Last week, I published an essay on this blog with my thoughts about the Circle Line’s new wayfinding system. You can — and should — read that first before continuing the rest of this one. Find the link below.

Completing the Circle

Yesterday, the 2 July 2026, the LTA made a change to the STIS — the TV screens — in response to online feedback. I’m not satisfied with the new layout, so I’m writing an additional piece — this should be a short one — with my thoughts about the latest version.

Here’s what it looked like before:

And here’s what it looks like now.

The most obvious change is the service direction label has shrunk, which I think is a welcome change for visibility and legibility of the other information. But it hasn’t come with a correction to the maps — again, 32 out of 34 stations have at least one platform’s worth of maps that incorrectly reflect the service direction, and with no other service direction labels anywhere in the navigation system, a commuter who needs to know which platform he needs will now actually have to be directly under the STIS to get the relevant information. Now, the commuter could refer to the single route diagrams elsewhere on the platform, but that’s going to be a time-wasting and deliberate move — it’s not as immediate as good wayfinding should be. This means the commuter most likely has to refer to the PSD maps to confirm which direction the trains on that platform are heading in, in which case, he’s going to be misled by those maps that are not grounded in reality.

To the LTA, seriously, fix those maps.

The loss of the direction graphic altogether doesn’t help either. While I felt it could have been simplified and perhaps made smaller — again, for legibility, the elimination of the icon from the STIS makes it much harder for commuters to understand from afar which service direction that platform serves.

The new layout also seems to have removed the data for the subsequent train from the arrival screen. That causes a big issue. If a commuter arrives at a platform while a train is already there, he won’t be able to tell if the train after is one he can take or has to take — depending on his destination — without a prior grasp of the line’s service pattern, and he won’t be able to tell when that subsequent comes. This has a few effects.

First, if the train already standing at the platform is one the commuter could take, then he could be incentivised to rush for it since he can’t tell when the next train is coming. This is extremely dangerous and could result in commuters getting trapped by the train doors or the platform doors and potentially getting injured.

Second, if it isn’t, then the commuter simply does not know if the subsequent trains will be ones he can take to his destination. And because the Circle Line is a loop, and either direction can reach any station, it also means the commuter will not know whether it is advantageous for him to take the other direction if he’s headed for a destination on the other side of the Circle.

This information is crucial. I recommended the inclusion of more trains’ data in my previous piece; not only has that not been done, that layer of information has been removed outright. Why?

Another noticeable absence (which can be gleaned from the LTA’s promotional materials) is the word “Loop”.

On the previous version, the inclusion of the “Loop” label was meant to help commuters to differentiate between services that go the full circle and non-loop services that service the Dhoby Ghaut branch, though the label hasn’t been used as of the time of writing since loop services have yet to begin.

The new version uses a strange and very much unintuitive way to distinguish between looping and non-looping trains.

Non-looping trains are simply reflected with “ends at [Final Destination]”, without any indication of the key stations along their route, removing any sense of directionality except for the now shrunken direction label and the dissonant maps.

Looping trains on the other hand are now reflected as “via [Next Key Station]”. The commuter would have to know beforehand that “via” means the train loops, and if he doesn’t, well, it could be the difference between ending up on the Dhoby Ghaut branch and continuing round the loop, especially now that the subsequent train’s data has been removed from the arrival screen.

The new STIS layout is unintuitive. They tried to fix a problem, they overshot, and now things have actually regressed. With the opening of Circle Line 6 fast approaching us, the LTA has now created even more problems to fix. While previously there were only readability issues, now the new layout will also create legitimate wayfinding issues. It’s ironic that the LTA has billed the latest changes as “easier”.

The Circle Line is finally a circle!

On 12 July 2026, three stations — Keppel, Cantonment, and Prince Edward Road — will begin revenue service, finally completing a missing link between HarbourFront and Marina Bay stations.

As the opening looms, a new issue has emerged — wayfinding.

Until now, the Singapore MRT’s wayfinding has been mostly linear, and the LTA has found a formula for that which works (somewhat, but that’s a story for another time). But the completion of the Circle Line presents some new challenges.

How do you represent a circular route on a space that works best for linear diagrams? And with some trains now making full loops, how do you show that? How do you tell the commuter which way round the loop is quicker to their destination? How does the commuter tell which platform they need? How do you reflect the Circle Line’s complicated service pattern?

These are all questions the LTA tried to answer in their redesign exercise, and they’re also the questions we’re answering in this piece by contrasting the Circle Line’s new wayfinding with some international examples.

But first, my thoughts on the latest debacle. Inner/Outer Loop or Clockwise/Anticlockwise? That’s the debate that has been ongoing, not just in the transit sphere, but also amongst the general public.

Both the Yamanote Line in Tokyo and the Osaka Loop Line in Osaka use the Inner/Outer Loop convention. Think of a loop line as two concentric circles. The Inner Loop is the track on the inside and the outer loop is the one outside. Japan, and indeed Singapore, drive and walk on the left, so the Outer Loop is the clockwise one and vice versa.

On the other hand, the Clockwise/Counterclockwise convention is used by the Nagoya Municipal Subway’s Meijo Line — which we’ll come back to later, and it’s also what the LTA chose for the Circle Line — they did a survey and this is the option that won out.

There are various pros and cons to both that have been mentioned by people — such as the younger generation not knowing what clockwise/anticlockwise means (?), but I’m not going to address those.

Because — in my opinion — it doesn’t matter.

The average commuter doesn’t need to know the names of the directions — the Japanese systems afford this information the tiniest bits of space on signages and electronic displays. They just need to know if the train makes a full loop and which stations are easier to reach using either direction. You could do away with the name entirely — Berlin’s Ringbahn simply calls theirs the S41 and S42.

But whichever convention you choose, be consistent. The chosen convention should be reflected across all touchpoints. If the service is named anticlockwise on signage, the maps should show it going anticlockwise, and vice versa.

Yeah. The LTA doesn’t seem to quite have gotten that memo.

You’re about to see photos of the maps at three stations — Bayfront, Promenade, and Dakota. There’s one commonality running through the three stations; let’s see if you can guess what it is.

Did you catch it?

Yeah.

At these stations, the anticlockwise service is represented clockwise, and the clockwise service anticlockwise.

And these are just the most egregious ones. The maps at the vast majority of the other stations represent either direction wrong. In total, of the 34 stations on the Circle Line — including the 3 new ones, only 2 stations have both directions’ maps right. Not by design, though.

Kent Ridge station just happens to be at the boundary between the half of the line where the clockwise maps are wrong and the half where the anticlockwise maps are wrong. The other side of the Circle, Dakota station, is one of the three with incorrect maps in both directions. Prince Edward Road, meanwhile, owes it to its stacked platform arrangement, with the Circle Line platforms coincidentally being on the side where the maps are correct. The other stacked stations, Bayfront and Promenade, have their platforms on the side where the maps are wrong.

So why? How did this happen?

Well, in their little redesign exercise, there was one little detail, or rather a self-imposed — they didn’t poll this, strangely — requirement, that the LTA did their work around. The PSD maps are designed to prioritise reflecting the actual direction of train travel. That’s fair; as an additional layer of confirmation, it helps commuters make sure that they are boarding the right train.

But the LTA’s answers in their FAQs seem to suggest that there exists a dichotomy between representing the service direction accurately and showing the actual direction of train travel relative to a commuter on the platform.

That’s only true because they’ve forced the Dhoby Ghaut branch to remain at the bottom of the diagrams. To preserve the direction of train travel, this decision forces two things:

1. the map flips about its vertical axis — the northern arc has one orientation and the southern arc has another.
2. the two maps at each station are mere mirrors of each other.

That’s what creates the incongruency between the map representations and the actual service directions.

But what if they didn’t do this?

This is where the Meijo Line comes in. It’s probably the closest parallel I could find to the Circle Line. Like the Circle Line, it features a branch — in the form of the Meiko Line — and it also uses the clockwise/anticlockwise convention.

The Meijo Line’s approach allows the Meiko Line branch to rotate with the rest of the map. The opposite directions’ maps are not mirrors of each other, they’re a 180-degree rotation about the center of the circle. And there aren’t any weird flips. By doing this, the Meijo Line’s maps are able to capture both the service directions and the travel direction accurately.

This is what the LTA should have done.

But there are also other lessons the Meijo Line has for us. In the introduction, I raised the questions:

How do you represent a circular route on a space that works best for linear diagrams? And with some trains now making full loops, how do you show that? How do you tell the commuter which way round the loop is quicker to their destination? How does the commuter tell which platform they need? How do you reflect the Circle Line’s complicated service pattern?

In a properly working navigation system, the signs must allow the commuter to decide correctly which platform they need at a glance. That requires several layers of confirmation: the direction, a map of the stations, and the major stations ahead. These different layers serve to reinforce and complement each other — each of these aspects alone must be able to quickly inform a correct decision, and the others must be able to just as quickly corroborate that decision.

There are several navigational elements that carry these layers of confirmation — maps, signs, digital aids. The need for urgency on a metro system means these elements must all be quick and easy to understand and parse.

I’ve already discussed the directional element in the section before this. Other than the whole maps-not corroborating-with-signage issue, there’s nothing significantly glaring about the LTA’s approach, though personally, this information could probably be afforded less real estate on the digital signage.

As for the maps, the LTA chose to include the full circle on theirs — both those on top of the platform screen doors and the single route diagrams located in the middle of the platforms. This is in keeping with the wayfinding on the other lines. But that’s the issue. Those lines are linear. It’s fine to keep the whole line on the maps with the previous stations faded out because trains in that direction won’t call at those stations.

But that doesn’t hold for a loop.

Instead, what’s faded out on the Circle Line maps are all the stations for which the opposing direction is faster. That means every map has half the stations faded out. Keeping the entire line up on the maps is a waste of space, and it creates a lot of unnecessary and distracting clutter on the space-limited wayfinding surfaces of the Circle Line stations.

When I was transferring to the Meijo Line to collect material for this post, I came across an interesting diagram at Sakae station. Here, the Meijo Line meets the Higashiyama Line (which is where I was transferring from) perpendicularly. What’s more, the Meijo Line platforms here, as with many other Japanese subway stations, take on a side platform layout. This means the transfer linkways to each of the Meijo Line platforms are located at opposite ends of the Higashiyama Line platforms.

To help commuters make sure they’re going the correct way, there’s this:

What they’ve done here is genius. Truncating away the stations that are easier to get to via the other way round the loop creates more space for more useful details, or even just for bigger fonts and better visibility, and at the same time, the little arrow at the end still establishes that the line does, in fact, continue past the highlighted stations.

Of course, the Meijo Line still has full maps. But they’re not located at the decision points that commuters have to crowd around to parse.

As you can see, the space these maps are located on are better suited for full line maps. But these maps also encode other forms of information — connections and travel time. They aren’t for on-the-spot decision-making, but for the commuter that needs to plan out the rest of their journey.

You’ll notice something with these maps, and that’s the notation of key stations along the route. This is important on a loop line where there’s no one destination where trains end their runs. They serve as another confirmation of the direction trains run in. Commuters are typically familiar with the major stations along their route — usually the ones they need to transfer at — so providing this information helps people make sure they are heading the right way.

Loop lines in Japan typically provide more than one key station in the various navigational touchpoints. As seen above, the Meijo Line provides two or even three, but Tokyo’s Yamanote Line provides up to 6. Having more than one key station listed not only increases the utility of this layer of information to commuters, but it also adds a sense of linearity — basically it helps commuters understand which order the major stations come in.

Contrast this with the Circle Line, which only provides one station — the next nearest interchange station. If a first-time commuter is at Bishan station and he needs to go to Promenade station, the signs only tell him the trains are heading via Serangoon in one direction and via Caldecott in the other. If he doesn’t already know where Promenade is, then this forces him to have to first consult a map, most likely the ones on top of the PSDs. One person doing this is fine, but many people having to do the same thing causes human flow problems on the platform.

So that’s the physical signage.

We also need to talk about the digital signage. This is the dynamic stuff, the part of the navigational system that reflects the service pattern and status.

I don’t know what the final service pattern will be, at the time of writing. But what I do know is that there will be loop and non-loop services sharing the Circle Line. The non-loop services feed into the line’s Dhoby Ghaut branch — the opposing end should be Prince Edward Road station, and this matters for the digital signage because one hiccup can mean a commuter ending up on the branch instead of carrying on through the loop.

To be clear, my issue with the digital signage — they’re called STIS — is a graphic design one. I’m not trained in graphic design or anything, but even I can tell a subpar UI apart.

There’s too much emphasis on the direction — I’ve already said this somewhere in this piece. But this isn’t just me nitpicking. It’s because the direction name takes up lots of space on the screens that could be used for other information, especially considering how relatively unuseful the direction name is to the commuter.

It’s not the only thing taking up excessive space, mind you. That weird icon that’s meant to symbolize the direction and key stations, which you can’t really even read unless you’re right underneath the screens?

That needs to go too. Simplify it, make it smaller.

What really needs to take up more space on those screens are: whether the approaching train loops, the listed key stations, and the time the train is expected in. That’s the information that matters to the commuter.

And the screens really need to be able to display more than two trains, since there are at least two possible types of service calling at each platform. Showing three subsequent trains means at least one of the service patterns will be repeated, allowing commuters to get a better grasp of the service frequency.

So that’s the Circle Line’s wayfinding.

The underlying concepts are well-intentioned, but the execution needs to be refined. A common theme through this essay is that there’s a lot of focus on things that really don’t deserve it, and that takes away from the information that actually matters. Wayfinding isn’t just about relaying information, but about encoding it in a way that is intuitive and immediate. A station platform is a hectic place, and poor wayfinding will only make it worse.

While all MRT stations in Singapore feature platform screen doors — underground stations have full-height doors and overground stations have half-height ones, LRT stations don’t, instead featuring partial barriers. The positions on the platforms where the doors of the trains (approximately) align to are open, and that means passengers can still be pushed or accidentally fall onto the tracks.

The partial barriers were installed beginning in 2017, completing in 2018. Since then, there have been 5 incidents of injuries or fatalities sustained from commuters falling onto the tracks. And with each incident, calls for the installation of proper platform screen doors at LRT stations intensify. And every time, the LTA’s response is:

LRT stations are smaller than MRT stations, and that installing platform screen doors would require more space for power, communications and signal control rooms. This would take up more waiting space from commuters.

- Dr. Amy Khor, 2023 (Retrieved from Channel NewsAsia, 2023)

But before we talk about why platform screen doors matter, let’s first talk about what they are. A platform screen door is a barrier, including some form of door, that physically separates a platform from the adjacent railway. In general, there are 2 kinds of platform screen doors. Full-height ones, which extend above a person’s height, usually to the ceiling or roof of the station, and half-height ones, which extend to approximately shoulder-level.

The Singapore MRT is actually a revolutionary system on this front, being the first heavy rail metro system to employ full-height platform screen doors across all its underground stations. The platform screen doors allowed the underground stations to be climate-controlled, something truly valuable in Singapore’s warm and humid Equatorial climate.

Half-height platform screen doors came to overground MRT stations between 2009 and 2012, following a series of high-profile incidents on the North South and East West Lines.

The installation of the partial barriers on the LRT began in 2017, after a man fell onto the tracks earlier that year and in light of expected increases in ridership. By 2018, all the stations across the Bukit Panjang and Sengkang-Punggol LRT had them installed.

But interestingly, more incidents — again, 5 — have happened after the partial barriers were installed than prior to their installation — one in 2000, and the one in 2017. Why?

I can’t find evidence explaining this correlation online, and maybe it’s because Google is getting trashy, but the trend is real, and it shouldn’t be ignored.

My theory is that the barriers make the platform edge feel more tame than it is. That creates a misperception that the platform edge — or areas near it — are safe to navigate. People stand closer to it, and with population growth, that means more people are on the platform and on the platform edge. That increases the likelihood of someone being pushed or slipping and tripping onto the tracks.

Conversely, without the barriers, it becomes obvious that the platform edge is a precarious place to stand at, and so people avoid it naturally and are more careful when they do have to be near it. That reduces the chance of accidental falls onto the tracks, at least compared to the partial barriers.

The partial barriers are not safe. Safer than just an open platform edge, sure, but they don’t prevent falls. And seemingly, they lure people into letting their guards down.

As the population density around the LRT stations increases, one thing becomes clear:

The partial barriers need to go. The Singapore LRT needs real platform screen doors.

So why doesn’t it? and what can be done?

Well, while I’m usually very critical when it comes to the LTA, their argument:

LRT stations are smaller than MRT stations

is actually right. A lot of people, both laymen and not, seem to think it’s a comment about platform space vis-a-vis the doors themselves, but that’s not the issue, or at least it’s not the angle the LTA is coming from. What they actually mean is that LRT stations don’t have the space to accommodate the supplementary equipment that’s required — to make sure the doors open at all, and to make sure they open when they need to and don’t when they shouldn’t. The MRT stations either had this space factored in or had plenty of space to retrofit it in. The LRT stations don’t.

What these stations will need is some lightweight PSD validation technology — to check that doors are aligned before they open, and that they are closed before the train can move off.

SMRT, the operator of the Bukit Panjang LRT, are already working on one that uses LiDAR and AI technology, which they showed off on a social media post dated 2 April.

As far as I can tell, the new technology gets rid of the need for cumbersome signalling boxes by allowing the trains’ onboard computers — thanks to the CBTC signalling system — to talk directly to another computer mounted in the PSDs to validate when they are supposed to open and close.

LiDAR systems, together with the existing iSafe AI cameras, validate that the train is stopped and aligned.

Simpler technologies already exist for that function, though.

Tokyo’s Asakusa subway line runs a mix of rolling stock thanks to its interoperation with the Keikyu and Keisei networks. These various trains have varying door placements. To ensure that the correct doors open, a camera scans QR codes printed on the side of the trains to instruct the relevant doors to open. A similar system could work on the BPLRT. A QR code or barcode printed on one door could be interacted with to make sure that the train is aligned properly before the doors open and close.

World's First Platform Doors Using 2D Codes Installed at Toei Subway Stations | The Government of Japan - JapanGov -

I’m not super well-versed in signalling matters, so I also ran that through Google’s Gemini. You can see its thoughts in the link below.

Google Search

The Sengkang and Punggol LRT lines, on the other hand, don’t have CBTC, and so their trains aren’t able to communicate with PSDs directly the same way. Without the CBTC handshake, there’s no way to ensure the train doors are aligned to a particular tolerance. But what if they didn’t need to be? What if the whole platform edge was one big PSD?

This is where rope-type PSDs come in.

You might have seen these in Japan, or in videos of Japan, where they’re used for platforms with varying train types. They’re those ropes that get pulled up when a train arrives at a platform. They’re light, they’re cheap, and most importantly, they don’t require precise alignment to work.

Paired with a LiDAR technology that can detect when a train is stopped at a station, rope-type PSDs could be a solution for the Sengkang and Punggol LRT lines.

But maybe there’s an even more obvious solution — man those stations. Have someone constantly on the watch for people who accidentally (or not-so-accidentally) fall onto the tracks. In Japan, unmanned systems like the Sengkang and Punggol LRT must have platform screen doors. Without those doors, the Sengkang and Punggol LRT stations need to be manned to be safe.

So there are steps that can be taken. But the LTA shows no sign of taking them, and while both operators have deployed AI-based intrusion detectors, it’s currently unclear — as investigations are proceeding — if SMRT’s was working when the incident at Segar station happened. I wouldn’t put all my eggs in the AI intrusion detection basket.

As our LRT platforms are much smaller than MRT platforms, there is insufficient space to install equipment and cabling to support platform screen doors that can operate in sync with train doors, without substantially affecting the space available for commuters. — LTA, 22 May 2026, as quoted in CNA, The Straits Times, and Mothership.

Frankly, I’m disappointed. A life has been lost. But the LTA’s latest response: almost exactly the same words that SMS Amy Khor said 3 years ago.

Transit topology is an important aspect of transport planning. It refers to how a network is laid out, and it has implications on the network effects that a transit system enjoys. Today, we’re talking about two main classes of transit lines. We’re going to look at some examples of orbital lines, and we’re also going to talk about how radial and orbitals fit into a transit network. This essay will hopefully help to define radials and orbitals for future reference. I’ve already used these terms a few times in previous essays, but I’ve decided to explain them in full for those that aren’t in the know.

Let’s start with radials.

Radials (from the word radius) describe transit routes that start in a city center and extend out to suburban areas. Cross-city routes that run between suburban areas via the city center are also radials — they can be thought of as two radials combined.

Radials represent the fastest way from a location to the city center. Because suburban areas tend to be residential and city center areas tend to be commercial in nature, radial lines are often responsible for moving commuters during the morning and evening peak hours. They therefore experience intense use during the peak and less use during the offpeak.

A key defining factor of the traffic along a radial line is that it overwhelmingly flows in one direction during peak periods. In the morning, most commuters are heading towards the city center, and in the evening, most commuters are heading away from the city center. We can call this phenomenon the “peak direction”.

Because radials experience the highest use during peak periods, this is also when the most capacity on the line is mobilised. This is done by running trains closer together, and therefore running more trains.

Increased capacity during peak periods isn’t unique to radials — all transit routes experience the same peak periods, but it does cause a problem that is only seen on radials. The core of the problem is that a train heading towards the city center (and vice versa) full of commuters must also return whence it came empty. This is what is known as “reverse peak capacity”.

Reverse peak capacity can be simply understood as the capacity of a rail line that exists in the direction opposite to the peak direction. Because this is capacity that often goes unused, it is considered to be wasted. Urban planning strategies such as decentralisation often target reverse peak capacity, mobilising it to move people and increase the efficiency of the transit network.

Another key dampener to the reverse peak capacity problem is designing radial routes to also include centers of employment or destinations that are typically far away from the city center, such as ports and airports. London’s Piccadilly Line is an example of this.

While radials are the main engines of a transit network, a network that only consists of radials quicky becomes inefficient as the city grows.

In such a network, a commuter travelling from one suburb to another always has to pass through the city center, creating long and cumbersome journeys, and taking up valuable space that could be used by commuters closer to the city.

To alleviate this, most transit networks have orbitals. An orbital is a transit route that excludes the city center. They can be arcs, circles, semi-circles, and more loosely, chords that exist entirely outside the city center.

The main function of orbitals is to link radial routes together outside the city center, creating a way for commuters to travel from one suburban area to another without having to go all the way into the city center and back out.

Orbitals are a valuable tool in urban and transport planning. They make it easier for a city to decentralise by creating nodes outside the city center. They create valuable transport capacity within and around the city center, and also help make employment centers outside the city center more attractive.

Because orbitals link radials together, they experience passenger loads differently. The load pattern on an orbital is dominated by passenger exchange as commuters get on and off as they move between the radials. Thus, while orbitals experience smaller loads at any one time, they are often able to accommodate more trips overall.

One of the busiest rapid transit lines in the world, the Yamanote Line in Tokyo, is one of these orbitals. If the figures from JR East are to be believed, the Yamanote carries around 5 million passengers every day. The line’s high ridership is reflected in the service it runs, with an 11-car train with 5 doors per side per car arriving every 3–4 minutes all day.

Originally created as a bypass to Tokyo Station and then closed into a loop, its presence helped to spur on the growth of districts such as Shinjuku and Shibuya in the wake of the 1923 Great Kanto earthquake.

The Yamanote’s positioning as a ring encircling the central areas of Tokyo is undoubtedly a major contributor to its success. In the prewar era, the Ministry of Railways maintained a policy of not allowing private railways to construct railway lines within the Yamanote Line loop. This meant private railways terminated at stations along the Yamanote Line, turning those stations first into major transport hubs and then their surroundings into major urban centers.

It’s not a coincidence that the Yamanote Line passes through almost all of Tokyo’s major commercial districts, and it’s not a fluke that some of the busiest railway stations in the world — including the top three — are all on the Yamanote. It’s perhaps not an exaggeration to say that places like Shinjuku and Shibuya are only as bustling as they are today because of the Yamanote Line.

In this vein, orbitals are often the centerpiece of a city’s transit network. Many cities center their transit maps around their orbitals, and many people view their city’s orbitals as the boundary between the prestigious city center and the urban fringe. In other words, orbitals shape the way a city and its neighbourhoods are viewed.

New transit projects in mature transit systems are often focused on creating on orbital connections. Examples include Taipei’s Circular Line, the Greater Copenhagen Light Rail, London’s West Orbital, Paris’s Lines 15, 16, and 18, and Singapore’s very own Cross Island Line.

The Cross Island Line will be Singapore’s most influential railway project yet.

It will form an outer semicircle, featuring wider stop spacings and a higher top speed than the parallel Circle Line. As an express version of the Circle Line, it is expected to form an effective relief for the perenially overcrowded line. The first heavy-capacity line since the North East Line in 2003, the Cross Island Line will feature trains with 6 cars, each equipped with 5 doors per side, with provisions at stations for the trains to eventually expand to 8 cars as demand climbs.

The Cross Island Line is designed for capacity, something that was neglected during the planning and construction of its orange predecessor. It is an example of how to do an orbital right; its orange neighbour, not so much.

What the Cross Island Line and others before it proves is that the shape of a transit route is an important factor in designing its service levels, infrastructure, and equipment. It shapes (pun intended) how a transit route experiences demand, and therefore the kind of design that it needs to meet demand.

Categorising transit routes strictly into orbitals and radials is one way of understanding transit topology, but these are very broad categories, and really, we’re barely scratching the surface here — a radial-orbital transit system is but one kind out of many. The insights in this article should still form a basis for understanding all forms of transit networks — from the Complex Grid of Osaka to the Soviet Triangles of the Soviet metros.

The early 1980s were a heyday for Japanese horseracing. In 1983, Mr C.B. became the third ever to achieve the Japanese Classic Triple Crown for Colts — 19 years after the previous Triple Crown, Shinzan. But the next Triple Crown would not take 19 years. In an unprecedented show of dominance, a colt named Symboli Rudolf became the 4th Triple Crown in 1984, and the first to accomplish the feat while remaining undefeated.

Symboli Rudolf would go on to continue dominating the racing scene, ending his career with a total of 7 Grade 1 wins — a record that remained unbroken until 2020. This era would incite great excitement in the racing scene, as unprecedented feat after unprecedented feat was achieved.

But the high would not last.

As Japan headed into the late 80s, the vitality of the horseracing scene would decline. Now seen as “only for gamblers”, it would fail to attract interest from young blood.

All that began to change on 19 May 1987, when a grey horse named Oguri Cap steps into the gates of Kasamatsu Racecourse.

Many people have already told the story of Oguri Cap’s career. So instead, we’re going to focus on his influence on the Japanese horseracing scene. Unlike a lot of other famous Japanese racehorses, Oguri Cap’s story begins in the regional circuit. As a grey horse — at the time believed to be unable to run well (yes, horse racism is a thing), expectations were not high for him to become a successful racer. Coupled with his unimpressive pedigree and his leg troubles growing up, he was deemed to be rather unappealing as a foal, and would be brought, not into the prestigious national circuit, but the then-unglamorous track of the small town of Kasamatsu in Gifu Prefecture.

Oguri Cap’s career in the regional circuit can be summed up with one word: dominance. Winning 10 out of 12 races — only losing twice to a horse named March Tosho, Oguri Cap would make it to the central circuit in 1988.

His first test in the central circuit would be the Grade III Pegasus Stakes, held at the Hanshin Racecourse on 6 Mar 1988, where he crosses the line 3 lengths ahead of second place, Rugger Black.

After his next win, the Grade II Mainichi Hai, held on the same track on Mar 27 1988, expectations from fans and the industry are high for him to contend for the Classic Triple Crown.

The Classic Triple Crown is the highest prestige in Japanese horseracing. Designed to test a horse’s abilities over various tracks, distances and conditions, the Triple Crown consists of the Satsuki Sho, Tokyo Yushun (Japanese Derby), and the Kikuka Sho. These races are only open to horses in their classic year, meaning contenders only have one chance to compete for the title.

Throughout Japanese horseracing history, only 8 horses have achieved the Classic Triple Crown. Saint Lite in 1941, Shinzan in 1964, Mr C.B. in 1983, Symboli Rudolf in 1984, Narita Brian in 1994, Deep Impact in 2005, Orfevre in 2011, and Contrail in 2020. At the time of Oguri Cap’s entry to the central circuit, it had been 4 years since Symboli Rudolf had claimed the title. So when a horse as dominant as Oguri Cap showed up, many believed he had a chance to capture the title.

But there was a problem. Oguri Cap’s team had never registered him for the Triple Crown races, and they were now long past the registration deadline. Public outcry, from fans and media alike, ensued. How could the Classic Triple Crown, meant to determine the strongest Japanese horses, not include the one many felt had already reached the top?

But the Japan Racing Association, the government agency that is responsible for the central horseracing circuit, refused to budge. In the end, Oguri Cap would not run the Classic Triple Crown. This detail will become relevant later.

Oguri Cap continues to prove his worth throughout the summer of 1988. He sets new race records in the Grade II New Zealand Trophy and Takamatsunomiya Hai, before claiming yet another victory at the Grade II Mainichi Okan against the 1985 Japanese Derby victor Sirius Symboli.

Oguri Cap’s next bout is the Grade I Tenno Sho (Autumn), his first challenge for a Grade I title, and his first meeting with fellow gray horse, Tamamo Cross. Tamamo Cross was on a 6-win streak, including a win at the Grade I Tenno Sho (Spring). It’s not an exaggeration to say that he was Japan’s strongest racehorse at the time.

Unfortunately for Oguri Cap, Tamamo Cross emerges victorious, becoming the first to achieve both the Spring and Autumn editions of the Tenno Sho in the same year.

It would not be the last time the two met, with Oguri Cap placing behind Tamamo Cross and the American horse Pay the Butler in the Japan Cup, and then finally claiming his first Grade I victory at the end-of-year Grand Prix, the Arima Kinen.

In a time when the superstition that grey horses were not good racers existed, the competition between Oguri Cap and Tamamo Cross helped to shatter the belief. The electrifying rivalry between the two silver lightning bolts once again renewed fan interest. The JRA named Tamamo Cross the Horse of the Year and Best Older Male Horse, while Oguri Cap was awarded Best Three-Year-Old Colt.

As Japan stepped into the Heisei era, Oguri Cap would be met with a setback. An injury forces him out of commission for most of the year, and he returns to the turf on 17 September 1989, at the Grade III Sankei Sho All Comers, where he sets yet another race record.

The Heisei era brings with it two new stars: Super Creek, the 1988 Kikuka Sho winner, and Inari One, who, like Oguri Cap, had been promoted from a regional track — in this case, Oi. These three horses would compete intensely throughout 1989, and would earn a reputation as the “Heisei Big Three”.

In subsequent races, Oguri Cap is immediately thrown to the test against the Big Three, narrowly defeating Inari One to claim a 2nd win at the Mainichi Okan, and then finishing second behind Super Creek in that year’s edition of the Tenno Sho (Autumn).

It’s at this time that Oguri Cap’s current and third owner, perhaps troubled by his long downtime due to injury, makes a shocking announcement. Oguri Cap would be entered in the Mile Championship, held on 19 November 1989, and the Japan Cup, held on 26 November 1989. Two Grade I races, a week apart.

At the Mile Championship, Oguri Cap wins against the miler Bamboo Memory after an intense battle of speed.

The 1989 edition of the Japan Cup pitted Japan’s Oguri Cap and Super Creek against a star-studded field that included that year’s winner of the Prix de L’Arc de Triomphe, Caroll House, and the then-turf 2400m record holder, Hawkster.

This race would come to be known as “the 2:22:2 incident”. Under the lead of frontrunner and European champion Ibn Bey, the pace of the 1989 edition of the Japan Cup would shatter the Japanese records for distances of 1800m, 2000m, 2200m. As the pack rounds the 4th corner, the frontrunners tire out, but the New Zealand mare Horlicks continues to lead. Against the thunderous cheers of the Japanese fans, their star Oguri Cap rushes out from the pack in a desperate attempt to catch her, but he narrowly misses the lead by a fraction of a second.

Under the high pace of this Japan Cup, Horlicks sets a new world record for 2400m on turf, coming in at 2:22:2. There’s no doubt this is Oguri Cap’s best race, one where he showed off his monstrous speed to almost claim the world record, despite having run another major race just a week prior.

Perhaps exhausted from his prior races, Oguri Cap performs poorly at the Arima Kinen that year, placing 5th, the lowest in his career thus far.

For his achievements in the year 1989, Oguri Cap is awarded the JRA Special Award.

Entering the year 1990, Oguri Cap’s popularity only continues to grow. After a period of recuperation, he returns to the racetrack by setting a new race record at the Yasuda Kinen. But his performance only declines from here.

He finishes the Takarazuka Kinen in second, the Tenno Sho (Autumn) in sixth, and the Japan Cup in 11th.

Despite that, the year 1990 marked the peak of the “Oguri Boom”. His second owner, who had been disqualified from the JRA because of tax evasion, had trademarked his name and likeness, and had begun producing Oguri Cap plushies. In total, it is estimated that 11 million of these plushies were sold — equivalent to 9% of the Japanese population.

Unprecedented for Japanese horseracing, Oguri Cap, perhaps because of his natural airheadedness, had immense popularity amongst young female fans. The phenomenon of the “Oguri Gal” — young girls that appeared at the racetrack to support Oguri Cap — is well-documented.

Coupled with his story as a hero from the countryside and his sheer dominance in his career, Oguri Cap helped to massively renew interest in the Japanese horseracing scene. Even today, if you ask anyone from that era — even if they weren’t involved in horseracing — if they’re familiar with any racehorses, there’s a good chance they will bring up Oguri Cap.

His career, and his growing popularity, would culminate in the 1990 Arima Kinen. After his poor performance at the Japan Cup, many fans believed his prime was over, and fearing a potentially life-ending injury, demanded his owner to retire him from racing, even going as far as sending death threats. Despite the outcry, the owner decided to run Oguri Cap in one last race.

The 1990 edition of the Arima Kinen would feature rivals such as Mejiro Ryan — third place at that year’s Kikuka Sho, White Stone — who placed second in that same race, Osaichi George — who Oguri Cap had lost to in the Takarazuka Kinen, Mejiro Ardan — second place at the Tenno Sho (Autumn), and Yaeno Muteki — first place at said Tenno Sho. Paired with the new star jockey Take Yutaka, Oguri Cap would enter the gates of the Arima Kinen as the 4th favourite to win.

But the betting odds were in no way a reflection of the amount of supporters that had gathered to see Oguri Cap’s retirement run. Nakayama Racecourse, where the Arima Kinen is held every year, is rated for 160,000 spectators. On the 23 December 1990, it is estimated that 177,779 people showed up — a record still unbroken to this day.

Oguri Cap, against all odds, would finish in first, achieving a 4th Grade I victory. Many had showed up that day to simply watch him run, but few had expected that the monster of the past would win one last time.

神はいる。そう思った (JRA, 2011)

While it is customary for the audience to celebrate the winning jockey, for the first time in history, the audience at the 1990 Arima Kinen called out Oguri Cap’s name. A thunderous “Oguri Call”, forever immortalised in the memory of Japanese horseracing.

The entire stands shook with cheers before the finish line, and I will never forget the thunderous Oguri call that followed, sung by 177,000 people. As a horse racing professional, I consider it a lifelong treasure to have been able to witness that moment.

— JRA Chairman Takeyuki Tsuchikawa

For his achievements, Oguri Cap is named Horse of the Year and Best Older Male Horse by the JRA, while the National Racing Association — who manage the regional racing circuit — award him their NAR Grand Prix Special Award.

In 1991, Oguri Cap is inducted into the JRA Hall of Fame, a recognition of his ability and his contribution to the continued longevity of the horseracing scene.

While other star racehorses typically have one or two retirement ceremonies, Oguri Cap’s fans would get three opportunities — in Kyoto, Tokyo, and his hometown of Kasamatsu — to bid him farewell as he left the racing world. It’s recorded that the ceremony in Kasamatsu increased the town’s population from 20,000 to 40,000 that day.

In 1992, the JRA finally budges on late registration for the Classic Triple Crown, establishing a rule that will benefit late entries, albeit at a much higher monetary penalty.

The “Oguri Cap Rule” would be used by two of Japan’s subsequent racing legends. T.M. Opera O would win the 1999 Satsuki Sho, and then go through the year 2000 season undefeated, including a Grand Slam of the senior class Grade I races — an unprecedented and unseen since feat that earned him the nickname “Conqueror of the Century’s End”. Kitasan Black would place 3rd in the 2015 Satsuki Sho and win the Kikuka Sho, going on to finish his career with 7 Grade I wins, including a world record time at the 2017 Tenno Sho (Spring).

After retirement, Oguri Cap is transferred to the Yushun Stallion Station to begin his life as a stud, but he was never able to produce offspring of his caliber and was retired as a stud in 2007.

On 3 July 2010, Oguri Cap suffers a fracture to his tibia in an apparent pasture accident, and is subsequently euthanized, aged 25.

He is remembered through statues made in his likeness at the Yushun Stallion Station and at the Kasamatsu Racecourse — who to this day organize a race titled the Oguri Cap Kinen yearly — and a museum at the Tokyo Racecourse detailing his achievements.

Today, the 27 March 2026, marks 41 years since Japan’s greatest star was born. His legacy lives on in the thousands of racehorses that grace Japan’s tracks — it’s often said his popularity helped to revive Japanese horseracing.

Since 1981, the JRA has produced a line of “Hero Posters” to commemorate its star racers. While many of these feature grandiose taglines, Oguri Cap’s poster simply says “ありがとう” (lit. “Thank You”). At the bottom, the following poem (which I’ve translated) can be found:

駆けた全4万1100m。燃えた全41分16秒6。

あの有馬記念、あのマイルチャンピオンシップ。無念の天皇賞、ジャパンカップ。そして復活の有馬記念。

数々の感動と記録を残して怪物はターフを去っていった。

心に刻み込まれた栄光と惜敗の全ドラマ。その感動は決して消えはしない。

オグリキャップ。これからも、そしていつまでも、その名は心のなかで永遠に走り続けて行くだろう。

ありがとう。夢を、感動を、ありがとう。

The 41100m you ran, the 41:16:6 you burnt through.

That Arima Kinen, that Mile Championship. The regret of the Tenno Sho, of the Japan Cup. And then that Arima Kinen where you revived.

Leaving us with numerous emotions and memories, the monster has left the turf.

But the drama, of your glory and defeats, remains forever engraved in our hearts. We will never forget those emotions.

Oguri Cap. Now and forevermore, that name will forever remain in our hearts.

Thank you. For the dreams, for the emotions. Thank you.

Thank you, Oguri Cap. We will always remember you.

Author’s Note: I was originally going to leave this unlisted, but I also want Oguri Cap’s story to be spread as far and as wide as possible. Will return to transport stuff from the next post.

There’s been much excitement over the RTS Link as we get closer to its opening date on 31 December 2026. And understandably so.

The RTS Link is supposed to revolutionise cross-border travel. It‘s been touted as the world’s first cross-border rapid transit system, and is meant to help relieve the Causeway, particularly needed with the creation of the Johor Bahru — Singapore Special Economic Zone.

The new cross-border railway starts at Woodlands North station, where it meets the Thomson-East Coast Line, and crosses the Straits of Johor to reach Bukit Chagar station, which will be a short 400m walk via an elevated walkway from Johor Bahru Sentral station — Johor Bahru’s main intercity station. Both RTS Link stations will feature integrated CIQ facilities, meaning passengers will be able to clear both nations’ Customs and Immigration at their point of departure. That will certainly make it more convenient to travel between the two cities.

But how transformational will the RTS Link truly be? What are the challenges that it will face in trying to attract ridership? Is the RTS Link truly as great as it sounds?

There are a few flaws with the RTS Link operation in its current form, but these are all things that can be fixed after it opens and the operating company has warmed up.

For starters, the line’s carrying capacity has come under scrutiny. Using four-car trainsets similar to those used on Kuala Lumpur’s future Shah Alam Line, the line will run a 5-minute headway in the offpeak — ramping up to 3.6 mins in the peak. With each train rated for 1,069 passengers, the RTS Link has been touted as being able to carry 10,000 passengers per hour per direction.

Sounds like a lot, but the Malaysia-Singapore border is the busiest land crossing in the world. Woodlands Checkpoint alone processes more than more 350,000 people every day.

According to the operator, the RTS Link will operate from 6 am to midnight daily, slightly shorter than the typical metro system. Over the 18-hour operating period, the RTS Link will be able to carry a maximum of 10,000 * 2 * 18 = 360,000 people.

But that’s not how the real world works. The vast majority of travellers using the Causeway are workers from Johor Bahru commuting into Singapore in the morning peak periods, and out during the evening peak. That means the RTS Link will realistically only make a dent of about 30,000 people each way — assuming 3-hour peak periods.

Capacity wise, the RTS Link is severely lacking. While it is still projected to reduce Causeway traffic by up to 35%, the brunt of the burden will still be on public buses, and on private transport. But this is an issue that can easily fixed by… just running more trains. With a 3.6 min headway and 5 min journey time, and assuming a 2 min dwell time, the RTS Link will need only 2 trains to run a full service.

They have 8.

Theoretically, they could run a sub-one min headway, though the signalling system will likely limit it to 90s, as with most modern metro systems. With a 90s headway, the RTS Link would be able to carry around 40,000 passengers per hour per direction. That’s a significant improvement.

Another obstacle, in my opinion, may be the fares. Malaysian Transport Minister Anthony Loke has indicated they will be between $5 and $7 both ways.

This is a detraction from past precedent.

On the KTM Shuttle Tebrau, fares have been $5 for bookings made in Singapore and RM5 for bookings made in Malaysia. The RTS Link fares, on the other hand, don’t seem to account for purchasing power parity, as the same fare is charged regardless of where you’re coming from.

$5 to $7 doesn’t sound like an incredulous amount of money, but at today’s exchange rate, that’s around RM15.40 to RM21.57. While these fares have rightfully been pointed out to be cheaper than driving into Singapore, they’re a far cry from the fares charged on public buses between the two countries. So it remains to be seen how the RTS Link will affect demand on bus services across the Causeway.

That being said, as with service frequency, these fares can always be adjusted after operation begins.

What will require far more political will to solve than these issues are the connecions on either side. The RTS Link cannot succeed as a standalone project. It must be supplemented by effective public transport on both sides of the border to be successful. That also means the quality of public transport on either side of the border will be a major sore point. Right now, public transport in Johor Bahru is… skeletal, to say the least. And even on the Singapore side, travel time and reliability will be a concern.

Despite its sizeable urban population of 858,000 and region population of 1.8 million, public transport in Johor Bahru consists solely of just regular buses. For comparison, its population, measured both ways, puts it ahead of Zurich, Switzerland — which has an S-Bahn and a tram system. For its size, Johor Bahru absolutely needs a rapid transit system, or the RTS Link will just be sending people to JB to get stuck, either in traffic congestion, or for buses that will get stuck in traffic congestion.

The silver lining — it is getting one. City officials recently revealed their Johor Bahru Draft Local Plan 2035, including plans for a city-wide transport network. The first of these routes is already in operation, being part of the Electric Train Service to Kuala Lumpur Sentral.

The second service, the Komuter Selatan, will operate on the same tracks, functioning as a local service. It was meant to open in February 2026, but I can’t confirm if it has begun operations at the time of writing, but I welcome any readers to confirm this for me.

As part of these projects, new stations will be added along the corridor to improve local access. In the nearer future, we can also expect frequency upgrades to the ETS service as double-tracking works are completed further up the line.

Another exciting project under this plan is the Komuter service from Tanjong Pelepas to Pasir Gudang. Both of these are major centers of employment, being the main port and the main petrochemical hub respectively. And notably, this route already exists as a freight-only line, and will only require some upgrades and station-building to be fit for passenger use, meaning it can be built cheaply and much quicker than the other components of this proposal.

The RTS Link obviously needs no introduction, so we can move on to the remaining 4 services.

The Skudai, Iskandar, Pasir Gudang, and Tebrau Transit Corridors are perhaps the most interesting components of the plan.

Under the plan, these routes will begin life as a Bus Rapid Transit system, featuring dedicated bus lanes that will allow buses to fly past Johor’s horrible traffic.

In the long term, the plan is to eventually build an elevated Autonomous Rapid Transit network, creating a “metro” system for the city and the wider region. The development model here is interesting. The BRT system will function not dissimilarly from pre-metro systems in Europe, providing transit service even while the ART infrastructure is under construction.

I’ve seen a post on social media proposing something similar for TEL5/DTL3e while those projects are under construction, and I think it’s a decent idea going forward — though not necessarily for TEL5/DTL3e specifically.

And of course, the ART system could eventually evolve into a true rail system, though I will say this is not likely — the ART tech does have benefits that true rail cannot enjoy.

The overall plans look sound — there are no complaints I can make about the plans themselves — there’s even an airport rail link (which you know I love), but the implementation and operation will be a different story. Johor Bahru could be looking at a decent, high-capacity rapid transit network if this is done right.

Johor Bahru is working with a clean slate here, and that means they can easily build their “ideal” transit network. But that won’t hold true for Singapore, where we must deal with the limitations of the existing network even as we plan for the future.

Perhaps the biggest limitation of the Singapore network is connectedness. Woodlands North RTS station will be at the heart of a new business and commercial district, meant to help local firms position themselves to benefit from the Johor-Singapore Special Economic Zone.

It will be connected to the rest of the Singapore network by means of the Thomson-East Coast Line. The TEL connects Woodlands North to destinations such as Woodlands Town Centre, Orchard Road, the CBD, and Changi Airport. It also brings commuters to key transport arteries, such as to the North East Line for the Punggol Digital District, or the North South Line for the Jurong Lake District.

But a lack of other connections to Woodlands North means passengers will be funnelled through the TEL section between Woodlands and Woodlands North, demanding high capacity on this section. Many trips will require complicated, transfer-heavy routes through the MRT system, and will certainly suffer from long travel times.

And with the TEL being one of Singapore’s worst performing MRT lines, there are also resiliency concerns if Woodlands North has to rely solely on the TEL. After all, when — not if — the TEL goes bust, people will be stuck at Woodlands North without an alternative.

Woodlands North cannot just be a forced transfer point for travellers between Malaysia and Singapore. Like JB Sentral on the other side of the Causeway, Woodlands North must become a transport hub in and of itself.

There are no concrete plans at the time of writing for more connections to Woodlands North station, though the Land Transport Master Plan 2040, published in 2019, alludes to the Seletar Line serving Woodlands North.

I’ve written my own proposal on the Seletar Line and North Shore Line that will improve transport access from Woodlands North to Yishun, Punggol, Seletar Airport, Pasir Ris, and Changi Airport, and my MRT wishlist (you can find this in my profile) also includes a proposal to extend the Downtown Line to Woodlands North — I’ll probably flesh this out soon. With these proposals in place, journeys to the east and west will become more seamless, with enhanced connections to other MRT lines.

The RTS Link is certainly different, and while it may yet revolutionise cross-border travel, its long-term success will hinge heavily on the development of public transport networks on either side of the border. While many of the improvements required are a long way away, they will almost certainly be powerful, albeit delayed, in making the RTS Link truly transformational.

It’s the Year of the Horse, so I thought I’d cover a relevant topic today. While my past pieces on this blog have been more focused on the transport aspect of urbanism, this time I want to explore the history and future of an important chapter of Singapore’s story.

So why am I writing about horseracing? It’s certainly a bit strange, from a blog that has been yapping about trains all this time.

Well, like many others, I’ve been hit by the Umamusume virus. Cute anime horse girls led me into a rabbit hole that now includes obsessing over past Japanese horse races.

It made me realize that Japanese horseracing is very popular. Racecourses there are more than just that. They’re a major third place — a place for socialisation. Parents bring their kids to the track for fun days out — Tokyo Racecourse even has a Shinkansen-themed attraction for kids. Every racing weekend, tens of thousands of people flock to the major racecourses to watch these majestic animals pitted against each other.

And as with any other sport, people form emotional connections with the horses they watch at the racecourse. Throughout modern Japanese history, there have been countless racehorses loved for their abilities, their stories, or simply as a source of hope for people.

At its heart, Umamusume brought to a global audience the emotional stories behind Japan’s star racehorses.

Oguri Cap, the “Idol Horse” whose unprecedented and unseen-since popularity revived Japanese horseracing.

Haru Urara, whose 113 consecutive losses brought her to reverence as the “Shining Stars of Losers Everywhere”.

Silence Suzuka, the “Otherworldly Runaway” who ran so fast his legs and life shattered on the racetrack.

Rice Shower, the “Jet-Black Assassin” whose victories were met with scorn and punished with tragedy.

The duo of T.M. Opera O, the “Overlord of the Century’s End” who finished the year 2000 undefeated and became the first — and still the only — to achieve a Grand Slam of Japan’s senior class Grade 1 races, and Meisho Doto, his challenger who came second to the Overlord in five G1 races and finally achieved triumph at the Takarazuka Kinen of 2001.

All these stories came to life, and a new generation was exposed to them, learnt them, and were once again inspired by them.

The global release of the Umamusume mobile game, together with the airing of the franchise’s pièce de résistance, Cinderella Gray, led to a widespread growth of interest in the horseracing sport beyond Japan. In Malaysia and Thailand, Umamusume fans attended races dressed in cosplay. In Indonesia, the release of the Umamusume game and the resulting explosion in horseracing fan numbers made it in time for the new community to witness and celebrate their 4th ever Triple Crown horse, King Argentin.

That wasn’t the case in Singapore. While the Umamusume explosion happened in mid-2025, Singapore’s last race was held in late-2024, with the Singapore Turf Club confirming it will return the land that Kranji Racecourse sits on to the government for redevelopment by March of 2027. This meant Singapore’s horseracing scene never experienced the renaissance that other countries’ did.

The closure of Kranji Racecourse, marking the end of the sport here, gave me one question:

What are the stories in Singaporean horseracing that no longer have a platform to be told to the new generation?

And what now? What’s next for Kranji Racecourse?

Like other Asian countries, horse racing arrived in Singapore by means of the British.

For the 10 years after their arrival, informal races were held on the Field (now the Padang).

A History of The Padang

Formal horse racing began on 4 October 1842. Singapore’s first racecourse, Serangoon Road Racecourse, would be located at Farrer Park, which opened in December that year. Over the course of its history, it witnessed the creation of some of Singapore’s most prestigious racing events. In 1880, the first-ever Singapore Derby was held here. In 1924, the first-ever Singapore Gold Cup, a domestic Grade 1 race, was run here.

Serangoon Road Racecourse lasted for almost a hundred years. Today, you can still see clues of the area’s past, with road names such as Race Course Road and Race Course Lane.

Serangoon Road Racecourse closed in 1933. It was replaced by the new Bukit Timah Racecourse, located on a large 140 hectare site in the affluent Bukit Timah district.

In 1972, Queen Elizabeth II visited the racecourse while on a state visit to Singapore. In her honor, the Grade 1 Queen Elizabeth II Cup was established. In 1981, Bukit Timah Racecourse bore witness to history when Irene Pateman and Paula Wagg from Australia became the first female jockeys to be granted local permits to ride. In 1991, the racecourse saw the advent of the Grade 1 Raffles Cup.

Bukit Timah Racecourse moved in 1999, with the land it sat on being redeveloped for residential use, replaced by the new Kranji Racecourse. The old grandstands became the commerical development Turf City, which itself closed in 2023 to make way for the construction of the eponymous station on the Cross Island Line. The site will be further redeveloped into a housing estate.

The new Kranji Racecourse opened on 4 March 2000. That same year, Singapore launched its Singapore Triple Crown series. It would also be the birthplace of key feature races such as the Kranji Mile, Singapore Airlines International Cup, and the KrisFlyer International Sprint.

Able to seat 30,000 people, Kranji Racecourse would be the largest facility to date.

It is during this chapter of Singapore’s racing history that its greatest racehorse emerged. Little information is available online about Singapore’s racehorses, but Rocket Man seems to be the sole exception. Despite a career riddled with injury, he managed 20 wins — including the Grade I Dubai Golden Shaheen — and 5 placings, plus a 4th place finish at the Grade I Sprinters’ Stakes in Japan, out of 27 starts. He is often regarded as Singapore’s strongest.

Rocket Man, Singapore's most famous horse, calls it a day

Edit: Rocket Man passed away in South Africa on Mar 10, 2026, from colic. He was 21. We mourn the departure of one of Singapore’s greatest racing legends.

Singapore racing icon Rocket Man dies

But the history of horseracing in Singapore would soon come to an end. On 5 June 2023, the Singapore government announced that the Singapore Turf Club will cease its operations and return the land used for the racecourse for redevelopment.

Kranji Racecourse held its last race, the Grand Singapore Gold Cup, on 5 October 2024, marking the end of the 182-year long story of horse racing in Singapore. The racecourse will be demolished by 2027. You can still see it in person, at the time of writing, but it isn’t expected to stay intact for long.

You can find some pictures of the racecourse facilities here, though the racecourse is no longer publicly accessible.

Occupying an area of 1.24 square kilometers, the site of the former Kranji racecourse is expected to become a housing estate, with several of the racecourse’s structures preserved to serve a second life as estate amenities.

I’ll talk more about transport in the area some other day, but for what it’s worth, the additional homes created through this redevelopment will not likely overwhelm the existing Kranji MRT station. After all, that station is equipped to handle the demand from the racecourse when it was in operation.

We can expect some additional investments in the area to boost first-and-last-mile connectivity, and potentially create new links, but we will not be seeing the same amount of investment as districts such as the Jurong Lake District. I am hoping, at least, that the new district will be designed with car-lite goals in mind.

I’m personally excited to see how the new district will incorporate the area’s equine history. Perhaps the old track could remain as an open space, or the grandstand could be redeveloped into a mall or other third space. It would also be lovely if the new district included a platform for the stories of Singapore’s racehorses to be told once again.

And of course, I’m also hoping the area will integrate well with nature, particularly with green spaces such as the Rail Corridor and Central Catchment Nature Reserve in the vicinity.

Kranji Racecourse will be a different kind of urban space in its second life. While it’s certainly a pity to see the local horseracing scene go, especially when, in hindsight, I know it was on the cusp of a renaissance, the new housing development will continue to shape Singapore’s urban landscape in new ways. Done right, it will be a new vibrant space in the north. And hopefully, it will be able to do Singapore’s equine heroes justice.

Good public transport is an important part of getting people out of their cars. There’s countless benefits to a car-lite city — fewer traffic deaths, less pollution, reduced carbon emissions, quieter cities, less space needed for roads and more that can be used for everything else.

But for this vision to happen, public transport must be attractive enough of an option to convince drivers to ditch their cars.

Despite what the LTA might want you to believe, public transport here isn’t quite there yet. We can see this reflected in COE prices, in the demand for ride-sharing services, and in comments expressing general dissatisfaction that you might come across on the Internet.

There’s still more that can be done for a more seamless, more user-friendly, more integrated public transport system.

Transport planning can be likened to choosing tools out of a toolbox.

Each tool has its own role, and so having more tools to choose from makes it easier to finish the task at hand.

In transport terms, this means weaponising more modes of transport to move people. Multi-modality gives people choice, and it also makes sure that different needs are accommodated to by different levels of service.

Singapore needs to develop more of these tools.

We’re working there, of course, with the LTA now looking into autonomous vehicle shuttles that have the potential to be used on fixed routes as buses or as demand-responsive transport.

We are also aggressively expanding active mobility infrastructure, though I personally disagree with the way it is implemented here. Regardless of the shortcomings of the LTA’s cycling “paths”, though, active mobility is a very useful tool for the first and last mile, helping to bridge gaps that are not covered by the public transport network.

But more can be done.

If Singapore is to sustain a car-lite population, public transport capacity must increase, especially in the hyperlocal space. This means we have to develop more tools beyond just buses. Ultimately, buses lack capacity, and so, tools that can move more people are sorely needed.

We must also be creative in the types of tools we develop. I previously wrote about the potential of water transport, but Singapore must be able to make use of any and all technologies available to it.

But it’s important to choose the right tool for the right job.

This is something Singapore hasn’t been particularly great at.

Like when we chose people movers typically seen in airports and theme parks to serve as neighbourhood transit.

Or when we chose light metro-sized trains for our orbital.

Choosing the wrong tool for any particular role can often lead to shortfalls — poorer capacity, higher costs of operation, poorer reliability, or even poorer comfort. All things that can stunt a public transport system and hurt its reputation.

But it’s also cardinal to use each tool for only one job.

After all, each tool performs best doing its own role. Just as a ruler wouldn’t do very well measuring temperature, each mode of transport can only do one thing.

When I’m thinking about transit planning, I keep in mind what I call the ‘transit hierarchy’.

The transit hierarchy relates the characteristics and goals of a proposed transit project to the ideal mode, and is a useful guide prescribing standards for vehicle capacity, service frequency, station spacing, and other qualities for each situation. Of course, individual qualities for each transit project can be tweaked based on the specific circumstances surrounding each rail line and its destinations.

But the key idea is that each transit project should only fall into one of these categories.

This is my main gripe with the Singapore MRT — it tries to do too many things at once.

The Singapore MRT is a metro system — designed for short-distance urban travel, but it also has to double as a commuter rail network, facilitating longer-distance travel from a suburb to the city center.

And it IS an excellent metro system. It’s very good for short journeys, such as Orchard to City Hall or Hougang to Punggol Coast.

But it does poorly at being a commuter rail system — the Singapore MRT isn’t designed to be one.

Commuter rail systems often place a heavier emphasis on comfort and speed, often featuring transverse seating and express services or longer spaces between stations, neither of which the Singapore MRT has.

The result — commuters from suburbs further away from the city center have to endure longer and more uncomfortable journeys even as their trains get more and more crowded as they travel closer to the city center. Many of these journeys are also really slow — a trip from Punggol to the city center takes around 40 minutes, for example.

This means the Singapore MRT is designed in a way that disenfranchises commuters from further out towns — such as Tampines, Sengkang, Punggol and Boon Lay.

This is why many people in Singapore clamor for express trains. What they’re really asking for is commuter rail; the average layman just doesn’t know what to call it.

But beyond speed and comfort, there’s a more fundamental reason why commuter rail is necessary for Singapore.

Capacity.

The human pursuit for faster transport has always been motivated by capacity demands. From the Shinkansen to the RER to our very own Cross Island Line to the BRTs of Latin America, express transit has always been targeted at increasing capacity. After all, capacity dictates how many people the transit system can carry, and for transport operators, how much money they can milk out of their transit system. Speed, while important and very nice to have, is but the means to unlocking more capacity.

Commuter rail being faster and more comfortable — especially for longer distance commuters — means it can attract ridership away from slower, less comfortable MRT lines. This effect allows commuter rail lines to separate commuters making longer journeys from those making short journeys. It opens up more capacity on the MRT lines for more trips to be made by public transport, and is especially important with Singapore’s main MRT lines already bursting at the seams — especially ones that dead-end like the NEL.

But back to the toolbox analogy. Just as it’s important to choose the correct tools and use them the right way, so too is it important to procure high quality tools and maintain them properly.

Transport infrastructure in Singapore must be high-quality, not haphazardly put together — re: our cycle “paths” which are nothing more than paint on a sidewalk. And it’s important to ensure their continued reliability, and to constantly renew, refine, and update these tools to fit the populace’s requirements.

So, what would Singapore look like if we managed our toolbox properly?

Well, I think it would look like Paris. Or at least the transit network will.

Now, Paris isn’t necessarily the first place that comes to mind when we’re talking about public transit. But why shouldn’t it?

After all, it’s the densest urban public transport network in Europe, and it’s one of two cities to have been awarded the Sustainable Transport Award more than once — the other is Bogota, Colombia.

So let’s take a stock take of what Paris has and is about to have.

There’s the Paris Metro with its 14 main lines and 2 branch lines, 5 RER lines, 15 tram routes — 6 of which are tram-trains, and 9 Transilien routes, alongside over 350 bus routes — of which 61 exist within Paris proper — and 3 BRT corridors. There’s also the newly opened Cable 1, Europe’s first urban cable car system, and the Montmartre funicular, which ascends the eponymous hill. This is of course, on top of regional, intercity, and high speed rail services throughout the Ile-de-France region.

Public transport had a modal share in the city center of 30%, whereas cars made up just 4.3%. In the same period, Singapore had a public transport modal share of 42%, but cars accounted for 36% of all trips.

Singapore's Progress Toward The Paris Agreement

Paris also boasts more than 1,000 km of cycle paths and routes. These active mobility rights-of-way make it easier and safer to cycle around Paris. Many of these cycle lanes are separated from other traffic — yes, including pedestrians, meaning they’re safe but still allow cyclists to attain some speed.

Paris is also rapidly pedestrianising streets, closing them off to private vehicles. They’ve also placed an outright ban on SUVs entering the city centre. All welcome policy changes that make Paris’ streets safer for everyone.

In 2022–2023, bicycles represented 11.2% of all trips in the city centre. Doesn’t sound like a lot, but this is also a small, safe, and pleasant enough city for 53.3% of such trips to be made on foot. And compared to Singapore’s measly 1%, Paris has actually been very successful at encouraging people to cycle.

Cycling in Paris - Wikipedia

Despite the already incredible size of its transit network, Paris continues to aggressively expand its transport network. By 2030, 4 more metro lines, lines 15 to 18, will open. Part of the Grand Paris Express project, these lines are targeted at connecting outer suburban areas — including Paris Charles de Gaulle airport — with the city center and other suburbs. Line 15, which forms an outer loop, will be the longest rapid transit tunnel when it opens.

There are also extensions to some older metro lines to link them up with the rest of the GPX project, including an extension to Paris Metro Line 14 which opened last year.

RER E will also be extended west once more, and by 2026, Parisians can expect another 180 km of cycle paths.

What’s amazing is that the bulk of this transport infrastructure — including the cores of all 5 RER lines, the metro lines, and most of the tram routes — lies within Paris’ city limits, an area comprising around 100 square kilometres. That’s an area 7 times smaller than Singapore.

Now, I’m not saying we need to have 7 times the transport infrastructure that Paris has, but it does teach us something: Singapore isn’t too small for more transit and faster transit.

And we certainly have room for more creativity in transit planning.

I was bored, so I decided to do something fun. AI has been a big topic recently, and especially generative AI. Like many others, I’m worried if AI will eventually take over my job.

So I decided to put ChatGPT to the test.

I asked it:

if you had the power to construct an MRT line anywhere in Singapore, where would it be?

And here are the results.

It’s an interesting proposal. The “Northern Arc Line” would serve a similar role to the Cross Island Line, but would be more focused on underserved areas such as Tengah and Seletar.

So the concept itself is fine.

But let’s take a closer look at ChatGPT’s proposed alignment.

There are a few details here that I’m pretty sure ChatGPT has hallucinated. Firstly, the NSL and JRL meet at Choa Chu Kang station. There isn’t even a station, current, future, or planned, called Choa Chu Kang North.

Similarly, Punggol Coast station isn’t on the Cross Island Line. I believe ChatGPT actually means Punggol station here.

I also asked ChatGPT to draw its proposal on Google Maps for me, but it couldn’t do that. Instead it gave me a list of coordinates, which I plotted out myself.

I’ve attached a copy of the plot below for your reference, but I’ll also be referring to screenshots from it later.

Northern Arc Line - Google My Maps

From the get-go, we can already see a major problem. The coordinates it’s given us for “Tengah Town Centre” station are actually inside Tengah Air Base. Pretty sure that location isn’t going to be a viable MRT station site for a long time.

The second location, which I said earlier:

I’m pretty sure ChatGPT has hallucinated.

is actually quite interesting. It is, in fact, not the existing Choa Chu Kang station, but a new site located between that station and the adjacent Yew Tee station. I guess ChatGPT also wants to extend the JRL there…

But it is too close to Yew Tee to actually justify building a new station.

The next location is actually fairly normal. After all, it’s actually sited next to the existing Bukit Panjang station.

It’s from here where ChatGPT’s proposal gets a little strange.

Right.

What’s with the really big detour to MacRitchie Reservoir? I can accept that any east-west line here would need to pass under the CCNR, but to this extent? And along an alignment where it would have to pass under the CRL’s already 70m-deep tunnels, no less.

The Yishun end of this detour is also weird. ChatGPT proposes a station located at…

So tantalizingly close to Khatib station.

The station after this is at Seletar Airport. Again, pretty normal, except that ChatGPT tries to put the station inside the airport perimeter. Directly under the runway, in fact.

The Sengkang West station isn’t in an entirely unreasonable location, though, as it is adjacent to the Fernvale LRT station.

And finally, the Punggol terminus.

Yeah, ChatGPT may not be so good at plopping down stations. Or knowing where existing ones are, for that matter.

But exact station locations aside, its proposal does make sense. So I probed it further to refine the idea.

I asked ChatGPT to prioritise connecting existing stations instead, and it came up with:

This is better, but this proposal still has some flaws. Mainly in that it still does that mega detour to Upper Thomson.

It also suggests these potential infill stations.

Which suggests the route will pass through Sengkang West and Seletar Aerospace Park between Sengkang and Punggol. If I was building this route though, I would probably build these stations together with the rest of the line, so that the detour to Seletar actually serves a purpose other than adding to travel time.

I now really wanted to get rid of the engineering nightmare that is the Upper Thomson detour, so I tried a few new prompts.

Redo this proposal with the same constraints but make the line as straight as possible

ChatGPT did not, in fact, straighten the line as it continued to keep the Upper Thomson detour.

Why Upper Thomson? Seems to be a bit of a detour

Then it tries to gaslight me with:

But why Upper Thomson instead of, say, Springleaf, which is further north?

It would seem that ChatGPT thinks Upper Thomson is a

compared to Springleaf. Interesting.

I also want to minimise tunnelling under the Central Catchment Nature Reserve, and in this sense, Upper Thomson is less preferable. Where would you route your proposal instead?

And now it cuts out the TEL interchange altogether.

I made some other refinements and questions, but I’m going to skip straight to my thoughts. You can find the entire exchange in the link below.

ChatGPT - MRT Line Proposal

So, my thoughts?

ChatGPT does okay, but it still requires human refinement. For transport planners, it is useful for coming up with a general alignment or list of intended destinations, but the minor refinements such as precise station locations still need human intervention.

Also it tends to hallucinate stuff, so those errors do need to be corrected for.

As for the proposal itself, I’ll continue to work on and evaluate it, and it should come up soon. Stay tuned.

If you’ve been in the transport space long enough, then you’ve probably come across the name “North Shore Line”. If you haven’t, then well, now you have.

The concept of the North Shore Line comes from the URA’s 2001 Concept Plan.

It’s a bit hard to decipher as is, so here’s a representation I found on Reddit.

Pay attention to the lime green line in the top-right corner of the map. That’s the North Shore Line.

Some parts of it may look familiar. The stretch between Seletar and Woodlands was announced as part of the Seletar Line in the 2040 Land Transport Master Plan (in pink), while the stretch between Punggol and Changi is part of the Cross Island Line and its Punggol branch.

But what about the section between Seletar and Punggol? Could that ever happen?

Today, I’m going to explore a proposal to extend the Punggol branch of the Cross Island Line, and we’re going to also discuss some additions to the Seletar Line route between the Seletar area and the Woodlands North station.

The Cross Island Line’s Punggol branch is weird. Despite being ostensibly a branch line, the line wil be operated independently, with its own set of platforms at Pasir Ris station. Trains will run back-and-forth between Pasir Ris and Punggol stations, serving Elias and Riviera stations en-route.

Although operated independently, the branch will still depend on the Cross Island Line’s Changi East Depot. For this reason, the connection with the main line at Pasir Ris points east towards the depot rather than west towards Hougang and Ang Mo Kio.

That restricts the branch from having through operations with the main line. Trains from Punggol can’t travel towards Hougang and Ang Mo Kio without reversing, and they can’t serve Loyang and Changi Airport Terminal 5 without sacrificing peak-period main line capacity.

As it is the branch will serve purely as a link between Punggol and Pasir Ris stations. Of course, this is a very essential connection. It bridges a gap between the East West Line and the North East Line, allowing residents in Sengkang and Punggol to travel to Tampines and Pasir Ris by rail. It’ll improve access to the Punggol Digital District, and it will allow residents of the northeast to reach Changi Airport quicker, via a transfer to the CRL main line.

As I’ve emphasised several times throughout the course of my blog writing, Changi Airport isn’t just a gateway to international destinations, it’s also a major employment centre. That makes it all the more necessary for it to have quick and efficient links to destinations across the island.

The Punggol extension is a step towards this. But there’s still a major gap to be filled.

If you’re living in Yishun or Woodlands today, your best option to get to the airport is to take service 858. Not the most pleasant, given it’s one of Singapore’s busiest bus routes.

Singapore’s Most Crowded Bus Routes: A deep dive into public bus data (Part 2)

It’s also Singapore’s longest, so it also just takes a long time.

Having to rely on buses for this (and other similar) journey(s) is a fact that will be true even after the Seletar Line opens in the mid-2040s, even if a significant portion can now be done by rail. That’s because of the sizable distance between the Seletar Line and Punggol station that will necessitate a bus transfer anyway.

But Singapore’s bus industry is facing a serious shortage of manpower. We certainly won’t be able to cough up enough buses to properly bridge that gap.

A modal shift to rail will be necessary. So that gaping hole between the Seletar Line and Punggol must be bridged by rail.

Nothing in this article is to be taken as an indication of official policy. This is purely speculative.

Therefore the Punggol branch needs to be extended to meet the Seletar Line. Based on current projections, this will most likely (conveniently) be around Seletar Airport.

The link to Seletar Airport will be transformational. With Changi Airport’s passenger capacity doubling with the opening of Terminal 5 in the 30s, the ability to utilise spare capacity out of Seletar Airport — a so-called “double hub” — for shorter routes will be valuable.

This year, the RTS Link to Johor Bahru will open, and that would open up opportunities to get Johorians to travel to international destinations via Changi Airport. Currently, the way LTA intends to do this is via the Thomson-East Coast Line, but, frankly, that line will take ages between Woodlands North and Changi Airport. Extending the Punggol branch to the Seletar Line (also projected to call at Woodlands North) will be far more attractive for airport-bound travellers (once the Seletar Line is built, anyway).

The proposed extension will also give residents in Punggol, Sengkang, Tampines and Pasir Ris access to employment opportunities in the industrial areas centered around Seletar Airport.

And while this proposal is transformational enough as it is, I’m going to take it a step further. But for this, I’m going to have to first put forth a prediction of where the Seletar Line stations between Seletar Airport and Woodlands North will be.

Right, here’s the radical part of my proposal.

From Seletar Airport, the Punggol branch is further extended, running along the Seletar Line. Along this corridor, the Seletar Line will run local, while the Punggol branch will provide express service.

You might be wondering how this is possible. Surely the roads here aren’t wide enough to accommodate 4 whole tracks?!

The answer is that they… aren’t. So instead of laying the tracks side by side, we simply stack each pair of tracks.

This way, we use the equivalent horizontal footprint of two tracks to accommodate 4. I’ve represented the layout as underground (hence the circles), but after Seletar Airport, the line could also be elevated in the same way. Elevating the two lines would reduce the engineering complexity of the construction, and make it cheaper and quicker to construct.

The stacked arrangement also presents another benefit. At express stations, it enables same-direction cross-platform interchanges between the two lines, allowing commuters to seamlessly transfer between local and express trains.

The extended Punggol branch, as the express service, will only stop at two of the intermediate stations along this section.

One of them will be Yishun East station, located at the junction between Yishun Ave 9 and Yishun Ave 8.

This station is chosen as the express station because it is directly east of Yishun Town Centre. That means quick rapid transit routes through Yishun Town can be provided, and with the express trains as a pull factor, can effectively relieve the North South Line. Other considerations include the existing population density around the station, as well as potential future developments provided for in the 2025 URA Master Plan.

The other station I’ve selected for express service is Sembawang Shipyard station.

With the area surrounding the old Sembawang Shipyard being announced as an upcoming residential precinct, and the Shipyard itself set to transform into a “mixed-use waterfront lifestyle district”, this part of the Seletar Line alignment will undoubtedly be in high demand.

The express service will free up capacity along the local Seletar Line for Yishun and Canberra residents. They may also be able to relieve pressure on downstream sections of the Seletar Line by directing commuters to the Cross Island Line — helping to free up space for Sengkang West residents, and they will give commuters from the RTS Link a quicker way to get to Pasir Ris, and therefore, Changi Airport. Oh, and the express service will also be a quick way of getting to the new Seletar Airport air-and-rail hub.

With these two phases built, the Punggol branch will be transformed into an essential corridor for residents of the north and northeast. It’ll also be an effective link to and between the two airports. This proposal is a big step to providing better orbital links, helping to relieve pressure on the Circle Line and Cross Island Line. It’ll also greatly benefit commuters coming off the RTS Link at Woodlands North, providing them with more options for travel to various destinations.

The North Shore Line will transform northern Singapore.