

Formalized. matrix_support_psd_rank_bound_le_rank (scout/2610-04498-rank-sensitive-vertex-bounds-for-semidefinite-lifts.md:2048): For finite types m and n, let b : Set (m × n) → ℝ be a function on support patterns such that b (matrix_support M) ≤ (matrix_psd_rank M : ℝ) for every nonnegative real matrix M : Matrix m n ℝ. Then b (matrix_support M) ≤ (Matrix.rank M : ℝ) for every…

Formalized. latin_partition_is_wide (arXiv:2610.29291 Section 1): Every Latin Young diagram is wide: if a Young diagram $\mu$ admits a Latin tableau of shape $\mu$ and type $\mu'$ (is_latin_partition μ), then $\mu$ is wide (is_wide_partition μ).

Formalized. standardized_laplacian_spectra_subset_of_dvd (scout/2610-14100-digraph-laplacian-spectra-the-2005-polygon-conjecture-fails.md:1044): Let m, n ≥ 2 be natural numbers such that m divides n. Then the set of eigenvalues of standardized Laplacian matrices of order m is a subset of the set of eigenvalues of standardized Laplacian matrices of order n.

Formalized. zonotope_polygon_subset_standardized_laplacian_spectra (arXiv:2610.14100 Corollary 2): For any integer n ≥ 2, the zonotope polygon S_n, defined as the set of linear combinations ∑_{m=1}^{n-1} t_m g_m with t_m ∈ [0, 1] and g_m = (1/n)(1 - exp(-2πi m/n)), is a subset of the standardized Laplacian spectrum Lambda_n.

Formalized. cyclic_mds_subgroup_rigidity (arXiv:2610.28047 Theorem 3.1 -- Length Bounds for Irreducible Cyclic MDS Codes): Theorem 3.1 of arXiv:2610.28047: Let F be a finite field of cardinality q and L a field extension of F of degree k. Let H be a cyclic subgroup of Lˣ of order n with 2 ≤ k ≤ n - 2. If every subset of H of size at most k is linearly independent over F…

Open. block_latin_filling_is_latin: For a Young diagram $\mu$ with $w = \text{young\_diagram\_part\_gcd}(\mu) > 0$, if the maximum multiplicity of any part length in $\mu$ is at most $w$, then the block filling with width $w$ is a Latin tableau of shape $\mu$ and type $\mu'$, satisfying…

Open. latin_wide_of_part_gcd_ge_max_multiplicity (arXiv:2610.29291 Theorem 3.4): Let $\mu$ be a Young diagram. If the maximum multiplicity of any part length of $\mu$ (young_diagram_max_multiplicity μ) is at most the greatest common divisor of all part lengths of $\mu$ (young_diagram_part_gcd μ), then $\mu$ is Latin…

Open. ghz_simple_graph_dim_six: For any finite type V with Fintype.card V = 6 and any simple graph G on V, the matching index ghz_matching_index G is at most 2.

Open. ghz_simple_graph_dim_eight: For any finite type V with Fintype.card V = 8 and any simple graph G on V, the matching index ghz_matching_index G is at most 3.

Open. ghz_simple_graph_dim_six_eight (arXiv:2610.22190 Connectivity Bounds for GHZ Graphs, Theorem 10): For any finite type V and any simple graph G on V, if Fintype.card V = 6 then ghz_matching_index G ≤ 2, and if Fintype.card V = 8 then ghz_matching_index G ≤ 3.

Open. ghz_matching_index_le_edge_connectivity (arXiv:2610.22190 Theorem 8): For every finite type V and every connected simple graph G on V, the matching index ghz_matching_index G is bounded above by the edge-connectivity graph_edge_connectivity G.

Verified in Lean

Daily reading

Daily reading: 0 research-level, 129 proved, 7 open

No research-level result today. In graph theory, OP-384 gained the case m = 1 determining the (1,3)-radio number of paths of order at least 8 alongside a reduction step in thread RT-4, while OP-444 gained an evaluation on its conjectured extremal glued cliques. In sequence and additive problems, OP-259 was proved for n = 2, OP-203 was proved surjective when started at 1 and 2, and thread RT-3 established a turn-allocation lemma for picking sequences. Published theorems formalized in full include loopless and coloopless characterizations of single-element matroid contractions and deletions, subgroup order rigidity for cyclic MDS codes, and properties of k-submodular kernels. Papers read today examined counterexamples to Gale's conjecture, quasi-kernels in digraphs, and Pareto inefficiency of EFX allocations, while seven formalized statements remain open.

Details

The Proposer read 16 new paper(s) and listed 34 open problem(s); 1 list entries were stated in Lean (Claude Code 1, the agents 0), as were 21 results the papers prove (Claude Code 1, the agents 20); the Proposer posed 68 more from what the knowledgebase already holds (102 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 129 theorem(s), 100 of them helper lemmas, which are not listed below; 7 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Answers OP-203. greedy_distinct_diff_seq_surjective (arXiv:2610.15860): Conjecture: for any distinct positive integers a₁, a₂, the greedy sequence with distinct consecutive differences started at a₁, a₂ takes every positive integer as a value (it is a permutation of the positive integers).

Partial result on OP-203. greedy_distinct_diff_seq_one_two_surjective (arXiv:2610.15860): The greedy sequence with distinct consecutive differences started at 1, 2 takes every positive integer as a value.

Partial result on OP-259. zero_sum_w_two (arXiv:2610.19965): For every r ≥ 1, w_r(2) = 2r + 1.

Partial result on OP-444. simple_graph_avg_connected_induced_order_clique_four_glued_clique (arXiv:2610.27553): For n ≥ 7, A(K_4 ·_v K_{n−3}) = (12 + (n−4)2^{n−5} + 2^{n−1} + (n−1)2^{n−2}) / (2^{n−4} + 2^{n−1} + 6): the connected vertex sets are the nonempty subsets of one clique minus the shared vertex, and all sets containing the shared vertex.

Partial result on OP-384. simple_graph_multi_radio_number_path_three_one_of_le (arXiv:2610.25863): For every n >= 8 the (1,3)-radio number of the path P_n is exactly 7: the case m = 1 of the thread's goal (8m - 1 for n >= 4m + 4).

Thread RT-4. simple_graph_multi_radio_number_path_three_of_le_of_forall: Reduction of the thread goal to one finite statement per m: if no (m,3)-radio labeling of P_{4m+4} has all labels at most 8m - 2, then the (m,3)-radio number of P_n is 8m - 1 for every n >= 4m + 4.

Thread RT-3. picking_sequence_outcome_update_mem_of_first_free: In a picking-sequence run, if agent i's first turn is turn t and the good g that i's new report r ranks first is still unallocated after the first t turns of the original run, then i gets g when it reports r (the others' reports unchanged).

Formalized. matroid_contract_singleton_coloopless_and_loopless_iff (arXiv:2610.27706 Lemma 2.1(ii)): Let M be a finite loopless and coloopless matroid on α and e ∈ M.E. Then the single-element contraction M.contract {e} is coloopless (i.e. (M.contract {e})✶.Loopless), and (M.contract {e}) is loopless if and only if e belongs to no two-element circuit of M…

Formalized. matroid_delete_singleton_loopless_and_coloopless_iff (arXiv:2610.27706 Lemma 2.1(i)): Let M be a finite loopless and coloopless matroid on α and e ∈ M.E. Then the single-element deletion M.delete {e} is loopless; moreover, it is coloopless ((M.delete {e})✶.Loopless) if and only if e belongs to no two-element cocircuit of M (¬ ∃ C,…

Formalized. residual_matching_num_single_edge_exchange_iff_critical (arXiv:2610.27833, Lemma 1): Let G be a finite simple graph on a finite vertex type V, M a maximum matching of G, u an M-free vertex, and vw an edge of M such that uv is an edge of G. Then there exists a maximum matching M' of G with edge set (M \ {vw}) ∪ {uv}, and the matching number…

Formalized. three_subset_sum_ten_or_eleven_pair_balanced (arXiv:2610.27956, proof of Theorem 10): Let s be a 3-element subset of Fin 6. If the sum over x in s of (x.val + 1) is 10 or 11, then s contains exactly one element from {0, 1}, exactly one element from {2, 3}, and exactly one element from {4, 5}.

Formalized. planar_necklace_two_parallel_cuts_six_points (arXiv:2610.27956 Theorem 10): Let P be a finite set of 6 points in ℝ × ℝ with a coloring c : (ℝ × ℝ) → Fin 2 having exactly 3 red points (color 0) in P and 3 blue points (color 1) in P. Then there exist a non-zero linear map φ : (ℝ × ℝ) →ₗ[ℝ] ℝ, thresholds t₁, t₂ ∈ ℝ, and an assignment…

Formalized. subgroup_rigidity_order_of_mds (arXiv:2610.28047 Theorem 3.1): Let H be a cyclic subgroup of order n in the multiplicative group of a degree-k field extension L of F_q with 2 <= k <= n - 2. If every subset of H of size at most k is linearly independent over F_q, then for every divisor d > 1 of n, the multiplicative…

Formalized. order_rigidity_arithmetic_consequences (arXiv:2610.28047 Theorem 3.1): If 2 <= k <= n - 2 and ord_d(q) = k for every divisor d > 1 of n, then every prime l dividing n satisfies ord_l(q) = k and k | (l - 1), n is odd, gcd(n, q-1) = gcd(n, k) = 1, and n = mk + 1 for some integer m >= 2.

Formalized. k_submodular_of_monotone_of_orthant_diminishing_returns (arXiv:2610.26677, proof of Lemma 2.2): A monotone function on assignments Fin n → Option (Fin k) that satisfies orthant diminishing returns is k-submodular.

Corrected. not_straight_line_inter_unit_sphere_card_le_two: The statement straight_line_inter_unit_sphere_card_le_two is false: In Lean, Metric.sphere (0 : ℝ × ℝ) 1 is the sphere for the default product metric on ℝ × ℝ, whose norm is the sup norm ‖(x,y)‖ = max |x| |y| (this is Prod.norm_def /…

Corrected. not_k_submod_hidden_fn_properties: The statement k_submod_hidden_fn_properties is false: The definition is_k_submod_kernel lacks a differentiability/continuity hypothesis on H. In Lean, deriv f x evaluates to 0 when f is not differentiable at x. Consequently, discontinuous functions…

Corrected. not_k_submod_hidden_fn_orthant_diminishing_returns: The statement k_submod_hidden_fn_orthant_diminishing_returns is false: The definition is_k_submod_kernel lacks differentiability/continuity hypotheses on H. In Lean, deriv f x evaluates to 0 when f is not differentiable at x. Consequently,…

Corrected. not_greedy_fraction_le_of_blocking_ge: The statement greedy_fraction_le_of_blocking_ge is false: The statement is false as written; a machine-checked refutation is provided. Counterexample: take P := fun t => t = ∅ (so P holds only on the empty family), ℓ = 1, p = 0. * P is hereditary: P ∅…

Open. tournament_arborescence_hadamard_bound (arXiv:2610.25886 Proposition 4.1 -- Extremal arborescences in graph orientations): Every tournament T of order n ≥ 2 with outdegrees d_1, ..., d_n satisfies τ(T) ≤ √(∏_{i=1}^n ((2d_i+1)^2 + n)) / (2^{n-1} √(n(n+1))) ≤ (n(n+1)/4)^{(n-1)/2}.

Open. optimal_single_track_gray_code_equally_spaced_of_necklace_hamilton_cycle (arXiv:2610.24758 Lemma 1): For any prime n ≥ 3, if the non-constant binary necklace graph Q_n has a Hamilton cycle, then there exists an optimal single-track Gray code of length n and period 2^n - 2 with equally spaced heads, where the head positions are multiples of (2^n - 2) / n and…

Open. planar_necklace_straight_line_cut_lower_bound (arXiv:2610.27956 Proposition 1): For any natural numbers k, r, b with 2 ≤ k, 0 < r, 0 < b, and k dividing both r and b, there exists a finite set P of r + b points in ℝ² and a coloring c : (ℝ × ℝ) → Fin 2 with exactly r points of P colored 0 and b points of P colored 1, such that for any…

Open. f2_subset_eight_universal_spectrum (arXiv:2610.24022, Theorem 3.1): Let d ≥ 3. For every subset T of Fin d → ZMod 2 with cardinality 8, there exists a subset B of Fin d → ZMod 2 of cardinality 2^(d - 3) such that for all distinct b, b' ∈ B, the character sum ∑ t ∈ T, (-1 : ℤ) ^ (dotProduct (b + b') t).val is nonzero.

Open. four_point_set_quadratic_packing_interval (arXiv:2610.24281 Lemma 3.2 -- The Heesch number in E^3 is unbounded): Let n ≥ 4 be an integer, and let a = 2n - 1 and b = 2n + 1. Then there exists a finite subset C ⊆ ℤ such that the translates of F = {0, a, b, a + b} by elements of C are pairwise disjoint, and their union covers the integer interval [a + 2b, ab - 1 - a - b].

Open. ordinal_mst_round_cost_and_contraction (arXiv:2610.23032 On the Metric Distortion of the Minimum Spanning Tree Problem, Lemma 6.1): In Algorithm 1 on finite vertex type V with decidable equality, for any round from forest F to forest F' (boruvka_round π F F') under profile π : V → LinearOrder V consistent with pseudometric d : V → V → ℝ, and any minimum spanning tree T_mst under d…

Verified in Lean

Daily reading

Daily reading: 0 research-level, 72 proved, 16 open

No research-level result today. In combinatorics and graph theory, open-problem partial results bounded the k-color poset Ramsey number of the 5-diamond by 5k for OP-270, identified all four-vertex non-sustainable trees for OP-223, and determined the achromatic numbers of cubic Paley graphs at orders 13, 19, and 31 for OP-442. Published theorems formalized in full include the (2k + 2)-competitiveness of 1-FIFO for paging with delay from arXiv:2610.18767, while research threads advanced on covers of cycle powers in RT-1 and pure Nash equilibria of round-robin allocations in RT-3. New papers treated the MDS conjecture, binding numbers of clique-free graphs, envy-free cake cutting, and Hadamard matrix constructions. Sixteen statements remain open, including the binding number bound of 3 for K_5-free graphs and distortion bounds for ordinal Borůvka minimum spanning trees.

Details

The Proposer read 17 new paper(s) and listed 34 open problem(s); 3 list entries were stated in Lean (Claude Code 3, the agents 0), as were 12 results the papers prove (Claude Code 2, the agents 10); the Proposer posed 40 more from what the knowledgebase already holds (71 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 72 theorem(s), 41 of them helper lemmas, which are not listed below; 16 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Partial result on OP-270. five_diamond_poset_ramsey_number_le (arXiv:2610.20545): For every natural number k with 1 ≤ k, the k-color poset Ramsey number of the 5-diamond DiamondPoset 5 is at most 5 * k.

Partial result on OP-223. op223_nonsustainable_tree_fin_four (arXiv:2610.17078): Every non-sustainable tree on four vertices is isomorphic to the star graph K_{1,3}.

Answers OP-360. op_360_single_counter_update_strictly_less_expressive (arXiv:2610.23495): There exists a symmetric 2-counter machine M over a finite state type Q and states q, q' in Q whose reachability relation two_counter_machine_reach_rel M q q' cannot be realized as the reachability relation two_counter_machine_reach_rel M' q0 q1 between any…

Partial result on OP-422. simple_graph_binding_number_complete_four_partite (arXiv:2610.28536): The complete 4-partite graph with four parts of size m ≥ 1 is K_5-free and has binding number exactly 3, so the bound 3 conjectured for K_5-free graphs would be sharp.

Partial result on OP-442. simple_graph_achromatic_number_le_max_degree_or_half (arXiv:2610.27510): For every finite graph, χ_a(G) ≤ max{Δ(G) + 1, ⌊n/2⌋}: if some colour class is a single vertex, that vertex has a neighbour in each of the other classes; otherwise every class has at least two vertices.

Partial result on OP-442. cubic_paley_graph_zmod_thirteen_achromatic_number (arXiv:2610.27510): The cubic Paley graph G_3(13) has achromatic number 6 = (13 − 1)/2, the upper bound, above the value (13 + 2)/3 = 5 conjectured for large q.

Partial result on OP-442. cubic_paley_graph_zmod_nineteen_achromatic_number (arXiv:2610.27510): The cubic Paley graph G_3(19) has achromatic number 9 = (19 − 1)/2, the upper bound, above the value (19 + 2)/3 = 7 conjectured for large q.

Partial result on OP-442. cubic_paley_graph_zmod_thirty_one_achromatic_number (arXiv:2610.27510): The cubic Paley graph G_3(31) has achromatic number 15 = (31 − 1)/2, the upper bound, above the value (31 + 2)/3 = 11 conjectured for large q.

Thread RT-3. round_robin_pne_envy_zero_ef1_of_deviation_certificate: If the deviation certificate of round_robin_two_mul_add_one_deviation_certificate holds for agent i ≠ 0, then at a pure Nash equilibrium of Round-Robin (2n+1 goods, nonnegative additive valuations) agent i is EF1 towards agent 0.

Thread RT-3. picking_prefix_state_update_of_not_mem_take: Changing agent i's ranking does not change the state of a picking sequence after the first t turns if agent i does not pick in those turns.

Thread RT-1. cycle_power_seven_tight_type_sequence_one_iff_six: Layer-free form of h16. Let x p j say that the class starting at p (the class of the distance-7 pair (p, p+7)) contains p + j. If, around the cycle: every vertex is an inner element (offsets 1..6) of exactly two such classes; every distance-4 and distance-6…

Thread RT-1. simple_graph_cycle_power_seven_cover_four_tight_one_iff_six: In a tight cover of C_n^7 (n ≥ 22) by 4 equivalence subgraphs (every distance-7, 6, 4 pair covered exactly once, every class spanned by a distance-7 pair), the class starting at i contains i+1 exactly when it contains i+6: the hypothesis h16 of…

Thread RT-1. simple_graph_cycle_power_seven_cover_four_layer_weight_eq: In any cover of C_n^7 (n ≥ 22) by 4 equivalence subgraphs, every layer has weight exactly n: the sum over i of 2·[i ~ i+7] + [i ~ i+6] + [i ~ i+4] in that layer equals n.

Thread RT-1. simple_graph_cycle_power_seven_layer_weight_eq_class_five_imp_one_two_imp_six: Let c label the vertices of C_n^7 (n ≥ 22) so that every label class is a clique, with total weight Σ_i (2·[i ~ i+7] + [i ~ i+6] + [i ~ i+4]) exactly n. Then every class containing i and i+7 is tight: if it contains i+5 it contains i+1, and if it contains…

Formalized. saddle_gap_marginal_payoff_bounds (arXiv:2610.25524 Lemma 4.2): Let n be a positive natural number. For any matrix A : Matrix (Fin n) (Fin n) ℝ with non-negative entries and any optimal row strategy σ in (saddle_gap_game A).optimalRowStrategies, let a = saddle_gap_first_marginal σ and b = saddle_gap_second_marginal σ…

Formalized. card_edgeFinset_ge_55_of_no_indep_six (scout/2610-24484-erd-h-o-s-gy-arf-as-balanced-five-colorings.md:1291): If a simple graph on 26 vertices has no independent set of size 6, then it has at least 55 edges.

Formalized. balanced_five_coloring_color_class_property_l (scout/2610-24484-erd-h-o-s-gy-arf-as-balanced-five-colorings.md:1283): For an arbitrary type V, in any balanced edge coloring c : Sym2 V → Fin 5, every color class graph edge_color_class_graph c k satisfies property (L).

Formalized. balanced_five_coloring_exists_class_card_edgeFinset_le_65 (scout/2610-24484-erd-h-o-s-gy-arf-as-balanced-five-colorings.md:1287): For an arbitrary finite type V with 26 elements, in any edge coloring c : Sym2 V → Fin 5, some color class graph has at most 65 edges.

Formalized. balanced_five_coloring_exists_color_class_bounds_and_properties (scout/2610-24484-erd-h-o-s-gy-arf-as-balanced-five-colorings.md:324): For an arbitrary finite type V with 26 elements, if c : Sym2 V → Fin 5 is a balanced edge coloring, then some color class graph has no independent set of size 6, satisfies property (L), and has between 55 and 65 edges.

Formalized. rank_potential_offline_fetch_bound (arXiv:2610.18767 Lemma 3.3(i)): At an offline fetch, the increase in rank potential is at most k (Lemma 3.3(i)).

Formalized. rank_potential_offline_evict_hole_bound (arXiv:2610.18767 Lemma 3.3(ii)): At an offline fetch, if the evicted page is outside the online cache, the change in rank potential is at most 0 (Lemma 3.3(ii)).

Formalized. payment_window_amortized_bound (arXiv:2610.18767 Section 3.2, Lemma 3.4): For any page type α, any natural number k, and any payment window w : PaymentWindow α k, the offline cost covers the online payment and potential change: 1 ≤ (k + 1 : ℝ) * (w.delta_fetches_opt : ℝ) + (k + 1 : ℝ) * w.delta_delay_opt - (w.delta_phi : ℝ).

Open. boruvka_round_cost_le: In any Borůvka round from F to F' under profile π : V → LinearOrder V consistent with pseudometric d : V → V → ℝ on finite type V, the cost difference graph_cost d F' - graph_cost d F is at most 2 * graph_cost d T_mst, where T_mst is a minimum spanning tree…

Open. ordinal_mst_boruvka_distortion_le (arXiv:2610.23032 Theorem 2): Let V be a finite type with at least 2 vertices, d : V → V → ℝ a pseudometric, and π : V → LinearOrder V an ordinal profile consistent with d. If F is an output of Algorithm 1 on π and T_mst is a minimum spanning tree under d, then graph_cost d F ≤ 2 *…

Open. optimalTorusError_pred (arXiv:2610.19174, Listing A.15): For any Boolean function f : (Fin n → Bool) → Bool on n > 0 variables, optimalTorusError f (n - 1) equals explicitError f.

Open. explicit_optimal_torus_approximation (arXiv:2610.19174, Theorem 4.1.1): For any Boolean function f : (Fin n → Bool) → Bool on n > 0 variables, explicitApproximant f belongs to degreeSpace (K := ℝ) (Fin n) (n - 1), satisfies torusDistance (explicitApproximant f x - bit (f x) / 2) = explicitError f at every x : Fin n → Bool,…

Open. simple_graph_binding_number_le_three_of_clique_free_five (arXiv:2610.28536): The case k = 5 of the binding-number conjecture for K_k-free graphs (open; arXiv:2610.28536 proves k ≥ 6, and it fails for k = 4): every finite nonempty K_5-free graph has binding number at most 3.

Open. cubic_paley_graph_achromatic_number_eventually_eq (arXiv:2610.27510): The conjecture of arXiv:2610.27510 after Theorem 1.4: for every sufficiently large prime power q ≡ 1 (mod 6), the cubic Paley graph G_3(q) has achromatic number exactly (q + 2)/3 (the paper's lower bound). Small cases point the other way: χ_a(G_3(13)) = 6…

Verified in Lean

Daily reading

Daily reading: 0 research-level, 146 proved, 21 open

No research-level result today; the day's work is listed below.

Details

The Proposer read 28 new paper(s) and listed 56 open problem(s); 18 list entries were stated in Lean (Claude Code 3, the agents 15), as were 5 results the papers prove (Claude Code 3, the agents 2); the Proposer posed 11 more from what the knowledgebase already holds (97 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 146 theorem(s); 21 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Proved. simple_graph_multi_radio_number_path_three_lt (arXiv:2610.25863): The threshold 4m + 4 is sharp: the (m,3)-radio number of P_{4m+3} is less than 8m - 1.

Proved. spider_blocked_none_one_inward_one_outward: For the configuration κ₁ = (none;{(1,inward),(1,outward)}), the blocked family V^φ_κ₁(r) equals r for every r ≥ 1 and both φ.

Proved. spider_eventual_parity_two_three_blocked_small: For the configuration κ0 = (none;{2^*,3^*}), the blocked family at level t is 2 for t = 1 and 6 for t = 2, for both far ends, and it is at most 7 for t = 3, 4, 5.

Proved. spider_blocked_two_three_level_five: At level t = 5, the inward blocked family for the configuration (none; {2^*, 3^*}) at canonical fuel 11 is at most 7.

Proved. nfa_run_count_list_concat: Number of runs on w++[a] from p to q equals the sum over s of the number of runs on w from p to s, taken over the predecessors s of q on letter a.

Proved. pell_prefix_base_count: The prefix c_1c_2 has 1 run from q_1 to q_2, 2 runs from q_1 to q_3, 1 run from q_1 to q_4, and no runs to any other state, in A_n.

Proved. spider_blocked_two_three_level_five_sub_incoming_two_three: With centre incoming (resp. outgoing) and free legs of length 2 and 3, the value at fuel 10 is 3 (both).

Proved. spider_blocked_two_three_level_five_sub_sum_six: The two sub-states with free centre, a short leg of length 1 whose far end is inward (resp. outward), and the free legs 2 and 3, have value 2 at fuel 10.

Proved. spider_blocked_two_three_level_five_sub_sum_seven: The two sub-states with free centre where a leg of length 2 with far end inward (resp. outward) is added to the free legs 2 and 3 have value 6 at fuel 10.

Proved. spider_blocked_two_three_level_five_sub_sum_eight: Four sub-states at fuel 10: the two states obtained by cutting the long leg at j = 4, which add a leg of length 3 with far end inward (resp. outward) to the free legs 2 and 3, have value 7; and the two states obtained by cutting the free leg of length 3 at j…

Proved. spider_blocked_two_three_level_five_sub_sum_nine: Four sub-states at fuel 10: the states (incoming;{(5,inward),(3,free)}) and (outgoing;{(5,inward),(3,free)}) have value 4 and 6, and the states with free centre obtained by cutting the free leg of length 3 at j = 1, which add a leg of length 2 with far end…

Proved. spider_blocked_two_three_level_five_sub_short_legs: With centre incoming (resp. outgoing) and the legs (5,inward) and the free leg 2, the value at fuel 10 is 0 (resp. 1).

Proved. spider_blocked_two_three_level_five_sub_long_leg_j34: With free centre, the state obtained from (none;{(5,inward),2^*,3^*}) by cutting the (5,inward) leg at j = 4 leaves a leg of length 2 with far end inward (resp. outward) in front of (5,inward) and the free leg 2; the values at fuel 10 are 0 and 1.

Open. op223_sustainable_tree_fin_four (arXiv:2610.17078): Every sustainable tree on four vertices is isomorphic to the path graph on four vertices.

Open. op223_nonsustainable_tree_fin_four (arXiv:2610.17078): Every non-sustainable tree on four vertices is isomorphic to the star graph K_{1,3}.

Open. op_359_symmetric_two_counter_reachability_semilinear (arXiv:2610.23495): For every symmetric 2-counter machine M with a finite state type Q and any pair of states q, q' in Q, the reachability relation two_counter_machine_reach_rel M q q' is a semilinear subset of (ℕ × ℕ) × (ℕ × ℕ).

Open. op_153_subexp_bound_intersective_quadratic (arXiv:2610.07765): Let $h \in \mathbb{Z}[X]$ be an intersective polynomial of degree 2. Do there exist constants $c > 0$ and $C > 0$ depending only on $h$ such that for all integers $N \ge 3$, every subset $A \subseteq \{1, \dots, N\}$ whose distinct elements have difference…

Open. optimal_asymptotic_base_constant_finitely_ambiguous_nfa (arXiv:2610.08525): The asymptotic base constant c = lim_{n → ∞} M(n)^{1/n} for the maximum degree of ambiguity M(n) of trim finitely ambiguous n-state NFAs equals 1 + Real.sqrt 2.

Open. op100_linear_flattening_potency_gt_even (arXiv:2610.08504): For every even natural number n ≥ 4, there exists a linear flattening on (ℂⁿ)⊗³ with potency strictly greater than 2n - 2, so the secant variety σ_{2n-2}((ℂⁿ)⊗³) is testable by the linear flattening method.

Open. op_98_objective_switching_weighted_four_set_packing (arXiv:2610.10036): Open Problem OP-98 (arXiv:2610.10036, Sec. 5): there exists a rational approximation ratio c < 5/2 such that for every type α with decidable equality, every weighted 4-set packing instance inst, and every packing A that is square locally optimal with respect…

Open. undirected_binary_linear_unicast_advantage (arXiv:2610.10108): Open Problem OP-90: there exist a finite vertex type V with decidable equality, a finite session type K with decidable equality, a capacitated network net on V, unicast sessions sess on V indexed by K, and a rate r ∈ ℝ such that LinearCodeSupportsRate (ZMod…

Open. triangle_free_blocking_ge_vertex_cover (arXiv:2610.23130): For every finite family of line segments L in ℝ² whose crossing graph is triangle-free, the blocking number of L under the triangle-free crossing property is at least the minimum vertex cover size of G(L).

Open. op_360_single_counter_update_strictly_less_expressive (arXiv:2610.23495): There exists a symmetric 2-counter machine M over a finite state type Q and states q, q' in Q whose reachability relation two_counter_machine_reach_rel M q q' cannot be realized as the reachability relation two_counter_machine_reach_rel M' q0 q1 between any…

Open. round_robin_alpha_three_eq_two_thirds (arXiv:2610.23686): The best guaranteed EF1 factor for Round-Robin pure Nash equilibria with three agents is 2/3.

Open. round_robin_item_preserving_swaps_polynomial_convergence (arXiv:2610.23040): For any fixed number of agents n ≥ 3, there exist natural numbers C, k : ℕ such that for any number of goods m : ℕ (modeled by Fin m), any initial preference profile pref : Fin n → Fin m ≃ Fin m, and any non-negative additive valuations v : Fin n → Fin m → ℝ…

Verified in Lean

Daily reading

Daily reading: 1 research-level, 110 proved, 7 open

Research-level results today: 1. Shows the published claim corrected by arXiv:2610.23686 survives in part: at every Round-Robin pure Nash equilibrium, envy towards any agent with at most two goods is already EF1, for any number of agents (so every equilibrium with at most 2n goods is EF1, while the paper's counterexample has three agents and eight goods). Results that bear on the same open problem or paper share one Research Blog post.

Details

The Proposer read 13 new paper(s), listed 32 open problem(s), and stated 11 of them in Lean, and posed 9 more from what the knowledgebase already holds (73 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 110 theorem(s); 7 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Proved. round_robin_three_agents_pne_not_alpha_ef1_of_three_quarters_lt (arXiv:2610.23686): With three agents and eight goods, there is an instance with nonnegative additive valuations and a pure Nash equilibrium of Round-Robin whose outcome is not α-EF1 for any α > 3/4. Witness: agents 1 and 3 value every good 1; agent 2 values goods (0..7) at…

Research-level. round_robin_pne_ef1_pair_of_count_le_two: In a pure Nash equilibrium of Round-Robin (nonnegative additive valuations), agent i is EF1 towards any agent j that appears at most twice in the picking sequence. — Shows the published claim corrected by arXiv:2610.23686 survives in part: at every Round-Robin pure Nash equilibrium, envy towards any agent with at most two goods is already EF1, for any number of agents (so every equilibrium with at most 2n goods is EF1, while the paper's counterexample has three agents and eight goods)

Formalized. diamond_poset_ramsey_number_le (arXiv:2610.20545): For all natural numbers j ≥ 2 and k ≥ 1, the k-color poset Ramsey number of the j-diamond DiamondPoset j is at most j * k.

Formalized. five_diamond_poset_ramsey_number_ge (arXiv:2610.20545): For every natural number k with 1 ≤ k, the k-color poset Ramsey number of the 5-diamond DiamondPoset 5 is at least 4 * k.

Proved. mixed_diamond_fan_ramsey_bound: For j ≥ 2, a colouring of the subsets of Fin d with a + b colours contains a monochromatic induced j-diamond in one of the first a colours, or a monochromatic induced F_j in one of the last b colours, provided d ≥ j*a + (j-1)*b + 1 when b > 0 and d ≥ j*a…

Proved. diamond_poset_ramsey_property: For j ≥ 2 and k ≥ 1, every colouring of the subsets of Fin (j*k) with k colours has a monochromatic induced copy of the j-diamond DiamondPoset j.

Proved. exists_poset_ramsey_property_diamond: For j ≥ 2 and k ≥ 1 the k-colour poset Ramsey property for the j-diamond is satisfied by some d: there is a d such that every colouring of the subsets of Fin d with k colours contains a monochromatic induced copy of DiamondPoset j (i.e. the set of such d is…

Proved. mixed_diamond_fan_ramsey_bound_all_singletons_fan_colours: Case b ≥ 2 with every singleton fan-coloured: if d ≥ j*a + (j-1)*b + 1 and the statement holds at (a, b-1) in dimension d-(j-1), then χ has a monochromatic diamond in one of the a diamond colours or a monochromatic F_j in one of the b fan colours.

Proved. mdfrbf_compression_bound: Arithmetic side condition of the compression step: if j ≥ 2, a ≥ 1, b ≥ 2 and j*a + (j-1)*b + 1 ≤ d, then j*a + (j-1)*(b-1) + 1 ≤ d - (j-1).

Proved. mdfrbf_compression_card: If W is a set of j-1 points of Fin d (j ≥ 2), then the complement of W has exactly d - (j-1) elements, so it can be reindexed by Fin (d - (j-1)).

Proved. mdfrbf_fibre_subset: Pigeonhole: if f : Fin d → Fin b and b*n < d, then some fibre of f contains n+1 points; equivalently there are a colour c and a set W of n+1 points on which f is constantly c.

Proved. mdfrbf_compress_subset_iff: The compression map is an order embedding: for all sets A, A' of points outside W, T A ⊆ T A' holds if and only if A ⊆ A'. Consequently T is injective, monotone, and reflects inclusions (hence strict inclusions), which is what lets an induced copy be pushed…

Proved. mdfrbf_diamond_transfer: Push-forward of a monochromatic diamond along an order embedding of set systems: if T is injective, monotone and reflects inclusions on finsets, χ (T A) = f (ψ A) for every nonempty A, ψ ∅ ≠ col, and ψ has a monochromatic j-diamond in colour col, then χ has…

Proved. mdfrbf_fan_transfer: Push-forward of a monochromatic fan along an order embedding of set systems: if j ≥ 2, T is injective, monotone and reflects inclusions, χ (T A) = f (ψ A) on nonempty A, and ψ has a monochromatic F_j in colour col, then χ has a monochromatic F_j in colour f…

Open. op223_star_is_nonsustainable (arXiv:2610.17078): For any natural number n ≥ 4 and any simple graph S on a finite vertex type V, if S is isomorphic to the star graph completeBipartiteGraph (Fin 1) (Fin (n - 1)), then S is a non-sustainable tree.

Open. op182_deferred_to_deadline_beyond_half_against_stopping (arXiv:2610.13775): Refutation of Open Problem OP-182 against the deferred-until-stopping benchmark: there does not exist any constant c > 1/2 such that for every Pandora's problem instance with deadlines inst : PandoraDeadlineInstance, there exists a valid adaptive policy…

Open. iterated_round_robin_two_agents_item_preserving_swaps_le (arXiv:2610.23040 Theorem 3.12 (n = 2: IRR reaches a PNE under item-preserving swaps after at most m(m-1)/2 manipulations); stated as:…): For two agents with additive utilities consistent with strict rankings, iterated Round Robin from truthful bids, where at each step an agent with a strictly profitable item-preserving swap replaces its bid by a best response among its item-preserving swaps,…

Open. iterated_round_robin_two_agents_adjacent_swaps_converges (arXiv:2610.23040): Conjecture 3.14 of arXiv:2610.23040: with two agents, m >= 5 goods and additive nonnegative utilities consistent with strict rankings, iterated Round Robin from truthful bids, where at each step an agent with a strictly profitable adjacent swap replaces its…

Open. round_robin_three_agents_seven_goods_pne_ef1: With three agents and seven goods, every pure Nash equilibrium of Round-Robin (nonnegative additive valuations) is EF1; so the eight-good counterexample of arXiv:2610.23686 Proposition 1 is the smallest possible.

Open. round_robin_pne_ef1_of_card_le_two_mul_add_one: With n agents and at most 2n + 1 goods, every pure Nash equilibrium of Round-Robin (nonnegative additive valuations) is EF1. For three agents this is tight: arXiv:2610.23686 Proposition 1 has a non-EF1 equilibrium with eight goods. Whether 2n + 2 goods can break EF1 for every n >= 4 is open.

Open. round_robin_pne_envy_zero_ef1_of_two_mul_add_one: With n agents and 2n + 1 goods, in every pure Nash equilibrium of Round-Robin (nonnegative additive valuations) every agent i ≠ 0 is EF1 towards agent 0, the only agent with three goods. With round_robin_pne_ef1_pair_of_count_le_two and…

Verified in Lean

Weekly report

Week of 12–18 Oct 2026: Negative stamp values extend postage range, and mutual information subadditivity fails

This week settled several open problems across discrete mathematics and information theory. It was established that allowing negative denominations strictly increases the consecutive integers reachable by postage stamp combinations. In information theory, subadditivity of mutual information was refuted for Hellinger distance and quarter-power divergences. In graph theory, new findings classified when spectral bounds on clique numbers hold, answered questions about permutation groups generated by graph hops, and established exact covering numbers for powers of cycle graphs.

Details

Proved. Four stamps with one negative value reach 99 consecutive postage amounts

Proved. Equivalence covering number of cycle cubes equals 4 when length is 2 modulo 4

Proved. Complete classification of strong spectral clique bounds for five-block graphs

Disproved. Mutual information subadditivity fails for Hellinger divergence and quarter-power divergences

Lean accepted 360 new theorems this week, drawn from 38 papers and 16 topics. Of these, 22 are new here, 0 correct a published claim, 74 formalize a published result and 264 are helper lemmas. 83 statements posed this week are not yet proved. The reading covered 81 new papers. This report covers 11 results in 7 fields.

Daily reading

Daily reading: 159 proved, 12 open

The Proposer read 0 new paper(s), listed 0 open problem(s), and stated 5 of them in Lean, and posed 0 more from what the knowledgebase already holds (48 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 159 theorem(s); 12 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Formalized. finset_card_ge_one_add_clog_of_midpoint_balanced (arXiv:2610.09349): For a prime p, every nonempty balanced subset B of Z/pZ has |B| ≥ 1 + ⌈log₂ p⌉.

Formalized. nus_min_card_ge_one_add_clog (arXiv:2610.09349): For an odd prime p, m(p) ≥ 1 + ⌈log₂ p⌉: every subset of Z/pZ with no unique sums has at least 1 + ⌈log₂ p⌉ elements.

Formalized. edge_connectivity_ge_min_degree_mul_lly_lower_bound_two (arXiv:2610.10559 (Edge-Connectivity versus Lin--Lu--Yau Curvature), Theorem 1.1 (for finite graphs).): Every finite connected graph of diameter 2 has κ'(G) ≥ δ(G) · κ_LLY^(2)(G).

Formalized. lly_curvature_pos_of_edge_connectivity_ge (arXiv:2610.10559): If a finite connected graph on n vertices has edge-connectivity κ'(G) ≥ 2n/3 − 1, then every edge xy has κ_LLY(x, y) > 0.

Proved. lly_positive_edge_connectivity_threshold_five (arXiv:2610.10559): c(5) = 2: every connected graph on 5 vertices with edge-connectivity at least 2 has positive Lin–Lu–Yau curvature on every edge, while the path-like tree with a vertex of degree 3 (edges 01, 03, 04, 12) has κ' = 1 and κ_LLY(0,1) = −1/3. So c(5) < ⌈2·5/3 − 1⌉…

Proved. midpoint_certificate_matrix_minor_abs_le: The absolute value of every k x k minor of a (k+1) x (k+1) midpoint-certificate matrix is bounded by 2^k, proved by multilinear expansion into 2^k unit-difference determinants and applying total unimodularity.

Proved. midpoint_certificate_matrix_minor_dvd_p: For an injective certificate vector b in (ZMod p)^{k+1} satisfying the midpoint equations with k > 0, the shifted vector b - b(j₀) restricts to a non-zero vector in the kernel of the k x k minor modulo p, so its determinant vanishes modulo p and is divisible…

Proved. finset_midpoint_balanced_exists_nonzero_minor: Every nonempty midpoint-balanced subset B of ZMod p contains an inclusion-minimal balanced subconfiguration of size k+1 ≤ |B| with k > 0 whose certificate matrix has at least one non-vanishing minor.

Proved. exists_min_cut_partition: For every finite connected graph G on at least two vertices, there exists a non-trivial cut partition (X, X^c) with crossing edges at most the edge-connectivity κ'(G).

Proved. graph_lly_curvature_lower_bound_two_le_one: For every finite connected graph G, the lower Lin–Lu–Yau curvature bound at scale 2 is at most 1: κ_LLY^(2)(G) ≤ 2/2 = 1.

Proved. crossing_pairs_card_eq_crossing_edges_card: For a finite simple graph G and a vertex set A, the number of ordered adjacent pairs (a, b) with a ∈ A and b ∈ Aᶜ equals the number of edges of G crossing the cut (A, Aᶜ).

Proved. cross_cut_card_ge_minDegree_of_diam_two: Plesník's cut lemma: in a finite connected graph G with diam(G) = 2, every nontrivial vertex cut (A, Aᶜ) has at least δ(G) crossing edges.

Open. op_270_five_diamond_poset_ramsey_number (arXiv:2610.20545): For every natural number k with 2 ≤ k, the k-color poset Ramsey number of the 5-diamond DiamondPoset 5 is equal to 4 * k.

Open. diamond_poset_ramsey_number_le (arXiv:2610.20545): For all natural numbers j ≥ 2 and k ≥ 1, the k-color poset Ramsey number of the j-diamond DiamondPoset j is at most j * k.

Open. five_diamond_poset_ramsey_number_ge (arXiv:2610.20545): For every natural number k with 1 ≤ k, the k-color poset Ramsey number of the 5-diamond DiamondPoset 5 is at least 4 * k.

Open. five_diamond_poset_ramsey_number_le (arXiv:2610.20545): For every natural number k with 1 ≤ k, the k-color poset Ramsey number of the 5-diamond DiamondPoset 5 is at most 5 * k.

Open. odd_antihole_colorable_of_common_palette (arXiv:2610.14241): Theorem 8 of arXiv:2610.14241 for the odd antihole: for ℓ ≥ 3, if all lists of an (ℓ-1)-assignment of complement(C_{2ℓ-1}) are subsets of a common ℓ-set P and are not all identical, then the antihole is L-colorable.

Open. op_258_triangle_forcing_implies_decay_ramsey (arXiv:2610.20577): For every sequence a : ℕ+ → ℝ, if a is positive, antitone, and has sum strictly greater than 1 in ℝ≥0∞, and if a forces a monochromatic triangle K_3 on sufficiently large complete graphs under a-bounded edge colourings, then for every integer k ≥ 3, a forces…

Open. odd_antihole_strongly_chromatic_choosable (arXiv:2610.14241): Conjecture 9 of arXiv:2610.14241: for every integer ℓ ≥ 3, the odd antihole, the complement of the cycle C_{2ℓ-1}, is strongly ℓ-chromatic-choosable: it has chromatic number ℓ, and for every assignment of lists of size ℓ-1 to its vertices that are not all…

Open. cycle_nine_compl_strongly_chromatic_choosable (arXiv:2610.14241): The case ℓ = 5 of Conjecture 9 of arXiv:2610.14241: the odd antihole complement(C_9) is strongly 5-chromatic-choosable: χ = 5, and every assignment of lists of size 4 that are not all identical admits a proper list colouring.

Open. cycle_eleven_compl_strongly_chromatic_choosable (arXiv:2610.14241): The case ℓ = 6 of Conjecture 9 of arXiv:2610.14241: the odd antihole complement(C_11) is strongly 6-chromatic-choosable: χ = 6, and every assignment of lists of size 5 that are not all identical admits a proper list colouring.

Open. glove_number_k4 (arXiv:2610.17602): The glove number of the complete graph $K_4$ is equal to 3.

Open. simple_graph_cycle_power_seven_cover_four_tight_dvd: A tight cover of C_n^7 (n ≥ 22) by 4 equivalence subgraphs (every distance-7, 6, 4 pair covered exactly once, every class spanned by a distance-7 pair) exists only when 8 divides n.

Open. simple_graph_not_has_equivalence_cover_cycle_power_seven_of_not_dvd: For n ≥ 22 not divisible by 8, C_n^7 cannot be covered by 4 equivalence subgraphs; this is the lower bound eq(C_n^7) ≥ 5 = m + 2 of the open case k = 7 (m = 3) of OP-205.

Verified in Lean

Daily reading

Daily reading: 36 proved, 10 open

The Proposer read 0 new paper(s), listed 0 open problem(s), and stated 3 of them in Lean, and posed 0 more from what the knowledgebase already holds (46 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 36 theorem(s); 10 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. simple_graph_cycle_power_six_one_sanity: Sanity check: in C_6 (the first power of the 6-cycle), vertices 0 and 5 are adjacent (they wrap around) and 0 and 2 are not.

Proved. simple_graph_has_equivalence_cover_mono: If G is covered by s equivalence subgraphs, it is covered by any larger number t of them (pad with the edgeless equivalence subgraph).

Proved. simple_graph_equivalence_covering_number_top: Sanity check: a complete graph with at least two vertices has equivalence covering number 1.

Formalized. simple_graph_equivalence_covering_number_cycle_power_of_le (arXiv:2610.15635 (Covering powers of cycles with equivalence subgraphs), Theorem 1.2 (the case n ≤ 2k + 1; the Lean adds n ≥ 2 so that the…): If 2 ≤ n ≤ 2k + 1 then every two vertices of C_n are at cyclic distance at most k, so C_n^k = K_n and eq(C_n^k) = 1.

Proved. simple_graph_equivalence_covering_number_cycle_power_forty_six_three (arXiv:2610.15635): eq(C_46^3) = 4: the smallest case of the open problem (k = 3, n = 46 even, 4 ∤ 46, 46 ≥ 45), where the paper only knows eq ∈ {3, 4, 5}.

Proved. mixed_graph_has_cycle_directed_triangle: Sanity check: the directed triangle 0 → 1 → 2 → 0 has a 3-cycle.

Proved. mixed_graph_has_girth_directed_triangle: Sanity check: the directed triangle has girth 3.

Proved. mixed_graph_is_valid_directed_triangle: Sanity check: the directed triangle is a mixed graph (no loops, no antiparallel arcs).

Formalized. mixed_graph_card_ge_of_is_regular (arXiv:2610.14110 (New mixed cages, bounds and computational results), Theorem 3.1 (the lower bound n(r,z,g) ≥ r + 2z + 1 used for g = 3 in…): Every (r, z)-regular mixed graph on a finite nonempty vertex set has at least r + 2z + 1 vertices.

Formalized. mixed_cage_order_girth_three (arXiv:2610.14110 (New mixed cages, bounds and computational results), Theorem 5.1.): For all r ≥ 1 and z ≥ 1, the minimum order of an (r, z, 3)-mixed cage is n(r, z, 3) = r + 2z + 1.

Formalized. mixed_cage_order_one_girth_three (arXiv:2610.14110): The case g = 3 of the conjecture: n(1, z, 3) = N(z, 3), i.e. 2z + 2 for every z ≥ 1 (both parities agree).

Proved. latin_square_is_add: Sanity check: the addition table of Z/n is a Latin square.

Open. simple_graph_equivalence_covering_number_cycle_power_ge (arXiv:2610.15635 (Covering powers of cycles with equivalence subgraphs), Lemma 2.2 (the lower bound of Theorem 1.2).): For k ≥ 1 and n ≥ 2k + 2, eq(C_n^k) ≥ 1 + ⌈log₂(k + 1)⌉.

Open. simple_graph_equivalence_covering_number_cycle_power_dvd (arXiv:2610.15635 (Covering powers of cycles with equivalence subgraphs), Theorem 2.6.): For k ≥ 1 and q ≥ 2, the k-th power of the cycle on q(k + 1) vertices has equivalence covering number exactly 1 + ⌈log₂(k + 1)⌉.

Open. simple_graph_equivalence_covering_number_cycle_power_two_pow_even_not_dvd (arXiv:2610.15635): Let k + 1 = 2^m with m ≥ 1, and let n ≥ max{2k + 2, k²(k + 2)} be even with k + 1 not dividing n. Then eq(C_n^k) = 2 + ⌈log₂(k + 1)⌉ = m + 2 (the larger of the two values the paper leaves open).

Open. mixed_cage_order_one_le (arXiv:2610.14110 (New mixed cages, bounds and computational results), Theorem 4.2 (with Definition 4.1).): For z ≥ 1 and g ≥ 3, n(1, z, g) ≤ N(z, g), where N(z, g) = z(g − 1) + 2 for even z and (z + 1)(g − 1) for odd z: the circulant M(z, g) on 2m = N(z, g) vertices with arcs of jumps 1..⌈z/2⌉ and m+1..m+⌊z/2⌋ and the antipodal edges is a (1, z, g)-mixed graph.

Open. mixed_cage_order_one_eq (arXiv:2610.14110): For all z ≥ 1 and g ≥ 3, the minimum order of a mixed graph that is 1-regular in edges and z-regular in arcs with girth g is n(1, z, g) = N(z, g), where N(z, g) = z(g − 1) + 2 if z is even and (z + 1)(g − 1) if z is odd.

Open. latin_square_exists_principal_isotope_avoiding (arXiv:2610.20008): If L and L' are any Latin squares of the same order n > 1, there are permutations π, ρ of {1, ..., n} with L'(π(i), ρ(j)) ≠ L(i, j) for all i, j: some principal isotope of L' avoids L.

Open. postage_stamp_number_two_eq (arXiv:2610.19470 (Positivity and negativity for h-bases for n), Theorem 1.1 (classical; cited there).): The classical postage stamp number for three stamps: n_h(2) = ⌊(h² + 6h + 1)/4⌋ for every h ≥ 1.

Open. postage_stamp_number_neg_two_eq_of_mod_four_eq_two (arXiv:2610.19470 (Positivity and negativity for h-bases for n), Theorem 3.3 (with the value of n_h(2) from Theorem 1.1).): If h ≥ 2 and h ≡ 2 (mod 4), then n♭_h(2) = n_h(2) = (h² + 6h)/4.

Open. postage_stamp_number_neg_two_eq_of_mod_four_eq_zero (arXiv:2610.19470 (Positivity and negativity for h-bases for n), Theorem 3.4 (with the value of n_h(2) from Theorem 1.1).): If h ≥ 8 and h ≡ 0 (mod 4), then n♭_h(2) = n_h(2) = (h² + 6h)/4.

Open. postage_stamp_number_three_lt_neg: For four-element sets (k = 3) and every h ≥ 8, sets containing a negative integer do strictly better than nonnegative ones: n♭_h(3) > n_h(3). This reverses the three-element case, where n♭_h(2) ≤ n_h(2) for every h.

Verified in Lean

Launch

TCSlib complete, and 581 declarations from QuantumInfo

The seven TCSlib roots left out on 8 October now compile on our Mathlib, so the 59 declarations they held back came in: among them the Hamming and quantum Hamming bounds, the quantum Singleton bound, list-decoding capacity, BLR soundness, Newman's theorem, Kruskal's optimality and the Johnson-Lindenstrauss lemma. Statements are as the library writes them; only proofs were adapted to our Lean version.

Details

QuantumInfo, the quantum information part of PhysLib, gave 581 declarations, 387 of them theorems, each checked interactively and then built offline with lake build like every other import. The rest of the library stays out for now.

Daily reading

Daily reading: 29 proved, 8 open

The Proposer read 14 new paper(s), listed 27 open problem(s), and stated 5 of them in Lean, and posed 0 more from what the knowledgebase already holds (37 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 29 theorem(s); 8 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. simple_graph_is_color_connecting_edge_set: In a connected edge-coloured graph, the whole edge set is colour-connecting.

Proved. simple_graph_is_color_connecting_mono: A set of edges containing a colour-connecting set is colour-connecting.

Proved. simple_graph_wild_number_le_card_edge_finset: The wild number of a finite connected edge-coloured graph is at most its number of edges.

Proved. simple_graph_wild_number_le_of_is_color_connecting: The wild number is at most the size of any colour-connecting set.

Formalized. wild_number_complete_bipartite_min_proper_coloring (arXiv:2610.22142 (Wild number for spanning trees in complete and complete bipartite graphs), Theorem 4.2. The Lean takes proper colourings…): Let 2 ≤ m ≤ n. For every proper edge colouring of the complete bipartite graph K_{m,n} with n colours (these are exactly its minimum proper edge colourings, since its chromatic index is its maximum degree n), the wild number is n.

Formalized. wild_number_complete_graph_min_proper_coloring_ge (arXiv:2610.22142): Let n ≥ 3 and let ρ be a minimum proper edge colouring of K_n (a proper colouring with χ'(K_n) colours). Then wld_ρ(K_n) ≥ ⌊4⌈n/2⌉/3⌋.

Formalized. wild_number_complete_graph_two_coloring_range (arXiv:2610.22142 (Wild number for spanning trees in complete and complete bipartite graphs), Theorem 3.1 (with 0 ≤ wld automatic in ℕ).): Let n ≥ 4. Every (surjective) 2-edge-colouring of K_n has wild number at most n − 2, and every integer 0 ≤ k ≤ n − 2 is the wild number of some 2-edge-colouring of K_n.

Proved. digraph_edge_count_eq_cut_add: For any vertex set A, the edges of a digraph split into those inside A, from A to its complement, from the complement to A, and inside the complement.

Proved. allocation_is_envy_freeable_of_envy_free: An envy-free allocation is envy-freeable (with zero payments).

Proved. allocation_is_ef1_of_envy_free: An envy-free allocation is EF1.

Formalized. exists_submodular_instance_no_ef1_envy_freeable_allocation_with_bobw_lottery (arXiv:2610.14994 (Best-of-both-worlds fair division via subsidies), Theorem 5.1 (the instance: t = (5, 8, 11, 0), u = (9, 1, 0, 10) on…): There is an instance with three agents and four indivisible goods, each agent having a normalized monotone submodular valuation, in which no complete allocation is both EF1 and envy-freeable, yet some lottery over complete allocations is ex-ante envy-free…

Proved. group_is_complete_mapping_id_zmod_three: In the cyclic group of order 3 the identity map is a complete mapping (x ↦ x² is a bijection in odd order).

Open. wild_number_complete_graph_min_proper_coloring_eq (arXiv:2610.22142): Conjecture: for every n ≥ 3 and every minimum proper edge colouring ρ of K_n, wld_ρ(K_n) = ⌊4⌈n/2⌉/3⌋.

Open. digraph_exists_bipartition_both_cuts_ge_ceil_div_six (arXiv:2610.21381): Let G be a finite loopless digraph with m edges and minimum outdegree at least 2. Then there is a vertex bipartition V = A ⊔ B with e(A, B) ≥ ⌈m/6⌉ and e(B, A) ≥ ⌈m/6⌉.

Open. digraph_exists_bipartition_both_cuts_ge_min_outdegree_bound (arXiv:2610.21381): Conjecture: for every integer d ≥ 2, every finite loopless digraph with m edges and minimum outdegree at least d has a vertex bipartition V = A ⊔ B with both e(A, B) and e(B, A) at least ⌈(d − 1)m / (2(2d − 1))⌉.

Open. poset_nonempty_order_iso_of_card_order_hom_eq_of_connected (arXiv:2610.20802 (Two short proofs of the A^A theorem), Lemma 2.4.): Let X and Y be finite posets (possibly empty). If for every finite nonempty connected poset Z the number of order-preserving maps Z → X equals the number of order-preserving maps Z → Y, then X and Y are order-isomorphic. (A poset is connected when its…

Open. subcubic_three_vertices_max_degree_le_one_exists_rich_neighbor_five_coloring (arXiv:2610.21868): Every finite simple graph of maximum degree at most 3 in which each vertex of degree 3 has at most one neighbour of degree 3 (Δ(G[V_3]) ≤ 1) admits a rich-neighbor 5-edge-colouring.

Open. subcubic_three_vertices_max_degree_le_two_exists_rich_neighbor_five_coloring (arXiv:2610.21868): Question: does every finite simple graph of maximum degree at most 3 in which each vertex of degree 3 has at most two neighbours of degree 3 (Δ(G[V_3]) ≤ 2) admit a rich-neighbor 5-edge-colouring?

Open. subcubic_connected_ne_complete_four_exists_rich_neighbor_five_coloring (arXiv:2610.21868): Conjecture: every finite connected simple graph of maximum degree at most 3 other than K_4 admits a rich-neighbor 5-edge-colouring.

Open. mutual_power_divergence_information_subadditive_of_two_thirds_le (arXiv:2610.21231): Conjecture: for every β ∈ [2/3, 1), the mutual f_β-information (power divergence) is subadditive under Y ← X → Z: Ĩ_{f_β}(X; YZ) ≤ Ĩ_{f_β}(X; Y) + Ĩ_{f_β}(X; Z). (It fails at β = 1/4 and β = 1/2.)

Verified in Lean

Daily reading

Daily reading: 25 proved, 6 open

The Proposer read 25 new paper(s), listed 49 open problem(s), and stated 4 of them in Lean, and posed 3 more from what the knowledgebase already holds (33 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 25 theorem(s); 6 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Formalized. chain_graph_two_independence_number_eq_upper_two_star_forming_number (arXiv:2610.03785 (Upper k-star-forming sets, k-independence, and upper domination), Theorem 6.2 (the case k = 2 of the question whose case…): For every finite chain graph G, the dissociation number β₂(G) equals the upper 2-star-forming number SF₂(G).

Proved. simple_graph_is_hop_inv: The inverse of a hop is a hop (the graph is undirected).

Proved. simple_graph_is_hop_top_iff: In a complete graph, the hops are exactly the derangements (fixed-point-free permutations).

Proved. simple_graph_is_hop_mem_leap_group: Every hop lies in the leap group.

Proved. simple_graph_kayak_paddle_adj_zero_k: In KP(k, m, l), vertex 0 of the first cycle is adjacent to vertex k, the first vertex after it on the connecting path.

Proved. kayak_paddle_leap_group_not_eq_top_or_alternating (arXiv:2610.18463): No: KP(3, 4, 1) (a triangle and a 4-cycle joined by an edge, 7 vertices) admits a hop, but its leap group is neither S_7 nor A_7. Every hop of KP(3,4,1) maps the triangle {0,1,2} to itself, so the leap group fixes that set (it has order 24).

Proved. kayak_paddle_leap_group_ne_of_even_cycle (arXiv:2610.18463): If one of the two cycles of the kayak paddle graph KP(k, m, l) has even length, its leap group is neither the symmetric nor the alternating group on its vertices.

Formalized. mondrian_seven_rects_defect_example (arXiv:2610.19193 (A Bounded-Defect Construction for the Mondrian Art Problem), Example 1.): For (p, q) = (13, 3) (n = 315) the seven rectangles have defect 280: the largest area is 238×60 = 14280 and the smallest 70×200 = 14000.

Proved. mondrian_seven_rects_is_partition: For positive integers p, q of the same parity with q < p < 7q, the seven rectangles R₁, …, R₇ have positive integer sides, partition the 105q × 105q square, and are pairwise noncongruent.

Proved. mondrian_seven_rects_defect: With E = p² − 19q², the defect of the seven-rectangle dissection is 140|E|.

Formalized. mondrian_seven_rects_partition_defect (arXiv:2610.19193 (A Bounded-Defect Construction for the Mondrian Art Problem), Proposition 1 (with the partition property stated before it).): Let p, q be positive integers of the same parity with q < p < 7q, n = 105q and E = p² − 19q². The seven rectangles R₁, …, R₇ of (2)–(3) have positive integer sides, partition the n×n square and are pairwise noncongruent, and the defect of this partition is…

Formalized. mondrian_seven_rects_defect_ge (arXiv:2610.19193): Within the seven-rectangle family, the defect is always at least 280 (in particular never 0): if p, q are odd then p² − 19q² ≡ 2 (mod 4), and if both are even then 4 ∣ E with E ≠ 0, so |E| ≥ 2.

Open. simple_graph_leap_group_card_odd_iff (arXiv:2610.18463 (The group generated by a graph's derangements), Proposition 2.16. 'Disjoint union of odd-length cycles' is stated as:…): Let G be a finite simple graph that admits a hop. Then the leap group Λ₁(G) has odd order if and only if G is a disjoint union of cycles of odd length, i.e. G is 2-regular and every connected component has an odd number of vertices.

Open. kayak_paddle_leap_group_of_odd_cycles (arXiv:2610.18463): If both cycles of the kayak paddle graph KP(k, m, l) have odd length, its leap group is the symmetric or the alternating group on its N = k + m + l − 1 vertices.

Open. mondrian_liminf_min_defect_seven_eq (arXiv:2610.19193): Conjecture: for Mondrian partitions of the n×n square into seven pairwise noncongruent integer rectangles, liminf_{n→∞} M₇(n) = 280, i.e. no sequence of seven-rectangle partitions has defect eventually below 280.

Open. mondrian_no_perfect_partition (arXiv:2610.19193): No square has a perfect Mondrian partition: there is no partition of an n×n square into at least two pairwise noncongruent integer rectangles all of the same area.

Open. zero_sum_w_two (arXiv:2610.19965): For every r ≥ 1, w_r(2) = 2r + 1.

Open. zero_sum_w_eq_question (arXiv:2610.19965): Does w_r(n) = 2r(n−1) + 1 hold for every r ≥ 1 and n > 1?

Verified in Lean

Daily reading

Daily reading: 30 proved, 9 open

The Proposer read 30 new paper(s), listed 65 open problem(s), and stated 5 of them in Lean, and posed 2 more from what the knowledgebase already holds (39 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 30 theorem(s); 9 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. simple_graph_is_k_independent_empty: The empty vertex set is k-independent, for every k.

Proved. simple_graph_is_k_star_forming_univ: In a finite graph the whole vertex set is k-star-forming, for every k (there is no vertex outside it).

Proved. simple_graph_is_k_star_forming_mono: Every superset of a k-star-forming set is k-star-forming.

Proved. simple_graph_k_independent_is_k_independent_one_iff: A vertex set is 1-independent exactly when it is independent: no two of its vertices are adjacent.

Formalized. simple_graph_k_independence_number_le_upper_k_star_forming_number (arXiv:2610.03785 (Upper k-star-forming sets, k-independence, and upper domination), Theorem 2.3 (a theorem the paper cites from earlier…): For every finite graph G and every k, β_k(G) ≤ SF_k(G).

Formalized. simple_graph_k_independence_number_eq_card_of_degree_lt (arXiv:2610.03785 (Upper k-star-forming sets, k-independence, and upper domination), Proposition 3.1.): If every vertex of the finite graph G has degree less than k, then β_k(G) = SF_k(G) = |V(G)|.

Formalized. complete_bipartite_k_independence_number_eq_upper_k_star_forming_number (arXiv:2610.03785 (Upper k-star-forming sets, k-independence, and upper domination), Theorem 3.2.): For k ≥ 2 and a, b ≥ k, the complete bipartite graph K_{a,b} has β_k(K_{a,b}) = SF_k(K_{a,b}) = max{a, b, 2k − 2}.

Formalized. exists_triangle_free_k_independence_number_lt_upper_k_star_forming_number (arXiv:2610.03785 (Upper k-star-forming sets, k-independence, and upper domination), Proposition 7.1 (the explicit graph with edges 02, 08,…): There is a triangle-free graph H on 10 vertices with β_4(H) = 7 < 8 = SF_4(H).

Proved. simple_graph_is_chain_graph_complete_bipartite: The complete bipartite graph K_{a,b} is a chain graph (all vertices on one side have the same neighbourhood).

Formalized. perm_distinct_abs_diffs_ascent_count_sub_half_unbounded (arXiv:2610.15860 (A note on the Slater–Vélez permutation), Lemma 1.): For any permutation (a_n) of the positive integers whose consecutive absolute differences are pairwise distinct, A(X) − X/2 is unbounded above, where A(X) counts the v ≤ X whose successor in the sequence is larger than v.

Proved. greedy_distinct_diff_seq_one_two_two: The greedy sequence started at 1, 2 has third term 4 (3 is refused since |3 − 2| = 1 is already a difference).

Formalized. exists_submodule_finrank_inf_eq_of_finrank_inf (arXiv:2610.17631 (Complete mappings stabilizing Sylow 2-subgroups in central extensions), Lemma 4.1 (stated here for a finite-dimensional…): Let E, F be k-dimensional subspaces of a finite-dimensional vector space with dim(E ∩ F) = k − j, 0 ≤ j ≤ k. Then some k-dimensional subspace W has dim(E ∩ W) = dim(W ∩ F) = k − j.

Open. chain_graph_k_independence_number_eq_upper_k_star_forming_number (arXiv:2610.03785): Conjecture: for every finite chain graph G and every k ≥ 3, β_k(G) = SF_k(G).

Open. greedy_distinct_diff_seq_surjective (arXiv:2610.15860): Conjecture: for any distinct positive integers a₁, a₂, the greedy sequence with distinct consecutive differences started at a₁, a₂ takes every positive integer as a value (it is a permutation of the positive integers).

Open. greedy_distinct_diff_seq_one_two_surjective (arXiv:2610.15860): The greedy sequence with distinct consecutive differences started at 1, 2 takes every positive integer as a value.

Open. simple_graph_triangle_cover_number_le_two_mul_packing_number (arXiv:2610.15539): Tuza's conjecture: every finite simple graph G satisfies τ(G) ≤ 2ν(G).

Open. simple_graph_triangle_cover_number_le_two_mul_packing_number_of_max_degree_le_eleven (arXiv:2610.15269): Tuza's conjecture for graphs of maximum degree at most 11: every finite simple graph G with Δ(G) ≤ 11 satisfies τ(G) ≤ 2ν(G).

Open. simple_graph_triangle_cover_number_le_two_mul_packing_number_of_max_degree_le_ten (arXiv:2610.15539): Every finite simple graph G with maximum degree at most 10 satisfies τ(G) ≤ 2ν(G).

Open. tuza_iff_tuza_of_edge_count_ge (arXiv:2610.15539 (Tuza's conjecture under an edge-count condition), Proposition 1.2.): Tuza's conjecture holds for all finite graphs if and only if, for every c ≥ 0, it holds for all graphs on n ≥ n₀(c) vertices with at least n²/4 − cn edges.

Open. chain_graph_two_independence_number_eq_upper_two_star_forming_number (arXiv:2610.03785 (Upper k-star-forming sets, k-independence, and upper domination), Theorem 6.2 (the case k = 2 of the question whose case…): For every finite chain graph G, the dissociation number β₂(G) equals the upper 2-star-forming number SF₂(G).

Open. davenport_constant_int_box_two_eq (arXiv:2610.16562): Conjecture: for every m ≥ 2, D(⟦−m, m⟧²) = 4m² − q(m), and every minimal zero-sum sequence over ⟦−m, m⟧² of that maximal length is the image of S(m) under one of the eight symmetries of the square.

Verified in Lean

Daily reading

Daily reading: 36 proved, 3 open

The Proposer read 12 new paper(s), listed 26 open problem(s), and stated 2 of them in Lean, and posed 4 more from what the knowledgebase already holds (38 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 36 theorem(s); 3 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. card_le_twenty_nine_of_regular_eight_triangle_distinct (arXiv:2610.08478): Every 8-regular triangle-distinct graph has at most 29 vertices: its triangle-degrees are pairwise distinct and each is at most C(8, 2) = 28.

Proved. digraph_invert_invert: Inverting the same vertex set twice gives back the digraph.

Proved. digraph_invert_is_tournament: Inverting a vertex set of a tournament gives a tournament.

Proved. digraph_dijoin_is_tournament: The dijoin of two tournaments is a tournament.

Proved. digraph_inversion_number_eq_zero_of_is_acyclic: An acyclic digraph has inversion number 0.

Proved. digraph_is_acyclic_of_rank: A digraph with a numbering of the vertices that increases along every arc is acyclic.

Proved. digraph_directed_triangle_not_acyclic: The directed triangle 0 → 1 → 2 → 0 is not acyclic.

Proved. digraph_is_acyclic_dijoin: The dijoin of two acyclic digraphs is acyclic: a directed walk never comes back from the D₂ side to the D₁ side.

Proved. tournament_exists_decycling_family: Every finite tournament can be made acyclic by finitely many inversions (so its inversion number is attained).

Formalized. tournament_dijoin_inversion_number_le_add (arXiv:2610.13798): For finite tournaments D₁, D₂: inv(D₁ → D₂) ≤ inv(D₁) + inv(D₂), by inverting an optimal family of D₁ and then one of D₂.

Formalized. tournament_dijoin_inversion_number_le_symm_add_two (arXiv:2610.13798): For finite tournaments D₁, D₂: inv(D₁ → D₂) ≤ inv(D₂ → D₁) + 2 (with D₁, D₂ swapped this gives |inv(D₁ → D₂) − inv(D₂ → D₁)| ≤ 2): invert V(D₁) and V(D₂) first, which turns D₁ → D₂ into the reverse of D₂ → D₁, and a family decycles a digraph iff it decycles…

Formalized. tournament_inversion_number_le_dijoin (arXiv:2610.13798): For finite tournaments D₁, D₂: inv(D₁ → D₂) ≥ max(inv(D₁), inv(D₂)), since D₁ and D₂ are induced subtournaments of the dijoin and a decycling family restricts to one.

Open. tournament_dijoin_inversion_number_symm (arXiv:2610.13798): Conjecture (arXiv:2610.13798, Conjecture 4.1): for all finite tournaments D₁, D₂, inv(D₁ → D₂) = inv(D₂ → D₁).

Open. tournament_inversion_number_le_min_rank_add_one (arXiv:2610.13798 (On the failure of additivity of inversion numbers for dijoins), inequality (1.2), quoting an earlier paper.): For a finite tournament D: inv(D) ≤ tmr(D) + 1.

Open. poset_height_one_reconstructible (arXiv:2610.12690): Open problem (arXiv:2610.12690, Sec. 9, from an open problem in a monograph on ordered sets): every finite ordered set of height one with at least four elements is reconstructible from its deck: if P and Q are finite ordered sets of height one, |P| ≥ 4, and…

Verified in Lean

Daily reading

Daily reading: 39 proved, 2 open

The Proposer read 0 new paper(s), listed 0 open problem(s), and stated 1 of them in Lean, and posed 9 more from what the knowledgebase already holds (41 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 39 theorem(s); 2 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Formalized. cycle_wrap_set_compl_eq_Icc_and_sum (arXiv:2610.08889 (Consecutive Cycle Sums), Lemma 1. Wrap sets and standard intervals are written out as Finset.Icc unions; the empty…): For 1 ≤ i < j ≤ n, the complement in {1,…,n} of the wrap set W = {j,…,n} ∪ {1,…,i} of the n-cycle is the standard interval {i+1,…,j−1} (empty when j = i+1), and its element sum is T_n − W*, where T_n = n(n+1)/2 and W* is the element sum of W.

Proved. simple_graph_sum_choose_degree_two_le_choose_card_edgeFinset: In any finite simple graph, the number of pairs of edges sharing a vertex, ∑_v C(d(v),2), is at most C(|E|,2): two distinct edges share at most one vertex.

Formalized. simple_graph_is_tree_card_sub_two_le_sum_choose_degree_le_choose (arXiv:2610.12109 (Sharp linear Turán remainders for cycle and edge extremality), Lemma 2.5, the two inequalities (the equality cases are…): For a tree T on k ≥ 2 vertices, σ₂(T) = ∑_v C(d_T(v),2) satisfies k − 2 ≤ σ₂(T) ≤ C(k−1,2).

Formalized. simple_graph_is_tree_sum_choose_degree_eq_iff_path (arXiv:2610.12109 (Sharp linear Turán remainders for cycle and edge extremality), Lemma 2.5, equality case of the lower bound.): For a tree T on k ≥ 2 vertices, σ₂(T) = k − 2 if and only if T is isomorphic to the path P_k.

Formalized. simple_graph_is_tree_sum_choose_degree_eq_iff_star (arXiv:2610.12109 (Sharp linear Turán remainders for cycle and edge extremality), Lemma 2.5, equality case of the upper bound. The star…): For a tree T on k ≥ 2 vertices, σ₂(T) = C(k−1,2) if and only if T is a star K_{1,k−1}: some vertex r is adjacent to every other vertex and there are no other edges (T = starGraph r).

Proved. word_occurrence_coverage_subset: Every coverage Cov_w(x) is a set of positions of w, i.e. contained in [1, |w|].

Proved. word_occurrence_coverage_ababa_aba: The paper's example: in w = ababa the factor aba occurs at positions 1 and 3, so Cov_w(aba) = [1,3] ∪ [3,5] = [1,5] (letters a, b encoded as 0, 1).

Proved. word_occurrence_coverage_nonempty: The coverage of a nonempty factor of w is nonempty.

Proved. is_string_attractor_Icc: The set of all positions [1, |w|] is a string attractor of w.

Proved. not_is_string_attractor_empty: The empty set is not a string attractor of a nonempty word.

Proved. word_min_factor_coverages_subset: Every inclusion-minimal factor coverage is a factor coverage: H_min(w) ⊆ H(w).

Proved. exists_min_factor_coverage_subset: Every factor coverage D ∈ H(w) contains some inclusion-minimal factor coverage C ∈ H_min(w).

Open. simple_graph_is_tree_stacking_number_eq_tree_estimate (arXiv:2610.09650): Conjecture: for every finite tree T, stack(T) = estim(T).

Open. simple_graph_is_tree_tree_estimate_le_stacking_number (arXiv:2610.09650): For every finite tree T, stack(T) ≥ estim(T).

Verified in Lean

Weekly report

Week of 5–11 Oct 2026: Pairwise fairness limits for chores, exact poset query complexity, and refuted conjectures

Fair division saw the resolution of several open questions: dividing chores fairly among three people using pairwise shares is impossible in general, the exact welfare guarantee from maximizing product utilities for two people was pinned down, and a graph-connectivity conjecture failed. In algorithms, finding minimal elements in two-chain posets was proved to require strictly more comparisons than conjectured. In mechanism design, counterexamples disproved a dictatorship conjecture for facility location with scaling and refuted two published claims in agency and matching.

Details

Proved. PMMS allocations of chores can fail to exist for three agents

Proved. Price of maximum Nash welfare for two agents equals 27/23

Disproved. Graph linkedness conjecture for two-agent price of connectivity is false

Disproved. Deterministic query complexity for minimal elements of width-two posets is 2n − 3

Disproved. Dictatorship conjecture for facility location with scaling fails for known scaling functions

Lean accepted 372 new theorems this week, drawn from 34 papers and 24 topics. Of these, 47 are new here, 5 correct a published claim, 94 formalize a published result and 226 are helper lemmas. 36 statements posed this week are not yet proved. The reading covered 51 new papers. This report covers 16 results in 7 fields.

Daily reading

Daily reading: 42 proved, 5 open

The Proposer read 0 new paper(s), listed 0 open problem(s), and stated 2 of them in Lean, and posed 8 more from what the knowledgebase already holds (42 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 42 theorem(s); 5 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. finset_contains_trace_of_subset: Containing H as a trace is monotone: if F ⊆ F' and F contains H as a trace, so does F'.

Proved. finset_contains_trace_powerset_univ: The full power set of {0,…,h−1} contains every set system on {0,…,h−1} as a trace.

Proved. not_finset_contains_trace_empty: The empty set system contains no nonempty set system as a trace.

Formalized. finset_vc_dim_le_of_not_contains_trace (arXiv:2610.08151 (Polynomial-in-r bounds for forbidden traces of uniform hypergraphs), Proposition 1. VCdim is Mathlib's Finset.vcDim…): Let H be a set system on an h-element ground set, h ≥ 1. If a finite set system F is trace-H-free, then VCdim(F) ≤ h − 1.

Proved. simple_graph_triangle_degree_complete_fin_four: Every vertex of K_4 lies in exactly 3 triangles.

Proved. simple_graph_triangle_degree_bot: In the empty graph every vertex has triangle-degree 0.

Proved. simple_graph_triangle_degree_le_choose: td_G(v) ≤ C(deg v, 2).

Proved. not_triangle_distinct_complete_fin_four: K_4 is not triangle-distinct.

Formalized. simple_graph_triangle_degree_compl (arXiv:2610.08478 (Regular K_3-Irregular Graphs of Every Regularity at Least Nine), Lemma 2.1 (first part); stated over ℚ so that N r / 2…): Let G be an r-regular simple graph on N vertices. For every vertex v, td_{Ḡ}(v) = C(N − r − 1, 2) − N r / 2 + r² − td_G(v), where Ḡ is the complement of G.

Formalized. simple_graph_compl_regular_triangle_distinct (arXiv:2610.08478 (Regular K_3-Irregular Graphs of Every Regularity at Least Nine), Lemma 2.1 (consequence); follows from the complement…): If G is an r-regular triangle-distinct graph on N vertices, its complement is an (N − r − 1)-regular triangle-distinct graph.

Proved. simple_graph_power_one: G^1 = G.

Proved. simple_graph_is_well_covered_bot: The edgeless graph on a finite vertex set is well-covered (its only maximal independent set is the whole vertex set).

Open. exists_regular_eight_triangle_distinct (arXiv:2610.08478): There exists a finite simple 8-regular triangle-distinct graph: a graph on some finite vertex set {0,…,n−1} in which every vertex has degree 8 and the triangle-degrees of the vertices are pairwise distinct.

Open. card_le_twenty_nine_of_regular_eight_triangle_distinct (arXiv:2610.08478): Every 8-regular triangle-distinct graph has at most 29 vertices: its triangle-degrees are pairwise distinct and each is at most C(8, 2) = 28.

Open. forall_power_well_covered_diam_le_three_and_twin_free_bounded (arXiv:2610.09300): Let G be a finite simple connected graph with G^d well-covered for every d ≥ 1. Then diam(G) ≤ 3; moreover, there are only finitely many such graphs (up to isomorphism) of diameter exactly 3 without true twins (adjacent u, v with N[u] = N[v]) — stated as a…

Open. diam_le_three_of_forall_power_well_covered (arXiv:2610.09300): If G is a finite simple connected graph such that G^d is well-covered for all d ≥ 1, then diam(G) ≤ 3.

Open. exists_nonempty_regular_eight_triangle_distinct (arXiv:2610.08478): There exists a finite simple 8-regular triangle-distinct graph with at least one vertex: a graph on {0,…,n−1}, n ≥ 1, in which every vertex has degree 8 and the triangle-degrees are pairwise distinct.

Verified in Lean

Launch

2634 declarations from PhysLib, each compiled here

PhysLib, the Lean community's physics library, came in through aftd import under the same checks as TCSlib, EconCSLib and CSLib: each declaration is checked interactively and then built offline with lake build, and one that does not compile leaves no trace. Each carries the "imported Lean" badge, and its page names the authors, links the file at the commit we read, and gives the license (Apache-2.0).

Details

It gave 2634 declarations, 1604 of them theorems, in nine physics topics, four of them new: quantum field theory, condensed matter, fluid dynamics and cosmology. Among them: for a densely defined closable operator on a Hilbert space, the double adjoint is the closure; the Levi-Civita symbol of a permutation is its sign; the potential energy of the simple pendulum lies between 0 and 2mgℓ and vanishes exactly at the bottom of the swing; black-body spectral radiance vanishes at absolute zero; the quantized wavenumbers of a tight-binding chain lie in the Brillouin zone; the figure of merit of a thermoelectric material; the anomaly-cancellation conditions of the Standard Model with right-handed neutrinos and of the MSSM, with results on their solutions; and the Lorentz vectors and tensors, Weyl fermions, Euclidean group and calculus on space and time these rest on.

These are PhysLib's own formalizations, not new results. PhysLib is written for Lean v4.34.1, one version newer than ours: 2581 declarations compiled here unchanged and 57 needed a change, shown on each page.

This was a single pass over 10,282 declarations, stopped at about nine tenths. Of those tried and left out, 5041 do not compile against our Mathlib, 1448 depend on something that does not, and 268 use syntax extensions an import cannot carry. Five around the Minkowski matrix are left out because their module would take the place of a definition we already had. QuantumInfo, the library's quantum information half, is not imported yet. 1277 imported declarations have no docstring upstream, so their description is only the name for now.

Declarations

Daily reading

Daily reading: 27 proved, 5 open

The Proposer read 1 new paper(s), listed 2 open problem(s), and stated 2 of them in Lean, and posed 0 more from what the knowledgebase already holds (32 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 27 theorem(s); 5 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. graph_wasserstein_dist_le_of_plan: The Wasserstein distance is at most the cost of any transport plan.

Proved. graph_wasserstein_dist_ge_of_lipschitz: Weak Kantorovich duality: for probability measures μ, ν and any function f with f(a) − f(b) ≤ d(a, b), Σ f μ − Σ f ν ≤ W(μ, ν).

Proved. graph_idle_measure_sum_eq_one: μ_x^α is a probability measure (total mass 1) when x has a neighbour.

Proved. graph_wasserstein_dist_self_eq_zero: The Wasserstein distance from a nonnegative measure to itself is 0.

Proved. graph_edge_connectivity_le_minDegree: Edge-connectivity is at most the minimum degree, κ'(G) ≤ δ(G) (for at least two vertices).

Proved. one_le_graph_edge_connectivity_of_connected: A connected graph on at least two vertices has edge-connectivity at least 1.

Formalized. ediam_le_two_of_card_sub_one_le_two_mul_minDegree (arXiv:2610.10559 (Edge-Connectivity versus Lin--Lu--Yau Curvature), Lemma 4.1; diameter as Mathlib's extended diameter ediam, so the…): A graph on n vertices with minimum degree δ(G) ≥ (n − 1)/2 has diameter at most 2 (in particular it is connected).

Proved. graph_wasserstein_dist_comm: The Wasserstein distance is symmetric: W(μ, ν) = W(ν, μ).

Proved. graph_lly_curvature_comm: The Lin–Lu–Yau curvature is symmetric: κ_LLY(x, y) = κ_LLY(y, x).

Proved. graph_lly_curvature_eq_of_ollivier_eq: If κ_α(x, y) = (1 − α)c for all α in an interval (a, 1), then κ_LLY(x, y) = c.

Proved. graph_idle_measure_nonneg: μ_x^α is nonnegative for 0 ≤ α ≤ 1.

Proved. graph_lly_curvature_of_leaf: If x is a leaf of a finite connected graph whose only neighbour is y, then κ_LLY(x, y) = 2/d_y.

Open. edge_connectivity_ge_half_scale_mul_min_degree_mul_lly_lower_bound (arXiv:2610.10559): For every d ≥ 3 and every finite connected graph G of diameter at least d: κ'(G) ≥ (d/2) · δ(G) · κ_LLY^(d)(G).

Open. edge_connectivity_ge_three_halves_min_degree_mul_lly_lower_bound_three (arXiv:2610.10559): Every finite connected graph of diameter at least 3 has κ'(G) ≥ (3/2) · δ(G) · κ_LLY^(3)(G).

Open. edge_connectivity_ge_min_degree_mul_lly_lower_bound_two (arXiv:2610.10559 (Edge-Connectivity versus Lin--Lu--Yau Curvature), Theorem 1.1 (for finite graphs).): Every finite connected graph of diameter 2 has κ'(G) ≥ δ(G) · κ_LLY^(2)(G).

Open. lly_curvature_pos_of_edge_connectivity_ge (arXiv:2610.10559): If a finite connected graph on n vertices has edge-connectivity κ'(G) ≥ 2n/3 − 1, then every edge xy has κ_LLY(x, y) > 0.

Open. lly_positive_edge_connectivity_threshold_five (arXiv:2610.10559): c(5) = 2: every connected graph on 5 vertices with edge-connectivity at least 2 has positive Lin–Lu–Yau curvature on every edge, while the path-like tree with a vertex of degree 3 (edges 01, 03, 04, 12) has κ' = 1 and κ_LLY(0,1) = −1/3. So c(5) < ⌈2·5/3 − 1⌉…

Verified in Lean

Daily reading

Daily reading: 38 proved, 8 open

The Proposer read 30 new paper(s), listed 58 open problem(s), and stated 4 of them in Lean, and posed 0 more from what the knowledgebase already holds (41 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 38 theorem(s); 8 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. is_condorcet_winning_set_nonempty: In an election with at least one voter and at least one candidate, every Condorcet winning set is nonempty.

Proved. graph_matching_num_two_mul: Twice the matching number of G equals the supremum of twice the edge cardinalities of all matchings in G.

Proved. matching_two_mul_encard_le_card: For any matching M in a finite graph G, twice its edge cardinality is at most the number of vertices.

Proved. droop_valid_assignment_real_voters_card: For any Droop-valid Monroe assignment π of n voters to committee size k, the number of voters assigned to a real candidate (not none) is exactly n - n / (k + 1).

Formalized. single_peaked_median_is_condorcet_winner (Duncan Black, 'On the Rationale of Group Decision-making', Journal of Political Economy 56(1):23-34 (1948)): Under single-peaked preferences with asymmetric preferences, if an alternative m satisfies that strictly more than half of the voters have peak at least m and strictly more than half have peak at most m, then m is a Condorcet winner.

Proved. finset_has_no_unique_sums_univ_zmod_three: The whole of Z/3Z has no unique sums.

Proved. not_finset_has_no_unique_sums_zmod_five_pair: The set {0, 1} in Z/5Z has unique sums (0 + 0 has only one representation).

Proved. finset_is_midpoint_balanced_univ_zmod_three: Z/3Z is balanced: each element is the midpoint of the other two.

Formalized. finset_has_no_unique_sums_sumset_of_balanced (arXiv:2610.09349 (A near-quadratic lower bound for sets with no unique sums), Lemma A.2. Stated in any abelian group (the paper: F_p); B +…): If B is a nonempty balanced finite subset of an abelian group, then its sumset B + B has no unique sums.

Formalized. finset_is_midpoint_balanced_of_has_no_unique_sums (arXiv:2610.09349 (A near-quadratic lower bound for sets with no unique sums), Lemma B.1 (first statement: every NUS set in odd…): For an odd prime p, every subset of Z/pZ with no unique sums is balanced.

Formalized. finset_has_no_unique_sums_univ_zmod (arXiv:2610.09349 (A near-quadratic lower bound for sets with no unique sums), Sec. 1 (F_p itself is NUS: each s has (p+1)/2 ≥ 2 unordered…): For an odd prime p, the whole of Z/pZ has no unique sums.

Formalized. cycle_labeling_complete_range_five (arXiv:2610.08889 (Consecutive Cycle Sums), Sec. 1 (the 5-cycle is complete).): The 5-cycle labelled 1, 2, 3, 4, 5 in order is complete.

Open. finset_card_ge_one_add_clog_of_midpoint_balanced (arXiv:2610.09349): For a prime p, every nonempty balanced subset B of Z/pZ has |B| ≥ 1 + ⌈log₂ p⌉.

Open. nus_min_card_ge_one_add_clog (arXiv:2610.09349): For an odd prime p, m(p) ≥ 1 + ⌈log₂ p⌉: every subset of Z/pZ with no unique sums has at least 1 + ⌈log₂ p⌉ elements.

Open. nus_min_card_ge_c_log_sq (arXiv:2610.09349): Open question: there is an absolute constant c > 0 such that m(p) ≥ c (log p)² for all sufficiently large primes p.

Open. not_cycle_labeling_complete_range_of_ge_eight (arXiv:2610.08889 (Consecutive Cycle Sums), Theorem 1. The paper's proof uses explicit prime-counting bounds for n ≥ 599 and a computer…): For every n ≥ 8, the n-cycle labelled 1, …, n in order is not complete.

Open. exists_cycle_labeling_complete: Conjecture: for every n there is a placement of 1, …, n around the n-cycle that is complete.

Open. simple_graph_local_avg_dominating_order_stem_le (arXiv:2610.11446): Conjecture: if G is a graph of order n ≥ 2 without isolated vertices and v is an l-stem with l ≥ 1, then avd_v(G) ≤ (4n+1)/6, with equality if and only if l = 1 and G is star-like.

Open. simple_graph_local_avg_dominating_order_lt_of_two_le_stem (arXiv:2610.11446): If G is a graph of order n ≥ 2 without isolated vertices and v is an l-stem with l ≥ 2, then avd_v(G) < (4n+1)/6.

Open. simple_graph_local_avg_dominating_order_le_two_mul_add_one_div_three (arXiv:2610.11446): Conjecture: if the finite graph G of order n has no isolated vertices, then for every vertex v the dominating sets containing v have average size at most (2n+1)/3.

Verified in Lean

Daily reading

Daily reading: 27 proved, 4 open

The Proposer read 10 new paper(s), listed 16 open problem(s), and stated 4 of them in Lean, and posed 1 more from what the knowledgebase already holds (13 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 27 theorem(s); 4 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Formalized. matrix_det_eq_permanent_of_char_two (Valiant 1979, 'Completeness classes in algebra', STOC 1979, Sec. 3): For every commutative ring R of characteristic 2 and any square matrix M over R indexed by a finite type n with decidable equality, the determinant of M equals the permanent of M.

Proved. matrix_permanent_mul_of_char_two: For every commutative ring R of characteristic 2 and any square matrices A and B over R indexed by a finite type n with decidable equality, the permanent of the matrix product A * B equals the product of the permanent of A and the permanent of B.

Formalized. mv_polynomial_eq_zero_of_eval_eq_zero_on_finset (Shpilka & Yehudayoff 2010, Lemma 2.1; Schwartz 1980; Zippel 1979): Let R be an integral domain and p be a multivariate polynomial in n variables over R. If S is a finite subset of R with cardinality strictly greater than the total degree of p, and p evaluates to zero at every point of the product S^n, then p is the zero…

Proved. matrix3_tensor_rank_not_le_iff_ge_succ: For any commutative semiring R and any natural number r, the tensor rank of 3x3 matrix multiplication over R is not at most r if and only if every valid triad decomposition of 3x3 matrix multiplication over R has length at least r + 1.

Formalized. condorcet_paradox (Condorcet 1785, Essai sur l'application de l'analyse à la probabilité des décisions rendues à la pluralité des voix): The Condorcet paradox: there exists a preference profile of 3 voters over 3 alternatives, where each voter has a strict linear order of preferences, such that no alternative is a Condorcet winner.

Proved. is_condorcet_winning_set_mono: Condorcet winning sets are upward closed: if S is a Condorcet winning set and S is a subset of T, then T is also a Condorcet winning set.

Proved. condorcet_dimension_le_two_of_le_three: Every election with at most three candidates and at least one voter has a Condorcet winning set of at most two candidates.

Proved. matrix_permanent_from_blocks_zero_one_two: For any commutative semiring R and any square matrices A over m and D over n, the permanent of the block matrix with A in the top-left, 0 in the top-right, C in the bottom-left, and D in the bottom-right equals the permanent of A times the permanent of D.

Proved. droop_monroe_satisfies_droop_jr: Every committee W of size k selected by the Droop Monroe rule satisfies Droop-JR: every voter group S of size strictly greater than n/(k+1) that jointly approves some candidate contains a voter who approves at least one candidate in W.

Proved. matrix_permanent_succ_row_zero: For any commutative semiring R, any natural number n, and any square matrix A of size (n+1) x (n+1) over R, the permanent of A equals the sum over all columns j : Fin (n+1) of A 0 j times the permanent of the submatrix obtained by deleting row 0 and column…

Proved. matrix_permanent_succ_row: For any commutative semiring R, any natural number n, any square matrix A of size (n+1) x (n+1) over R, and any row index i : Fin (n+1), the permanent of A equals the sum over all columns j : Fin (n+1) of A i j times the permanent of the submatrix obtained…

Proved. card_piFinset_filter_apply_mem: For any type α with decidable equality, any positive integer n, any finite subset S of α, any subset T of α with T ⊆ S, and any coordinate index i₀ : Fin n, the cardinality of the set of points in the product S^n whose i₀-th coordinate belongs to T equals…

Open. matroid_sbo_common_list_chromatic_eq (arXiv:2610.07318): Open question (arXiv:2610.07318, Question 3.2): for any two loopless strongly base-orderable matroids M₁, M₂ on a common finite ground set, the common list chromatic number equals the common chromatic number, χ_ℓ(M₁, M₂) = χ(M₁, M₂).

Open. matroid_common_list_chromatic_le_add_const (arXiv:2610.07318): Open question (arXiv:2610.07318, Question 3.1): there is an absolute constant C such that χ_ℓ(M₁, M₂) ≤ χ(M₁, M₂) + C for every pair of loopless matroids on a common finite ground set. (The paper also asks whether C = 1 works; its Theorem 1.1,…

Open. tarski_query_complexity_five_le_log_sq (arXiv:2610.07055): Open question (arXiv:2610.07055, the case k = 5 of the gap left by its Theorems 2 and 3): Tarski(n, 5) = O(log² n), i.e. there is a constant C such that for every n ≥ 2 some deterministic query algorithm finds a fixed point of every monotone map [n]^5 →…

Open. matroid_symmetry_gap_lt_eight_seventeenths (arXiv:2610.07387): Open question (arXiv:2610.07387, after Corollary 3.4 and Appendix A): there is a strongly symmetric instance of non-negative submodular maximization over a matroid — a non-negative submodular f on a finite ground set, a matroid M and a permutation group G…

Verified in Lean

Launch

2934 declarations from EconCSLib and CSLib, each compiled here

Two more Lean libraries came in through aftd import, under the same checks as TCSlib: each declaration is checked interactively and then built offline with lake build, and one that does not compile leaves no trace. Each carries the "imported Lean" badge, and its page names the authors, links the file at the commit we read, and gives the license.

Details

EconCSLib, a library of economics and computation, gave 1822 declarations, 999 of them theorems, in nine topics. Among them: von Neumann's minimax theorem for matrix games and the existence of a mixed Nash equilibrium; weak and strong duality of linear programming; Gale–Shapley, with stable matchings closed under join and meet; the revelation principle for Bayesian mechanisms; cut-and-choose is envy-free, and EFX implies EF1; the core of a coalitional game lies within its imputations.

CSLib, the Lean community's computer science library, gave 1112 declarations, 566 of them theorems, in nine topics. Among them: Shannon's bound that a perfectly secret scheme needs at least as many keys as messages, and the perfect secrecy of the one-time pad; reconstruction in Shamir's secret sharing; a URM program halts exactly when it is normalizable; the basics of PAC learning (version spaces, shattering, VC dimension, and how sample complexity moves with ε, δ and the class); and the algebra of formal languages and automata.

These are the libraries' own formalizations, not new results. Our part is compiling them again against our Lean and Mathlib: 2693 declarations compiled unchanged and 252 needed a change, shown on each page. To do this, imports now follow the exact dependency graph Lean reports, carry a library's local notation with the declarations that use it, join mutual definitions, and move a name we already had under the library's namespace.

This took five passes, each fixing the causes the one before had found. 347 of the 2169 declarations we asked of EconCSLib and 1881 of the 2993 of CSLib still do not compile here. Most depend on something that does not; many of the rest pass our interactive check but not the offline build, because they lean on an instance or a grind lemma their module does not import. They are left out. About 900 imported declarations have no docstring upstream, so their description is only the name for now.

Declarations

Launch

853 declarations from TCSlib, each compiled here

The knowledgebase now takes Lean written elsewhere. aftd import reads a declaration from another repository at a fixed commit and compiles it here under the same checks as a proof found by the Prover: first the interactive check, then an offline lake build. A declaration that does not compile leaves no trace. Each imported declaration carries an "imported Lean" badge, and its page names the authors, links the file at that commit, and gives the license.

Details

The first source is TCSlib, a Lean library of theoretical computer science. We asked for 44 of its results and took each together with everything it depends on: 853 declarations in 13 topics, 509 of them theorems. 33 of the 44 results compiled here.

Boolean function analysis: Parseval and Plancherel, Friedgut's junta theorem, Bonami's lemma, Arrow's theorem, BLR completeness, and degree at most decision-tree depth.

Communication complexity: the fooling-set, log-rank and discrepancy lower bounds, the exact complexities of equality, disjointness and one-way indexing, and Pinsker's inequality.

Codes: the Singleton, Gilbert–Varshamov and binary Johnson bounds.

Online learning: the halving, weighted-majority and Hedge bounds, the minimax theorems, and no-regret play converging to coarse correlated equilibria.

Cryptography: completeness, special soundness and honest-verifier zero knowledge of Schnorr's protocol.

Reductions: 3-SAT to Clique and to 3-coloring, NAE-SAT to 3-coloring, and Subset Sum to Partition.

These are formalizations by TCSlib's authors, not new results. Our part is compiling them again against our newer Lean and Mathlib. 802 declarations compiled unchanged. 51 needed a change, for example because a Mathlib lemma had been renamed or a tactic now behaves differently. Each declaration's page shows what was changed.

Eleven of the 44 results are not imported yet. Three are left unproved (sorry) in TCSlib itself: KKL and both forms of the MRRW bound. The other eight do not yet compile against our Mathlib: the Hamming bound, list-decoding capacity, the quantum Hamming and Singleton bounds, BLR soundness, Newman's theorem, the optimality of Kruskal's algorithm, and Johnson–Lindenstrauss. They will be added as they are repaired.

Declarations

Daily reading

Daily reading: 22 proved, 9 open

The Proposer read 5 new paper(s), listed 10 open problem(s), and stated 6 of them in Lean, and posed 0 more from what the knowledgebase already holds (25 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 22 theorem(s); 9 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. catch_up_outcome_neg_eq_loss_iff: Negating a Catch-Up outcome yields a loss if and only if the original outcome is a win.

Formalized. catch_up_value_eq_loss_iff (Catch-Up: A Game in Which the Lead Alternates (2015)): For any nonempty finite set S of natural numbers, the game value catch_up_value S is CatchUpOutcome.loss if and only if for every element x ∈ S, the second player wins from the resulting position, i.e., catch_up_value_aux (S.erase x) 0 x false =…

Proved. catch_up_value_aux_empty_eq_draw_iff: In a Catch-Up position with an empty set of remaining pieces, the game value for the mover is CatchUpOutcome.draw if and only if the mover's score equals the opponent's score.

Proved. graph_matching_num_two_mul_le_card: In any finite simple graph G on vertex set V, twice the matching number is at most the number of vertices, 2 * graph_matching_num G ≤ |V|.

Formalized. konig_matching_vertex_cover_num (Diestel, Graph Theory, Theorem 2.1.1): König's theorem: in any finite bipartite simple graph G, the matching number is equal to the vertex cover number.

Formalized. gf8_oracle_t_count_ge_20 (arXiv:2610.01024): The slice-rank floor of arXiv:2610.01024 Proposition 59 at k = 3: every parity representation of the GF(8) multiplication oracle's syndrome has at least G(3) = 20 parities (every nonzero slice of the multiplication tensor has rank 6, so the column code is an…

Proved. matroid_strongly_base_orderable_freeOn: A free matroid (every subset of the ground set independent) is strongly base-orderable.

Proved. matroid_common_colorable_freeOn_one: A free matroid paired with itself has a common coloring with one color.

Proved. not_matroid_common_colorable_loopyOn_zero: A nonempty ground set has no common coloring with zero colors.

Formalized. matroid_common_colorable_of_list_colorable (arXiv:2610.07318): If (M₁, M₂) is k-list-colorable then it has a common coloring with k colors (take every list to be {0, …, k−1}); hence χ(M₁, M₂) ≤ χ_ℓ(M₁, M₂).

Formalized. matroid_common_list_colorable_one_of_colorable_one (arXiv:2610.07318): If (M₁, M₂) has a common coloring with one color (the whole ground set is independent in both), then it is 1-list-colorable: χ_ℓ = χ when χ ≤ 1.

Proved. k4_forest_aug: The forests of K₄ satisfy the matroid augmentation axiom.

Open. matroid_sbo_common_list_colorable_two (arXiv:2610.07318): If M₁ and M₂ are strongly base-orderable matroids on a common finite ground set and (M₁, M₂) has a common coloring with two colors, then (M₁, M₂) is 2-list-colorable.

Open. matroid_sbo_common_list_chromatic_eq (arXiv:2610.07318): Open problem (arXiv:2610.07318, Question 3.2): for every two strongly base-orderable loopless matroids M₁, M₂ on a common finite ground set, the common list chromatic number equals the common chromatic number: χ_ℓ(M₁, M₂) = χ(M₁, M₂).

Open. matroid_symmetry_gap_lt_eight_seventeenths (arXiv:2610.07387): Open problem (arXiv:2610.07387, after Corollary 3.4): there is a strongly symmetric instance of non-negative submodular maximization under a matroid constraint — a matroid M on a finite ground set, a non-negative submodular f and a permutation group G —…

Open. matroid_common_list_chromatic_additive_gap_bounded (arXiv:2610.07318): Open problem (arXiv:2610.07318, Question 3.1): there is a constant C such that χ_ℓ(M₁, M₂) ≤ χ(M₁, M₂) + C for every two loopless matroids M₁, M₂ on a common finite ground set.

Open. matroid_common_list_chromatic_le_chromatic_add_one: Open question (arXiv:2610.07318, Question 3.1, second part): for every two loopless matroids M₁, M₂ on a common finite ground set, χ_ℓ(M₁, M₂) ≤ χ(M₁, M₂) + 1.

Open. linear_flattening_potency_le (arXiv:2610.08504): For n ≥ 2, every linear flattening L of ℂⁿ ⊗ ℂⁿ ⊗ ℂⁿ satisfies Pot(L) ≤ 2n − 2 + 2/(n+1) = 2n²/(n+1): for every tensor T some pure tensor x ⊗ y ⊗ z has (n + 1) · rank L(T) ≤ 2n² · rank L(x ⊗ y ⊗ z).

Open. linear_flattening_potency_gt_even_exists (arXiv:2610.08504): Open problem (arXiv:2610.08504, Sec. 1.2, question (ii)): for some even n ≥ 4 there is a linear flattening L : ℂⁿ ⊗ ℂⁿ ⊗ ℂⁿ → Mat_{u×v}(ℂ) with potency greater than 2n − 2, i.e. Π(n) > 2n − 2, so that the linear flattening method gives nontrivial equations…

Open. tarski_query_complexity_five_le_log_sq (arXiv:2610.07055): Open problem (arXiv:2610.07055, Sec. 1): Tarski(n, 5) = O(log² n), i.e. there is a constant C such that for all n ≥ 2 some deterministic algorithm finds a fixed point of every monotone map [n]^5 → [n]^5 with at most C (ln n)² queries.

Open. spr_opt_diag_le_sqrt (arXiv:2610.07623): Open problem (arXiv:2610.07623, after Theorem 4.1 and Remark 6.2): the labeler can guarantee opt(n, n) = O(√n), i.e. there is a constant C with opt(n, n) ≤ C √n for every n ≥ 1.

Verified in Lean

Launch

A proof is listed only after its module builds

A result is now recorded as proved only once its Lean module has built offline with lake build, after the interactive check has accepted it. Until today the order was reversed: the result was recorded first and withdrawn if the offline build failed. While the build ran, and for good if the build itself crashed, a proof that had not passed it could be listed as proved. The same rule now covers new definitions and renamed results.

Details

Explanations in words no longer hold back the day's results. The results are published first, and their explanations follow in a second update.

Daily reading

Daily reading: 52 proved, 7 open

The Proposer read 10 new paper(s), listed 16 open problem(s), and stated 7 of them in Lean, and posed 4 more from what the knowledgebase already holds (33 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 52 theorem(s); 7 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Formalized. is_regular_one (standard textbook result (automata theory): {ε} is regular): The language {[]} containing only the empty word is regular.

Formalized. is_regular_pow (Sipser, Exercise 1.31): For any regular language L and natural number n, the n-th power L^n is regular.

Proved. karp_reducible_refl: Every language L over a finite alphabet Γ Karp-reduces to itself: the reduction is the identity function on words, which is computed by the identity Turing machine in a constant number of steps, and w ∈ L if and only if w ∈ L.

Proved. schwartz_zippel_nonzero_count_ge: Let F be an integral domain, S a nonempty finite subset of F, n a positive integer, and p a nonzero polynomial in n variables over F of total degree d. Then the number of points of S^n at which p does not vanish is at least |S|^n − d·|S|^(n−1); equivalently,…

Proved. schwartz_zippel_union_bound_zero_set: Let F be an integral domain, S a nonempty finite subset of F, n a positive integer, and let k ≥ 0, d ∈ ℕ and p_1, …, p_k be nonzero polynomials in n variables over F whose total degrees are all at most d. Then the number of points x of S^n at which at least…

Proved. schwartz_zippel_repeated_trials_zero_count_le: Let F be an integral domain, S a nonempty finite subset of F, n a positive integer, and p a nonzero polynomial in n variables over F of total degree d. Then for every t the number of t-tuples (x_1, …, x_t) of points of S^n at which p vanishes at every one of…

Formalized. poisson_bracket_self (Arnold, Mathematical Methods of Classical Mechanics, §38; Goldstein, Poole & Safko, Classical Mechanics (3rd ed.), §9.4): For any scalar function f on the phase space E × E and any phase-space point qp, the Poisson bracket {f, f}(qp) vanishes.

Proved. canonical_partition_function_pos: For any nonempty finite configuration space Ω, Hamiltonian H : Ω → ℝ, and inverse temperature β, the canonical partition function is strictly positive.

Formalized. boltzmann_distribution_sum_eq_one (Friedli & Velenik, Statistical Mechanics of Lattice Systems (2017), Section 1.2): For any nonempty finite configuration space Ω, Hamiltonian H : Ω → ℝ, and inverse temperature β, the sum of the Boltzmann distribution over all configurations is equal to 1.

Proved. wave_equation_zero: The zero function is a solution to the one-dimensional wave equation.

Formalized. quantum_commutator_jacobi_identity (Hall, Quantum Theory for Mathematicians, Chapter 3; Reed & Simon, Methods of Modern Mathematical Physics I, ch. VIII): For any three continuous linear operators A, B, C on a complex inner product space E, their commutators satisfy the Jacobi identity: ⁅A, ⁅B, C⁆⁆ + ⁅B, ⁅C, A⁆⁆ + ⁅C, ⁅A, B⁆⁆ = 0.

Proved. catch_up_value_aux_icc_erase_ne_loss_of_ge: In Catch-Up on {1, ..., N} with even triangular sum, if Player 2 has a response y ≥ x such that Player 1 does not win from the resulting position with scores (x, y), then Player 2 does not lose after opening move x.

Open. maxcut_game_strong_equilibrium_exists (arXiv:2610.04948): Open problem (arXiv:2610.04948, Sec. 4): for every finite simple unweighted graph G and every k ≥ 3, the unweighted Max-k-Cut game on (G, k) has a strong equilibrium, i.e. a coloring σ : V → [k] from which no nonempty coalition can deviate so that every…

Open. maxcut_game_optimal_strong_equilibrium_exists: Open problem (arXiv:2610.04948, Sec. 4, second question): for every finite simple unweighted graph G and every k ≥ 3, some optimal coloring of the Max-k-Cut game on (G, k) is a strong equilibrium. This would imply that every such game has a strong equilibrium.

Open. maxcut_optimal_not_strong_degree_66 (arXiv:2610.04948): There is a finite simple graph of maximum degree at most 66 with an optimal coloring of the Max-3-Cut game that is not a strong equilibrium; hence Δ*(3) ≤ 65. (The graph G_3 of arXiv:2610.04948, Sec. 2, has 91 vertices and maximum degree 66, attained by the…

Open. mms_threshold_monotone_asymptotic (arXiv:2610.06125): Open problem (arXiv:2610.06125, Sec. 3): for monotone valuations over goods, μ(n) = n + log₂ n + o(log₂ n) as n → ∞; i.e. (μ(n) − n − log₂ n) / log₂ n → 0.

Open. condorcet_dimension_le_three (Social Choice and Welfare 44(3):493-517 (2015)): Open problem: every election with at least one voter has a Condorcet winning set of at most three candidates, i.e. the Condorcet dimension of every election is at most 3.

Open. ejr_committee_monotone_chain_exists (arXiv:2007.01795): Open problem (arXiv:2007.01795, open problems, Q2): every approval profile with at least one voter admits an ordering c_1, ..., c_m of all candidates such that for every k ≤ m the prefix {c_1, ..., c_k} satisfies EJR for committee size k; equivalently, some…

Open. temporal_path_snapshots_explore_n_log_n (arXiv:2610.06228): Open problem (arXiv:2610.06228, Sec. 7): temporal graphs whose snapshots are all spanning paths can always be explored in O(n log n) steps; i.e. there is a constant C such that every such temporal graph on n ≥ 2 vertices can be explored from any start vertex…

Verified in Lean

Proved

PMMS for chores with three agents, the price of MNW, and the linkedness conjecture

Three questions from fair-division papers are settled. Every result is checked by the Lean kernel.

4 declarations verified in Lean

PMMS for chores (arXiv:2609.10493). Pairwise-MMS allocations of chores were known to fail only through EFX, whose counterexamples need at least four agents. Three agents suffice: nine chores with strictly positive additive costs (101 minus the values of that paper's three-agent goods instance) admit no allocation that is PMMS for every agent (pmms_chores_three_agents_may_not_exist). With two agents a PMMS allocation always exists, so three is the least number of agents. The proof checks all 3^9 allocations, each against a stored violation certificate.

The price of MNW for two agents (arXiv:1905.04910, Theorem 5.4). The price and the strong price of maximum Nash welfare for two agents were known to lie between 27/23 and 5/4. Both equal 27/23. For additive normalized valuations over any number of goods, every MNW allocation has at least 23/27 of the optimal social welfare (mnw_two_agents_welfare_le), and the paper's instance shows that no smaller factor works (mnw_two_agents_price_lower_bound). The proof groups the goods into four classes by their owner under an optimal and under an MNW allocation and compares Nash welfare with five allocations built from these classes.

Price of connectivity (arXiv:1908.05433, Conjecture 3.10). The conjecture says that a graph that is (2,k−1)-linked but not (2,k)-linked has two-agent price of connectivity exactly 2k/(2k−1). It is false. A 4-connected graph on eight vertices is (2,3)-linked but not (2,4)-linked, so the conjecture predicts 8/7. Yet with vertex values (5,5,5,1,1,1,12,0) the maximin share is 15 while every split into two connected parts leaves the poorer part at most 13, so the price is at least 15/13 (poc_linkedness_conjecture_counterexample). Every biconnected graph with at most seven vertices satisfies the conjecture (computed, not in Lean).

Verified in Lean
  • pmms_chores_three_agents_may_not_existPMMS allocations of chores need not exist already for three agents: there are strictly positive additive costs of three agents for nine…
  • mnw_two_agents_welfare_leFor two agents with additive, nonnegative, normalized valuations (v_i(M) = 1) over any number m of goods, every maximum Nash welfare…
  • mnw_two_agents_price_lower_boundFor every c < 27/23 there are two agents with additive, nonnegative, normalized valuations over three goods and an allocation O such that…
  • poc_linkedness_conjecture_counterexampleCounterexample to arXiv:1908.05433 Conjecture 3.10 for two agents: there is a graph G on 8 vertices that is (2,3)-linked but not…

Weekly report

Week of 28 Sep – 4 Oct 2026: Voting games rigidity, persuasion constants, and limits of strategic stopping rules

This week settled questions across game theory, fair division, and voting. In voting games, swing tables were proved to determine weighted systems uniquely. In Bayesian persuasion, the optimal receiver-utility ratio was fixed at 3/2. In chore division, the worst-case cost of fair allocations was shown to stay strictly below one-half for two people. In strategic games, uniqueness of equilibria failed for publication choice, and static stopping rules failed to remain robust against disguised rewards. Four open implications among discrete fairness criteria were also resolved.

Details

Proved. Weighted voting games are uniquely determined by their swing table

Proved. The receiver-utility constant in bit-string persuasion is 3/2

Disproved. The Publication Choice Problem can have multiple pure-strategy equilibria

Partial. Hollow-Shell market has at most ⌊(n−1)²/4⌋ blocking pairs

Lean accepted 175 new theorems this week, drawn from 18 papers and 8 topics. Of these, 51 are new here, 3 correct a published claim, 30 formalize a published result and 91 are helper lemmas. 17 statements posed this week are not yet proved. The reading covered 3 new papers. This report covers 24 results in 8 fields.

Proved

Minimal elements of width-two posets, anonymous games, scaling effects and temporal EFX

Four questions from papers on algorithms, equilibria and mechanism design are settled or advanced. Every result is checked by the Lean kernel.

10 declarations verified in Lean

Sorting and selection in posets (arXiv:0707.1532, Section 4). The conjecture is that the deterministic lower bound (w + 1)n/2 − w for finding the minimal elements is tight. It is not. For width 2 the deterministic query complexity is exactly 2n − 3 (pm_exact_complexity), so no algorithm achieves 3n/2 − 2 + C for any constant C (pm_conjecture_false). The lower bound is an adversary on unions of two chains whose potential, (components − 1) + (alive elements − 2), drops by at most one per query (pm_lower_bound).

Anonymous games (arXiv:0710.5582). The conjecture is that sλ is a lower bound for pure approximate equilibria of λ-Lipschitz games with s strategies. It holds for s = 2, 3, 4. For s = 2 the optimal guarantee is exactly 2λ: every game has a pure 2λ-equilibrium, and matching pennies has none better (ag_two_exact). For s = 3 and s = 4, a cyclic five-player game with λ = 1/8 has no pure ε-equilibrium for any ε < 4λ (ag_three_lower, ag_four_lower). The cases s ≥ 5 remain open.

Facility location with scaling effects (arXiv:2402.18908, Conclusion). The conjecture is that strategyproof, surjective deterministic mechanisms are dictatorships. This is true when the mechanism does not know the scaling function and must be strategyproof for every continuous positive q (fse_scaling_oblivious_dictatorship). It fails when the mechanism knows q: for q(y) = 1 − 9y/10, under which preferences are not single-peaked, the leftmost mechanism is strategyproof, surjective and anonymous (fse_dictatorship_conjecture_false, fse_conjecture_false_qdep).

Temporal fair division on identical days (arXiv:2607.17224, Section 7). The question is the best approximation ratio for temporal EFX; a 1/2-TEFX algorithm is known for two agents. No ratio above 1/√2 can be guaranteed: two agents with identical values 0, 1 and 2 + 2√2 per round over three rounds admit no α-TEFX allocation for any α > 1/√2 (tfx_identical_days_upper_bound).

Verified in Lean
  • pm_exact_complexityExact deterministic complexity at width 2. For every n ≥ 2, the worst-case number of comparison queries needed by a deterministic…
  • pm_conjecture_falseThe conjecture of arXiv 0707.1532 (Section 4) that the deterministic lower bound (w+1)n/2 − w for finding the minimal elements is tight…
  • pm_lower_boundDeterministic lower bound for 1-selection at width 2. Every deterministic comparison algorithm that finds the set of minimal elements of…
  • ag_two_exactExact answer for two strategies. For anonymous games with s = 2 strategies the best possible pure approximation is exactly 2λ = sλ: every…
  • ag_three_lowerThree strategies: the lower bound exceeds sλ. The 5-player, 3-strategy cyclic game is normalised and λ = 1/8-Lipschitz, yet has no pure…
  • ag_four_lowerFour strategies: the lower bound sλ is attained. The 5-player, 4-strategy cyclic game is normalised and λ = 1/8-Lipschitz, yet has no pure…
  • fse_scaling_oblivious_dictatorshipThe dictatorship conjecture holds for scaling-oblivious mechanisms. If a mechanism into [0,1] does not depend on the scaling function and…
  • fse_dictatorship_conjecture_falseRefutation of the dictatorship conjecture of arXiv 2402.18908 (Section 8). For every number n ≥ 2 of agents there is a scaling function…
  • fse_conjecture_false_qdepRefutation in the model with the scaling function as input. The mechanism fse_F (which receives q) is strategyproof and surjective for…
  • tfx_identical_days_upper_boundIdentical days, two agents: no α-TEFX guarantee beyond 1/√2. For every α > √2/2 there is an identical-days instance with two agents having…

Proved

Two published claims fail: a concavification identity and a team-order bound

Two claims made inside published proofs are false. Both counterexamples are checked by the Lean kernel.

9 declarations verified in Lean

Generalized principal-agency (arXiv:2209.01146, WINE 2024, Section 5.3). The APX-hardness proof for costly information acquisition rests on Eq. (5.4): the maximum of u*(σ) − h(σ) over the simplex equals 1/k times the maximum of u* over the cube, where u*(σ) = Σ_i max{σ_i − Σ_j e_ij σ_j, 0} and h(σ) = max{‖σ‖∞ − 1/k, 0}. For the 0/1 adjacency matrix of a graph with independence number α, the two sides are exactly 1 − 1/α + 1/k and α/k (gpa_lhs_isGreatest, gpa_rhs_isGreatest). They differ whenever 2 ≤ α < k (gpa_eq54_fails_of_indep). On the three-vertex path they are 5/6 and 2/3 (gpa_eq54_false). The auxiliary claim that u* is at most 1 on the cube fails too (gpa_cube_max_le_one_false). Exact maximisation stays NP-hard, but the stated approximation factor needs a new argument.

The Team Order problem (arXiv:2605.21234, SAGT 2024, Theorem 4(2)). Near weight n/2, an optimal line-up should beat a maximum-weight matching M* by at most (4 + o(1))/(n + 1) · Σ_{e∈M*} (p_e − 1/2)². For every odd n there is an instance where the identity is the unique maximum-weight matching (team2_unique_maxWeight) and every one of its edges has probability 1/2, so the bound is 0. The identity wins with probability 1/2, while an optimal line-up wins with probability at least 1/2 + 1/40 (team_gap_unique). So the bound fails for every choice of the o(1) term, even if it is only required of some maximum-weight matching (team_thm4_part2_exists_false). A simpler instance with a tie gives a gap of 1/16 (team_thm4_part2_false).

Verified in Lean
  • gpa_lhs_isGreatestExact value of the left side of Eq. (5.4) of arXiv:2209.01146 on the independent-set instances: max_{σ ∈ Δ_k} u*(σ) - h(σ) = 1 - 1/α(G) +…
  • gpa_rhs_isGreatestExact value of the right side of Eq. (5.4) of arXiv:2209.01146 (before the factor 1/k): max_{x ∈ [0,1]^k} u*(x) = α(G).
  • gpa_eq54_fails_of_indepEq. (5.4) of arXiv:2209.01146 (max_{Δ_k} (u* - h) = (1/k) max_{[0,1]^k} u*) fails on every graph whose independence number α satisfies 2 ≤…
  • gpa_eq54_falseEq. (5.4) in the proof of Theorem 5 (hardness of costly information acquisition) of arXiv:2209.01146 is false: it is not true that for…
  • gpa_cube_max_le_one_falseThe auxiliary claim max_{x ∈ [0,1]^k} u*(x) ≤ 1 used in the proof of Eq. (5.4) of arXiv:2209.01146 is false: on the path with three…
  • team2_unique_maxWeightThe identity is the unique maximum-weight matching of the strengthened instance.
  • team_gap_uniqueQuantitative form of the strengthened counterexample: for every m, the identity is the unique maximum-weight matching, its edges are all…
  • team_thm4_part2_exists_falseEven the existential reading of Theorem 4 (2) of arXiv:2605.21234 is false, for every function ε in place of o(1): on the (2m+3)-player…
  • team_thm4_part2_falseTheorem 4 (2) of arXiv:2605.21234 is false, whatever function is used for its o(1) term: for every m, on the (2m+3)-player instance the…

Daily reading

Daily reading: 17 proved, 13 open

The Proposer read 0 new paper(s), listed 0 open problem(s), and stated 5 of them in Lean, and posed 1 more from what the knowledgebase already holds (30 new theorem statement(s) in all, each checked against its English in both directions). The Prover proved 17 theorem(s); 13 of today's statements are not yet proved and stay in the knowledgebase, marked as such.

Details

Proved. cnf_not_satisfiable_of_has_empty_clause: A CNF formula containing the empty clause cannot be satisfied by any assignment because the empty clause contains no literals to be made true.

Proved. eval_matrix3_decomposition_apply: Entry (i, j) of the evaluation of a list of 3x3 triads on A and B is the sum, over the triads, of (sum of u_kl A_kl) times (sum of v_kl B_kl) times w_ij.

Proved. matrix3_tensor_rank_le_23_of_comm_ring (arXiv:2610.01639): Over every commutative ring R, 3x3 matrix multiplication has a valid bilinear decomposition with 23 products (the classical 23-multiplication algorithm with integer coefficients 0, 1, -1); hence its rank over R is at most 23.

Proved. matrix3_tensor_rank_le_of_surjective (arXiv:2610.01639): If f : R → S is a surjective homomorphism of commutative semirings and 3x3 matrix multiplication over R has a valid triad decomposition with at most r products, then so does 3x3 matrix multiplication over S (apply f to every coefficient).

Proved. matrix3_tensor_rank_ge_of_surjective (arXiv:2610.01639): If f : R → S is a surjective homomorphism of commutative semirings and every valid triad decomposition of 3x3 matrix multiplication over S has at least r products, then the same holds over R.

Proved. matrix3_tensor_rank_zmod2_le_of_int (arXiv:2610.01639): A bilinear algorithm for 3x3 matrix multiplication with integer coefficients and r products gives one over the two-element field with r products; in particular a 22-product algorithm over the integers would answer the two-element-field question positively.

Proved. matrix3_tensor_rank_int_ge_of_zmod2 (arXiv:2610.01639): A lower bound r on the number of products of every bilinear algorithm for 3x3 matrix multiplication over the two-element field is also a lower bound over the integers.

Proved. gf8_oracle_syndrome_zero_zero_zero: Sanity check: the syndrome of the GF(8) oracle is 1 on {x_0, y_0, z_0}, since 1 * 1 = 1 has bit 0 set.

Proved. gf8_oracle_syndrome_same_register: Sanity check: the syndrome of the GF(8) oracle vanishes on a triple inside one register.

Proved. gf8_oracle_syndrome_singleton: Sanity check: the GF(8) oracle has no order-one syndrome.

Proved. gf8_oracle_t_count_le_29 (arXiv:2610.01024): The GF(8) multiplication oracle has a parity representation with 29 parities, so its T-count delta(U_3) in Hadamard-free {CNOT, T} circuits is at most 29 (an explicit certificate found by SAT search; the merged schoolbook expansion has 36).

Proved. ccz_layer_syndrome_one_block: Sanity check: for two CCZ gates the syndrome is 1 on the first block {0, 1, 2}.

Open. matrix3_tensor_rank_int_le_22 (arXiv:2610.01639): Open problem (arXiv:2610.01639, Sec. 8): there is a bilinear algorithm for 3x3 matrix multiplication with integer coefficients that uses 22 products, i.e. a list of 22 triads over the integers whose evaluation equals A * B for all integer 3x3 matrices A and…

Open. matrix3_tensor_rank_zmod2_le_22 (arXiv:2610.01639): Open problem (arXiv:2610.01639, Sec. 7, Table 1, row 0, interval [22, 23]): the tensor rank of 3x3 matrix multiplication over the two-element field is 22, i.e. there are 22 triads over ZMod 2 whose evaluation equals A * B for all 3x3 matrices A, B over ZMod…

Open. matrix3_tensor_rank_zmod2_ge_22 (arXiv:2610.01639): Every bilinear algorithm (valid triad decomposition) for 3x3 matrix multiplication over the two-element field uses at least 22 products. This is the main lower bound of arXiv:2610.01639 (row 0 of its table), proved there by a large computer-assisted Lean…

Open. gf8_oracle_t_count_le_23 (arXiv:2610.01024): Open Problem 10 of arXiv:2610.01024 presupposes that the T-count delta(U_3) of the GF(8) multiplication oracle (Hadamard-free {CNOT, T} circuits, diagonal Cliffords free) lies in [20, 23] and asks for its value: this is the upper end, that some set of at…

Open. gf8_oracle_t_count_ge_24 (arXiv:2610.01024): Negation of the premise of Open Problem 10 of arXiv:2610.01024: every set of nonzero parities on nine qubits whose order-1, 2, 3 moments equal the syndrome of the GF(8) multiplication oracle has at least 24 elements, i.e. delta(U_3) > 23, so the T-count is…

Open. gf8_oracle_t_count_ge_20 (arXiv:2610.01024): The slice-rank floor of arXiv:2610.01024 Proposition 59 at k = 3: every parity representation of the GF(8) multiplication oracle's syndrome has at least G(3) = 20 parities (every nonzero slice of the multiplication tensor has rank 6, so the column code is an…

Open. ccz_layer_parity_repr_card_ge (arXiv:2610.01024): The Hadamard-free floor of arXiv:2610.01024 (Corollary 83): every set of nonzero parities on 3m qubits whose order-1, 2, 3 moments equal the syndrome of the layer of m >= 1 disjoint CCZ gates has at least 6m + 1 elements; so every Hadamard-free {CNOT, T}…

Open. resolution_refutation_exists_of_unsat (arXiv:2610.02047): Completeness of resolution, the length-free part of the problem: every unsatisfiable CNF formula has a resolution refutation.

Open. clifford_t_circuit_matrix_nil: Sanity check: the empty circuit has the identity matrix.

Open. clifford_t_gate_matrix_t_sq: Sanity check: T^2 = S on every qubit.

Open. implements_ccz_layer_zero: Sanity check: with no CCZ gates and no ancillas, the empty circuit implements the (identity) layer.

Open. ccz_layer_t_count_below_6m_plus_1_with_hadamards (arXiv:2610.01024): Open problem (arXiv:2610.01024, abstract and Sec. 1): in general Clifford+T circuits, with Hadamards anywhere and any number of clean ancillas, some layer of m >= 2 disjoint CCZ gates can be implemented with strictly fewer than 6m + 1 T gates (the exact…

Verified in Lean

Proved

Publication choice can have several equilibria, and no static threshold is 1/2-robust against strategic signaling

Two questions from recent papers on strategic behaviour are settled, both negatively. Every result is checked by the Lean kernel.

11 declarations verified in Lean

The Publication Choice Problem (arXiv:2511.13678, AAAI 2026). Theorem 4.1 states that a binary-type problem with a non-competitive venue has a unique pure-strategy equilibrium, and Conjecture 4.2 extends this to many types. Both fail. Take α = 1/2, β = 2 and two venues. Types 1 and 8 have densities 1/3 and 2/3. The first venue costs 1 for both types. The second costs 128 for the low type and 8 for the high type. This instance satisfies Assumptions 1 and 2 and has three equilibria with different venue impacts (publication_choice_three_equilibria, publication_choice_binary_equilibrium_not_unique). The best response has a closed form (pcp_half_two_best_response). Each equilibrium comes from a root of a degree-9 polynomial, and the intermediate value theorem gives three roots. Types 1, 2 and 8 with densities 3/10, 1/20 and 13/20 do the same for Conjecture 4.2 (publication_choice_three_types_three_equilibria, publication_choice_many_type_equilibrium_not_unique). With two venues, the equilibrium condition reduces to one variable for any number of types (pcp_two_venue_equilibrium_of_fixed_point).

Prophet inequality with strategic reward signaling (arXiv:2409.18269, NeurIPS 2024, Remark 4.1 and §6). Is some threshold stopping policy 1/2-robust? For static thresholds, no. There are four players: X1 = 24/5 surely; X2 = 9 with probability 1/10, else 3; X3 = 1000 with probability 1/300, else 0; X4 = 40 with probability 1/25, else 0. Then OPT = 155039/15625. Under every static threshold and in every equilibrium of the players' signaling game, the searcher gets at most 721/150 < OPT/2 (spx_searcher_payoff_le). The threshold T = 9 attains this bound, a ratio of 450625/930234 ≈ 0.4844 (spx_searcher_payoff_at_nine, strategic_prophet_no_half_robust_static_threshold). The signaling game is formalised in full, with finite signal sets. Equilibria exist for every threshold (spi_two_point_equilibrium_exists). In every equilibrium, each reached player behaves as in Proposition 3.1 (spi_two_point_equilibrium_pinned). Whether a dynamic threshold can be 1/2-robust remains open.

Verified in Lean

Proved

Proportional representation: Monroe meets Droop, and temporal FJR

Three questions from recent work on proportional representation with approval ballots are settled. Every one is checked by the Lean kernel.

9 declarations verified in Lean

Monroe and Droop-JR (arXiv:2508.00811). The Monroe rule satisfies Droop-JR when the committee size k divides the number of voters n; the other case was open. It holds for all n and k (hare_monroe_satisfies_droop_jr). Suppose a group of more than n/(k + 1) voters approves a common candidate but nobody in the committee. Then some winner can be swapped for that candidate along a reassignment of the voters, and the Monroe score goes up (monroe_exists_heavy, assignment_exchange).

Droop Monroe (arXiv:2508.00811, Table 1). When k + 1 divides n, every Droop Monroe committee provides Droop-PJR+ (droop_monroe_satisfies_droop_pjr_plus), by the same exchange. The table's other open entries, Droop-EJR+ and Droop-FJR, can fail through ties. With n = 3, k = 2 and ballots {0, 1, 3}, {0, 1, 3}, {0, 2, 3}, every committee scores 2, and the winning committee {1, 2} violates both (droop_monroe_tie_violates_droop_ejr_plus_fjr). Whether some winning committee always provides them remains open.

Temporal voting (arXiv:2505.22513, App. A.3). BFJR implies Droop-FJR. Does the converse hold, possibly under extra assumptions? No, even when no voter approves every candidate in a round (temporal_droop_fjr_not_imp_bfjr). Take four voters, six rounds and candidates 0 and 1. Voter 0 approves 0 in the first three rounds, voter 1 approves 0 in the last three, and voters 2 and 3 approve 1 throughout. Electing 1 in every round provides Droop-FJR (tfjr_count) but not BFJR (tfjr_avoid).

Verified in Lean
  • hare_monroe_satisfies_droop_jrEvery committee selected by the (Hare) Monroe rule satisfies Droop-JR, for all n and k. This answers the open question of Justified…
  • monroe_exists_heavyCounting step: if (W, π) is Hare-valid and |S| > n/(k+1), then some c ∈ W satisfies |π⁻¹(c)| < min(|π⁻¹(c)|, |S|) + |π⁻¹(c) ∩ S|.
  • assignment_exchangeGeneric exchange step: for an assignment π of voters to β, a set S of voters each satisfied by c′ but not by π, and the renaming c ↦ c′…
  • droop_monroe_satisfies_droop_pjr_plusIf k + 1 divides n, every committee selected by the Droop Monroe rule satisfies Droop-PJR+. This settles the PJR+ entry for Droop Monroe…
  • droop_monroe_unsatisfied_groupLemma 2 of arXiv:2508.00811: if (k+1) divides n and π is an optimal Droop-valid assignment for W, no group of more than n/(k+1) voters,…
  • droop_monroe_tie_violates_droop_ejr_plus_fjrWith ties, the Droop Monroe rule can output a committee violating Droop-EJR+ and Droop-FJR even when k + 1 divides n: n = 3, k = 2,…
  • temporal_droop_fjr_not_imp_bfjrDroop-FJR does not imply BFJR in temporal voting, even when no voter approves every candidate in a round. Four voters, six rounds, two…
  • tfjr_countIn the counterexample, every nonempty group either contains voter 2 or 3, or has a member who approves nothing in at least ⌈(|T|+1)|S|/4⌉…
  • tfjr_avoidAny two rounds out of six miss some round among 0-2 and some round among 3-5 (checked by decide).

Proved

Four results in social choice: Plackett–Luce entropy, leximin and EF1, Banzhaf veto distortion, apportionment

Four questions from recent papers in social choice and fair division are answered, one of them in part. Every one is checked by the Lean kernel.

11 declarations verified in Lean

Plackett–Luce entropy (Preference Elicitation as Average-Case Sorting, AAAI 2021, §5). The Plackett–Luce model with weights 1, φ, …, φ^(m−1) was known to have entropy O(m) for φ ≤ 1/2, and conjectured to for all φ in (0, 1). The conjecture holds: the entropy is at most C(φ)·m with C(φ) = log(1/(1 − φ)) + log(1/φ)·φ/(1 − φ)² (plackett_luce_geometric_entropy_linear). Each choice step has entropy at most C(φ) (geometric_first_choice_entropy_le), and the chain rule adds the steps up (plackett_luce_entropy_eq).

Leximin among EF1 allocations (The Unreasonable Fairness of Maximum Nash Welfare, App. B). The conjecture that a leximin-optimal EF1 allocation need not be Pareto optimal is true (leximin_ef1_not_pareto_optimal). Four agents with additive values normalised to sum to 1 share six goods. The only leximin-optimal EF1 allocation is ({g2, g3}, {g4}, {g1, g5}, {g6}). Swapping g4 and g5 between agents 2 and 3 raises one utility and lowers none. The kernel checks all 4^6 allocations (lxv_check_eq_true).

Banzhaf veto distortion (How Banzhaf Makes a Victor, EC 2026, §5). For quota 1/2 with a strict majority, the paper shows a veto distortion of at least 6/5 and leaves the case open. It is at least 5/3 (banzhaf_strict_majority_veto_distortion_ge_five_thirds): for the target (3/10, 7/10, 0), the weights (1/2, 1/2, 0) are optimal and give the first player a veto. Whether the distortion is bounded remains open.

Apportionment (arXiv:2202.11061, §3.1; arXiv:2608.02759). For four states, quota and population monotonicity were known to be incompatible when the compared house sizes may differ. They are incompatible even when every comparison uses one house size (quota_population_monotone_incompatible_four_states_common_house): twelve population profiles at house size 8 force a population paradox. The kernel runs a pruned search over the 13,824 quota-compliant assignments (apx_kernel_check, apx_search_sound).

Verified in Lean

Proved

The receiver-utility constant in bit-string persuasion is 3/2

In the persuasion model of arXiv:2606.22226, a sender sees a string of bits and signals to a receiver who wants to guess every bit; the sender wants as many bits as possible guessed as 1. Let R_0 be the receiver's utility from the prior alone and R_max the best receiver utility among sender-optimal schemes. The paper proves R_max ≤ (3/2)·R_0 for every prior, gives a six-bit prior with ratio 39/31, and asks for the optimal constant (§5). It is 3/2 (persuasion_receiver_ratio_optimal_constant).

10 declarations verified in Lean

The lower bound comes from slice priors. Take 2m + 1 equally likely states and one coordinate for each m-element subset of them; a state has a 1 at a coordinate exactly when it belongs to that subset. Then R_max / R_0 = (3m + 1)/(2m + 2) (persuasion_slice_ratio). That is 13/10 at m = 4, and it tends to 3/2 (persuasion_receiver_ratio_sharp). The sender's best scheme pairs the true state with a uniformly random other state and recommends the coordinatewise OR (persuasion_pair_or_scheme_obedient). That no scheme serves the sender better comes down to a counting lemma. For nonnegative weights on 2m + 1 points, at most C(2m+1, m) − C(2m−1, m) of the m-subsets carry at least half the weight (persuasion_heavy_sets_card_le). The paper's upper bound is checked too (persuasion_max_receiver_utility_le).

Verified in Lean

Proved

Sybil-proof, safe procurement without winner-take-all, and QA pooling without convex exposure

Two questions from recent papers in mechanism design and forecast aggregation are answered. Both are checked by the Lean kernel.

10 declarations verified in Lean

Procurement (arXiv:2603.27779, §6). Is there a procurement mechanism that is Sybil-proof, ex-post safe and has equilibria that are not winner-take-all? Yes (exists_sybil_proof_ex_post_safe_non_winner_take_all_procurement). In the stake-share mechanism, agents post costless stakes. The work and a reward pool L + K·max(0, M − W) are split in proportion to stake, where W is the total stake. The mechanism is Sybil-proof (proportional_share_sybil_proof) and ex-post safe for every cost up to L (stake_share_ex_post_safe). No pure Nash equilibrium gives all the work to one agent (stake_share_pne_non_winner_take_all). When KM ≥ 3n²L, an equilibrium exists for every cost profile in [0, L]^n (stake_share_pne_exists). Safety for every cost is out of reach in this class: a proportional-share mechanism that is safe for all costs has no pure Nash equilibrium (proportional_share_no_pne_of_safe_all_costs).

QA pooling (arXiv:2102.07081, §7). The score of the quasi-arithmetic pool is concave in the experts' weights when the scoring rule has convex exposure. Without that assumption it need not be (tsallis_three_qa_weight_score_not_concave). For the Tsallis rule with parameter 3, two experts on three outcomes suffice, and concavity fails by 23/1024.

Verified in Lean

Proved

Five results in fair division: subsidies, connected chores, weighted PROPX, PROPm and the cost of EF1

Five questions left open in recent papers on fair division now have answers, complete or in small cases. Every one is checked by the Lean kernel.

18 declarations verified in Lean

Envy-free orientations with subsidy (arXiv:2502.13671). Agents are the vertices of a multigraph and the items are its edges. Valuations are monotone, with largest marginal value 1. A total subsidy of n − 1 always suffices, n − 2 is sometimes needed, and n − 2 is tight on simple graphs. On multigraphs n − 2 is not enough. Three agents can need 3/2 (multigraph_orientation_subsidy_three_halves, multigraph_orientation_subsidy_not_le_one), and four agents can need 13/5 (multigraph_orientation_subsidy_four_agents). In the three-agent instance, edge b joins agents 0 and 1, and parallel edges c and d join agents 0 and 2. Agent 1 values b at 1. Agent 2 values one of c and d at 1/2 and both at 3/2. Agent 0 values b at 1, c and d at 1/4 each, any two edges at 5/4 and all three at 3/2. Each of the eight orientations needs at least 3/2. For general n the gap between n − 2 and n − 1 remains open.

Connected EF1 and PO for chores (arXiv:2609.32545, §4). Connected EF1 allocations of chores on a path always exist. Can they also be Pareto optimal? Not always (connected_ef1_po_chores_incompatible). Take three agents and five chores, with costs (4, 2, 1, 2, 2) for one agent and (1, 4, 1, 2, 2) for the other two. Of the 63 connected allocations, 12 are EF1. Each of them is Pareto-dominated by a connected allocation with cost vector (5, 1, 2), (2, 1, 5) or (2, 5, 1).

Weighted PROPX and PO for chores (arXiv:2608.16130). With equal shares, PROPX and PO are compatible for two agents. With unequal shares they can conflict, and the known two-agent example uses shares 3/10 and 7/10. Shares much closer to equal fail too: 2/5 and 3/5 with five chores, and 9/20 and 11/20, and 3/7 and 4/7, with six (weighted_propx_po_chores_incompatible_two_fifths, weighted_propx_po_chores_incompatible_nine_twentieths, weighted_propx_po_chores_incompatible_three_sevenths). This suggests that every unequal share vector fails for two agents.

PROPm orientations (arXiv:2602.18098, §6). Do PROPm orientations of a multigraph always exist? Yes, whenever the simple graph underneath can be oriented with every out-degree at most 2 (propm_orientation_of_out_degree_two). So they exist on every multigraph with at most five agents (propm_orientation_exists_of_le_five). For each adjacent pair, one agent splits the parallel edges into an EFX partition under her own values (propm_efx_split) and the other agent chooses. Each agent cuts for at most two pairs, and each cut costs her at most half of one edge held by that neighbour. The PROPm slack is taken over the agent's relevant edges, as the paper does for PROPX.

The cost of EF1 for chores (arXiv:2410.15738). Each agent's costs sum to 1. The cost of EF1 is the worst-case gap between the cheapest EF1 allocation and the cheapest allocation. It was stated to be at least 1 − 1/n. For two agents it lies between 1/6 and 3 − 2√2 ≈ 0.172 (cost_of_ef1_two_le, one_sixth_le_cost_of_ef1_two, not_forall_cost_of_ef1_ge_one_sub_inv). For the upper bound, start from the cost-optimal allocation, where at most one agent envies. Move her chores to the other agent in order of cost ratio until she is EF1 (exists_ef1_social_cost_le_opt_add_two_agents). The lower bound comes from costs (0, 1/2, 1/2) and (1/3 − 2δ, 1/3 + δ, 1/3 + δ) (ef1_cost_gap_instance_spec). Computer search suggests the exact value is 1/6.

The counterexamples were found by computer search in exact arithmetic. The proofs were written by the operator (Claude), not by the agents.

Verified in Lean

Proved

All four open implications among fairness notions in arXiv:2502.02815 are settled

A recent paper (arXiv:2502.02815) maps which of 22 fairness notions in discrete fair division imply which, and closes with four implications it could not resolve. All four are now settled, and every answer is checked by the Lean kernel.

9 declarations verified in Lean

Open problem 1, additive goods with equal entitlements: does MXS imply EEF1? No (mxs_fair_not_imp_eef1_fair). Three agents share twelve goods worth 2, 2, 2, 1, 2, 2, 2, 3, 2, 2, 2, 3 to each of them, and the first agent holds the goods worth 1, 3 and 3. Splitting the goods as {2,2,2,1}, {2,2,2,3}, {2,2,2,3} is EFX for her, so her bundle, worth 7, meets her minimum EFX share. But the other two agents must share nine goods worth 2, so one of them gets five, which are worth 8 to her even after removing any good. The implication holds for two agents (Lemma 22 of the paper), so three is the fewest.

Open problem 2, additive goods with unequal entitlements: does APS imply PROP1? No, already for two agents and three goods (aps_fair_not_imp_prop1_fair).

Open problem 3, submodular goods with equal entitlements: does MXS imply PROP1? Yes (mxs_fair_imp_prop1_fair_of_submodular). If an MXS-fair bundle failed PROP1, take from each other bundle of the EFX certificate a good outside the agent's bundle with positive marginal value. EFX bounds what remains, and four submodularity lemmas then bound the value of all the goods strictly below itself.

Open problem 4, submodular goods with binary marginals: does pairwise MMS imply MMS? No, with three agents and six goods under a partition-matroid valuation (pmms_fair_not_imp_mms_fair).

The definitions follow the paper literally (its Definitions 2, 3, 5, 7, 8, 9, 10 and 11). The proofs were written by the operator (Claude) rather than found by the agents.

Verified in Lean
  • mxs_fair_not_imp_eef1_fairOpen problem 1 of arXiv:2502.02815 has a negative answer: for additive goods and equal entitlements, MXS does not imply EEF1, already for…
  • aps_fair_not_imp_prop1_fairOpen problem 2 of arXiv:2502.02815 has a negative answer: for additive goods with unequal entitlements, an allocation can give every agent…
  • mxs_fair_imp_prop1_fair_of_submodularOpen problem 3 of arXiv:2502.02815 has a positive answer: for submodular goods (monotone, v(empty) = 0) and equal entitlements, every…
  • pmms_fair_not_imp_mms_fairOpen problem 4 of arXiv:2502.02815 has a negative answer: for submodular goods with binary marginals and equal entitlements, pairwise MMS…
  • submod_union_marginal_leFor a monotone submodular set function and S inside T, adding a set U gains at most as much on top of T as on top of S.
  • submod_biUnion_le_sumA monotone submodular set function with v(empty) = 0 is subadditive over a finite family of sets.
  • submod_add_items_leFor a submodular set function, if every item of G lies outside A and adds at most delta >= 0 to A, then adding G to any superset W of A…
  • submod_zero_marginalsFor a monotone submodular set function, items whose marginal value over X is zero add nothing on top of any superset of X.
  • capped_count_marginalFor S inside T and j outside T, the capped count S -> min(#{x in S : p x}, c) has a marginal value at T at most its marginal value at S,…

Proved

A weighted voting game is determined by its swing table

A recent paper on power indices (arXiv:2607.07013, §10) asks how much of a simple game the power indices can see. Every semivalue power index (Shapley–Shubik, Banzhaf and the rest) uses only the swing table: for each player i and size k, the number T(i, k) of losing coalitions of size k that i turns into winning ones. Different simple games can share a swing table, but in an exhaustive census at n = 5 and searches at n = 6, 7 no weighted game ever did. Conjecture 10.2, swing rigidity, says this always holds: a weighted voting game is determined by its swing table among all simple games. The paper lists it first among its open problems.

14 declarations verified in Lean

The Lean kernel now confirms it: weighted_game_eq_of_swing_table_eq. It also confirms that the hypothesis is needed: the two five-player games of the paper's Example 10.1 are different, share a swing table, and so are both unweighted (exists_not_weighted_simple_games_ne_of_swing_table_eq).

The proof reduces the conjecture to Chow's theorem of 1961. Summing the swings at size k over all players gives (k+1)w(k+1) − (n−k)w(k), where w(k) counts the winning coalitions of size k; since w(0) = 0, the swing table fixes every w(k) and so the number |W| of winning coalitions. Each player's total swing count is 2A(i) − |W|, where A(i) counts the winning coalitions containing i, so the table fixes every A(i) too. Those are exactly Chow's parameters. If f is weighted with quota q and g is any simple game with the same parameters, the sum over all coalitions S of (f(S) − g(S))(w(S) − q) is zero while every term is at least zero, and a coalition that g wins and f loses would make its term positive; so g wins only where f does, and equal counts give f = g.

The proof and its lemmas were written by the operator (Claude) rather than found by the agents.

Verified in Lean

Proved

Two open implications among fairness notions are false

A recent paper (arXiv:2502.02815, §6) maps how fairness notions for indivisible goods imply one another and lists four implications it could not settle. The Lean kernel now refutes two of them with small counterexamples.

3 declarations verified in Lean

Open problem 2, aps_fair_not_imp_prop1_fair: for additive goods with unequal entitlements, giving every agent her AnyPrice share does not guarantee PROP1, even with two agents. Both agents value three goods at 2, 1 and 1, and the entitlements are 4/5 and 1/5. At prices 81/100, 19/200 and 19/200 the big good is outside the first agent's budget, so her AnyPrice share is at most 2, and the big good alone meets it. Her proportional share is 16/5, though, and even adding one more good leaves her at 3, so PROP1 fails.

Open problem 4, pmms_fair_not_imp_mms_fair: for submodular goods with binary marginals, pairwise MMS does not imply MMS. Three agents share the valuation min(#X, 1) + min(#Y, 2) on six goods split into X = {0, 1, 2} and Y = {3, 4, 5}. The allocation {0, 1, 3}, {2}, {4, 5} is pairwise-MMS-fair to every agent. Yet the second agent gets 1, while the partition {0, 3}, {1, 4}, {2, 5} shows that her maximin share is 2.

The other two implications, problems 1 and 3, are still open. Both counterexamples and their proofs were written by the operator (Claude), not found by the agents.

Verified in Lean
  • aps_fair_not_imp_prop1_fairOpen problem 2 of arXiv:2502.02815 has a negative answer: for additive goods with unequal entitlements, an allocation can give every agent…
  • pmms_fair_not_imp_mms_fairOpen problem 4 of arXiv:2502.02815 has a negative answer: for submodular goods with binary marginals and equal entitlements, pairwise MMS…
  • capped_count_marginalFor S inside T and j outside T, the capped count S -> min(#{x in S : p x}, c) has a marginal value at T at most its marginal value at S,…

Proved

Weakly consecutive sequences of prime-power length begin or end with 1

A permutation s of {1, …, k} is weakly consecutive if, whenever m divides s(i) and m divides i − j, m also divides s(j). A recent paper (arXiv:2401.09497) conjectures that every such sequence begins or ends with 1 (Conjecture 5.2).

9 declarations verified in Lean

The Lean kernel now proves this whenever the length k is a prime power: weakly_consecutive_one_at_end_of_isPrimePow. The general conjecture is still open.

The proof counts residue classes. If m divides the first term, every position congruent to the first one holds a multiple of m. That is ⌈k/m⌉ positions for the ⌊k/m⌋ multiples of m up to k, so m divides k. By reversal, the same holds for the last term. If a prime p divided both end terms, the first term's class would already use up all k/p multiples of p, and the last position lies outside that class. So the two end terms are coprime. If 1 sat strictly inside, both end terms would exceed 1, divide k and be coprime, which is impossible when k is a power of a single prime.

The proof and its lemmas were written by the operator (Claude), not found by the agents.

Verified in Lean

Proved

The cyclic Hollow-Shell market has at most ⌊(n−1)²/4⌋ blocking pairs

A recent paper (arXiv:2609.17418) studies the Shield number σ(n) of the stable marriage problem: the least, over all preference profiles with n men and n women, of the most blocking pairs a complete matching can have. The paper builds a cyclic instance C_n, the Hollow-Shell, in which man g ranks the women g, g+1, g+2, … around the circle and every woman ranks the men in exactly the opposite order. Its Theorem 5.3 gives a matching of C_n with ⌊(n−1)²/4⌋ blocking pairs, and §7.1 leaves open whether any matching has more. The Shield-Core conjecture is σ(n) = β(C_n) = ⌊(n−1)²/4⌋.

8 declarations verified in Lean

The Lean kernel now confirms that no matching of C_n has more: hollow_shell_blocking_pairs_ncard_le. So β(C_n) = ⌊(n−1)²/4⌋ exactly, and σ(n) ≤ ⌊(n−1)²/4⌋ for every n. The rest of the conjecture, that every profile has a matching with at least that many blocking pairs, is still open.

The proof reads each man's match as an arc running clockwise from him to his partner; a blocking pair is one arc strictly nested inside another. Cut the circle at 0 and let k be the number of arcs that cross the cut. Counting inversions of the matching on each side of the cut, and the crossings between the two sides, bounds the nested pairs by k(n−1−k), which is at most ⌊(n−1)²/4⌋.

The proof and its seven lemmas were written by the operator (Claude) rather than found by the agents.

Verified in Lean
  • hollow_shell_blocking_pairs_ncard_leThe upper half of the Shield-Core conjecture (arXiv:2609.17418, open): in the cyclic Hollow-Shell market C_n, every complete matching has…
  • perm_nesting_boundFor a permutation s of {0, ..., n-1} with k positions a such that s(a) < a: the number of pairs (a, b) with a < b and s(b) < s(a) and a, b…
  • perm_cross_inv_sumFor a permutation s of {0, ..., n-1}, let A be the positions a with s(a) < a. The number of pairs (a, b) with a in A, b not in A, b < a…
  • perm_card_left_gt_leFor a permutation s of {0, ..., n-1} and a position x with x <= s(x), the number of i < x with s(i) > s(x) is at most the number of j > x…
  • perm_inv_left_addFor a permutation s of {0, ..., n-1} and a position g, the number of x < g with s(x) > s(g), plus s(g), equals the number of x > g with…
  • perm_gap_sumFor a permutation s of {0, ..., n-1}, the number of pairs (a, b) with s(b) < a < b, plus the number of positions b with s(b) < b, plus the…
  • perm_card_val_ltFor a permutation s of {0, ..., n-1} and c in {0, ..., n-1}, exactly c elements x have s(x) < c.
  • cyclic_offset_eq_iteFor i, j in {0, ..., n-1}, the cyclic offset (j - i) mod n equals j - i when i <= j and j + n - i otherwise.

Launch

News on the home page

The home page now has a News section: launches and newly proved results, newest first. Every proof announced here links the declarations Lean accepted; nothing is called proved here that Lean has not checked.

Proved

Catch-Up on {1, …, N} is a draw for N = 3, 4, 7, 8 and 11

Catch-Up, introduced in 2015, is a two-player game: players take numbers from a shared set, and on each turn a player keeps taking until their total catches up with or passes the other's, and the higher total at the end wins. The paper that introduced it computed optimal play on {1, …, N} for every N up to 20 and found a draw exactly when 1 + 2 + … + N is even (its Table 2). The open conjecture is that this holds for every N.

6 declarations verified in Lean

This entry is a formalization of the paper's computation, not a new result: the Lean kernel now checks its draws for N = 3, 4, 7, 8 and 11. Each case is checked with decide +kernel (no native_decide) through a pruned search, catch_up_search, which is proved to agree with the game's value. Together with the parity theorem the day before, the paper's computed values are checked in Lean for every N ≤ 11. The general case is still open.

The search, its correctness proof and the split of the conjecture into these steps came from the operator rather than the agents; the agents proved the rest of the ladder.

Corrected on 6 October 2026: an earlier version of this entry did not say that the paper had already computed these values.

Verified in Lean
  • catch_up_value_icc_threeThe game value of Catch-Up played on the interval Finset.Icc 1 3 is CatchUpOutcome.draw.
  • catch_up_value_icc_fourCatch-Up on {1, 2, 3, 4} is a draw under optimal play (the case N = 4 of the Catch-Up conjecture), evaluated by the Lean kernel.
  • catch_up_value_icc_sevenCatch-Up on {1, ..., 7} is a draw under optimal play (the case N = 7 of the Catch-Up conjecture), evaluated by the Lean kernel.
  • catch_up_value_icc_eightCatch-Up on {1, ..., 8} is a draw under optimal play (the case N = 8 of the Catch-Up conjecture), evaluated by the Lean kernel.
  • catch_up_value_icc_elevenCatch-Up on {1, ..., 11} is a draw under optimal play (the case N = 11 of the Catch-Up conjecture), established by the kernel through the…
  • catch_up_search_eqFor every fuel n, list l, scores and opening flag, the pruned search with flag true returns true exactly when the computable mirror's…

Proved

Catch-Up: no draw when the total is odd

On any finite set of numbers whose sum is odd, Catch-Up cannot end in a draw: the two totals always differ by an odd amount. The 2015 paper that introduced the game already makes this observation (Sec. 2.6 and Appendix 5.1); this is a formalization of it. On {1, …, N} it is the easy half of the conjecture above, for every N at once.

2 declarations verified in Lean

Corrected on 6 October 2026: an earlier version presented this as settling half of the conjecture, without saying the paper already proves it.

Verified in Lean

Launch

The site is live

Every verified declaration now has its own page: the statement in words and in Lean, the full proof, what it builds on and what builds on it. Anyone with a GitHub account can submit a problem for the machine to attempt, or challenge a result whose name claims more than it proves.

13,593machine-verified declarations8467 theorems · 5126 definitions
44topics with resultsacross 7 of 7 domains
2,872theorems awaiting proofstated in Lean, not yet proved
$1,227API spend to dateabout 9¢ per verified declaration
Point a loop at a curriculum of theoretical domains. Let it state and prove, forever. Keep only what Lean accepts. Publish all of it, immediately, to everyone.

Theory is written to be read by people, and it is checked by people. That checking is the bottleneck: it does not scale, it is unevenly spread across subjects, and most of it is never written down in a form another person or machine can reuse.

Lean 4 and Mathlib remove the bottleneck for the checking. AFTD automates the other half — deciding what is worth stating, writing the statement, attempting the proof — a subject at a time, not a paper at a time. The unit of progress is one Lean declaration that anyone can import tomorrow.

Read the programme →

How it works

Three agents, one dependency graph, and Lean as the only judge

  1. 1

    Propose

    The Proposer works one topic at a time: it transcribes known results from the literature and proposes its own — generalizations, converses, bridges between topics.

  2. 2

    Round-trip

    One model reads only the Lean and says what it asserts; another reads only the English and formalizes it blind. Lean is asked first whether the two agree.

  3. 3

    Prove

    The Prover takes one statement at a time. When it is stuck it files the lemmas it needs and the graph grows a rung — it never quietly proves something weaker.

  4. 4

    Publish

    Lean elaborates it, #print axioms is clean, and it goes public with its full source, rebuildable by anyone with lake build.

The scheduler is not a language model. What to work on, when to give up and when to split a hard theorem into lemmas are deterministic rules over the graph; the models supply judgement, never control flow.

The curriculum

Seven theoretical domains

13,593 declarations across 56 topics.

open

Theoretical Computer Science

The mathematics of computation: what can be computed, at what cost, and with what guarantees. Machine models, complexity classes, algorithm…

6,800 declarations · 24 topics

open

Optimization & Operations Research

The theory of optimizing over structured sets: when an optimum exists, when it is certified by a dual object, and how fast an iterative method…

206 declarations · 5 topics

open

Probability & Mathematical Statistics

Measure-theoretic probability and the theory of statistical inference: limit behaviour, concentration, what an estimator can achieve, and what no…

38 declarations · 6 topics

open

Mathematical Logic & Foundations

The mathematics of formal systems themselves: what structures satisfy a theory, what a proof system can and cannot prove, and the set-theoretic…

104 declarations · 4 topics

open

Game Theory & Mathematical Economics

Strategic interaction and the theory of collective decisions: when equilibria exist, what aggregation rules are possible, and what mechanisms can be…

3,465 declarations · 7 topics

open

Theoretical Physics

The mathematical structure of physical theory: the variational principles that generate the equations of motion, the symmetries that constrain them,…

2,756 declarations · 9 topics

open

Number Theory

Elementary questions about the integers -- divisibility, factorizations and permutations constrained by divisibility -- including small open…

224 declarations · 1 topics

Take part

Pose, read, judge

01

Pose a problem

Send a statement through a GitHub issue. Once reviewed, the machine formalizes it, checks the formalization says what you said, and tries to prove it. Your handle stays on it.

How submitting works →
02

Read the proofs

Every verified declaration has its own page: the statement in words and in Lean, the full proof, what it builds on and what builds on it.

The knowledgebase →
03

Judge the verdicts

Correctness is settled by Lean. Everything else is yours to dispute: a name that claims too much, a trivial statement, a definition the subject would not recognise.

Challenge a result →

Community problems

All open problems →

No community problems yet. Pose the first one: a statement from any of the domains, with a reference if you have one.

Submit a problem

Recently verified

Knowledgebase →

If a finite unweighted graph admits a k-coloring in which each vertex has at most one same-colored neighbor, it admits a strong k-color equilibrium, also with at most one same-colored neighbor per vertex.

Equilibria in GamesElementary finite potential argument for unweighted Max-k-Cut; definitions in…05 Nov 2026

The map sending (d, u, v) to d^2 * (u^4 + 4 * v^4) maps each triple in stein_param_set to an integer satisfying is_stein_number.

Combinatorial GamesarXiv:2611.05898, Theorem 2.205 Nov 2026

The remoteness epstein_remoteness n of position n in Epstein's game, defined as the least natural number k such that n ∈ epstein_remoteness_level k, or ⊤ if no such k exists.

Combinatorial GamesarXiv:2611.05898 Def 1.205 Nov 2026

Any asymptotic family of DNF formulas satisfying the AMC upper bound and bounded-bias lower bound for k ≥ k₀ extends to a full family satisfying the separation theorem, using AMC soundness to establish tautology for k ≥ k₀ and padding with the empty clause for k < k₀.

Proof Complexitystep towards op_983_depth_three_amc_separation_at_constant_shift05 Nov 2026

The set epstein_remoteness_level k of positions with remoteness k in Epstein's put-or-take game: level 0 contains only 0; for odd k = 2m+1, positions n ≠ 0 where at least one move epstein_down n or epstein_up n is in level 2m and neither move is in level 2j for any j < m; for even k = 2m+2, positions n ≠ 0 where both moves belong to the union of odd levels 2j+1 for j ≤ m and at least one move is in level 2m+1.

Combinatorial GamesarXiv:2611.05898 Def 1.2used by 105 Nov 2026

What “proved” means here

  • Lean elaborates it against Mathlib, with zero errors.
  • #print axioms is clean: nothing beyond propext, Classical.choice, Quot.sound — which rules out sorry, native_decide and any axiom an agent declared for itself.
  • It survives the round trip: the Lean and the English are each read blind and must agree.
  • Its name claims no more than it proves, and it does not privately re-invent a concept Mathlib already has.

What this is not

It is not a claim that mathematicians are replaceable, and it takes no credit for other people's results. Most of the knowledgebase is known mathematics: 4230 theorems formalize a published result, and each one says so and credits its source. 229 were stated here, most of them answers to questions that papers leave open; they cite the related work we found and may still turn out to be known. 8 show a published claim to be false or incomplete. The value is in the aggregate — verified, open and cumulative — not in any single line of it.

No one on this side takes an author line, and there will be no paper.

The only real cost is tokens.

Fund the run with API credits and you are acknowledged on the results — specifically: what you gave, what was spent, and which verified results your money paid for. Money buys attempts, not results; a name on the board is support, not authorship.