Fewer seats? – How many seats do we want on our trains?

The debate about train seats has come up again, thanks to The Greens uncovering minutes of a meeting between Metro and the Department of Transport discussing the removal of train seats from Comeng trains. (MX story / Channel 7 story)

DOT was generally comfortable with the proposals as presented by MTM. Options to be assessed were only to include low cost options necessary to achieve a 900 load standard with no reduction in dwell time performance.

Comeng train interior

The proposal is to remove the third seat in groups of three, widening the centre aisle.

I make it 12 seats in a “T” carriage (trailer, with no driving cab) if they don’t widen the aisle at the far ends of the carriage, or 16 if they do, making it roughly a 13-16% reduction — though it would vary according to the type of carriage (Motor or Trailer, and EDI/M>Train refurbishment or Alstom/Connex refurbishment).

This would differ from the Connex trial layout, which took out more seats near the doors, but left more in the centres of the carriages, including a narrow aisle. Overall that layout removed more seats than seems to be proposed now.

As ever, it’s aimed at fitting more people on board (the minutes talked about an increase in the “load standard” from 798 per 6-carriage train to 900) and improving flow within the carriages, which would help station dwell times (the time taken to load and unload a carriage) — in this case, they’re specifically looking to be able to carry more people without increasing dwell times.

This latter point is important: if you want the train system to run more efficiently, with the maximum number of trains on the most congested parts of the network, eg the City Loop, at peak times, you need to improve dwell times. (Connex claimed in 2009 that the trial layout did help this.)

What kind of train system do we want?

I’ve just finished reading Jarrett Walker’s excellent book “Human Transit“, which ponders a lot of these kinds of issues. He calls them “plumber questions” — the kinds of questions a plumber asks a client. There’s not necessarily a right or wrong answer, but you do need to give some guidance as to the outcome you want.

In the case of train seats, it’s one of the questions related to what kind of train network we want. (Another was posed a year ago: Should every train run around the Loop?)

Broadly speaking…

Do we want a (small m) metro? Frequent services, aimed at more than just 9-5 CBD workers (eg including short suburban trips, counter-peak trips); fast dwell times for efficiency; less seats to maximise speed and capacity.

Or do we want a commuter rail service? Less frequency, particularly outside peaks; more seats because it’s primarily about long trips; primarily concerned with CBD trips, meaning just five stations have to handle huge passenger loads; but can lead to longer dwell times and lower peak frequencies because you don’t take advantage of metro efficiencies like more doors/less seats.

It’s not actually black and white. Melbourne is probably destined to remain somewhere in the middle.

The CBD outstrips public transport demand for all other destinations, and will continue to do so until traffic and parking demand is such that paid parking and gridlock becomes prevalent in the suburbs. (It’s getting there, but slowly.)

But there’s no reason we can’t have frequent (10 minute or better) trains all day everyday, just like real metros, supporting suburban non-work trips, and ensuring patronage is not just about peak hour, therefore providing a better return on the investment that’s been made in rail infrastructure, fleets and staff.

And remember, handling the booming 9-5 CBD commuter load better means optimising operations, including internal designs of carriages. The current designs from the early 2000s (before the boom) try to maximise seats, and in the face of surging demand, this has left passengers left behind on platforms, sometimes when there is space in the middle of carriages, because those aboard have not moved down. This is a direct result of narrow aisles and virtually nothing to hold onto except around the doorways.

How many seats?

Taken to extremes, seat removal might result in something like this:

Brisbane train

Where’s this? It’s Brisbane. I suspect few want to see that kind of outcome here.

Note that only some of the carriages are set out like this, with maximum standing space, whereas others have more seats:

Brisbane train interior

But it got me thinking… how do other cities design their carriages? What ratio of seats to carriage space do they have? I did a quick comparison, and came up with the following.

CityTrain typeCarriagesSeatsSeats/square metre
LondonUnderground 199562480.88
Hong KongEast Rail Line Metro Cammell126250.71
BrisbaneSMU 26064721.19
PerthB series63840.90
MelbourneComeng Alstom (current)65361.24
MelbourneComeng Alstom (proposed)64641.07

(Some of these are estimates, as I couldn’t find very reliable figures. An authoritative figure for seats in Perth’s trains was elusive, and the length of carriages sometimes included couplings, which aren’t part of the useable area inside. But you get the general idea, hopefully.)

The current changes to Melbourne carriage designs (first seen in the second series X’trapolis trains) are leading to wider aisles and more handholds to encourage people to move down, and help stop as many congegating in the doorways. But from what I can see, they still provide more seats than in many other big cities.

The Comeng proposal is similar (the seats per square metre figure will come down to about 1.07 by my calculations), though we’re not yet sure if it includes more handholds.

I think it’s probably a reasonable proposal, provided it includes more handholds along the carriage, and provided it’s accompanied by a service frequency boost (particularly outside peak hours, when there’s no problem with track or fleet capacity) so the total number of seats offered on each line doesn’t drop (or possibly even increases).

And even if there’s no frequency boost in peak, if you’re outraged by the idea of removal of any seats, consider this: In the face of continually rising patronage, would you rather be able to squeeze onto a train with 15% less seats, or be left behind on the platform?

What do you think?

By Daniel Bowen

Transport blogger / campaigner and spokesperson for the Public Transport Users Association / professional geek.
Bunurong land, Melbourne, Australia.
Opinions on this blog are all mine.

22 replies on “Fewer seats? – How many seats do we want on our trains?”

I am typically a non-peak traveller, and do not like driving. At age 64, I am also unable to stand for lengthy periods in trains. Removal of seats from trains would be a sentence to stay at home.

I can envisage not being able to get a seat for off peak travel soon. Removing seats sounds awfully like a cheap and easy fix as an alternative to improving services. Our trains and the system is not designed as a Metro style operation but as a suburban commuter system. Our trains don’t accelerate and brake quickly. I am not sure why Melburnians are expected to put up with overcrowded trains when other countries manage to do it far better.

Excellent analysis and observations.
I think it needs to be easier for passengers to walk through carriages on crowded trains, thus clearing people away from the door area. If this means removal of some seats, then so be it.

Although I can see the case for fewer seats in cramming more on board, I am still reluctant over the issue of having to stand all the way to Clayton/Westall/Spingvale on every trip home fo those of us with the misfortune of getting on at Richmond/South Yarra/Caulfield which would be the result of this move (and was the case we got the reduced-seat Comeng on our run – and this was at 4.30-ish) – which would make the train that bit less attractive an option.

Removing seats does make it easier to move around the carriage, but I’m not convinced it allows more people to fit on the train. My experience on the X’traps with the reduced seating is that it accomodates the same number of people down the isle as those with their old-style full complement.

The space to stand and keep your balance is about the same footprint as it is to sit, unless you’re happy to be pressed up against someone else on the train. This latter issue is enough to drive Melbournians back to their cars – I think many/most would prefer to sit in idle road traffic if it means keeping their personal space.

My problem with seats facing inwards to the train is it makes it difficult for people to stand close by. When seats face each other you can cross your legs around the person opposite you. But when the person opposite is standing this can make it very difficult as they need to keep their balance. The only real solution in reducing seats is to have no seats at all!

With regards to alighting the train, I think it has more to do with door layout. Wider isles can only do so much on the Comengs & X’traps when only 2 people can fit through the doorway at a time. The best example I’ve seen of where dwell times are ridiculously long due to alighting is Belgrave/Lilydale morning peak trains at on Platform 4 Parliament Station. 5th carriage middle door is directly opposite the main station exit. This causes a bottleneck with people trying to clamber on to the escalators, which makes it difficult for people to get off the train. This morning my train closed up doors and departed before those who needed to re-board the train could get back on! Sometimes I’ve seen people not be able to get off in time.

On a personal note, I try and catch the train earlier and earlier to get a seat. Why is it so important to me? The newer trains jump around so much that I worry about the long-term effect on my knees. I still consider myself as young, but my knees are noticeably stiff after a 30 minute journey standing on an X’trap. I’ve probably got at least another 20-30 years of riding the trains to work, so standing 30 minutes (or 60 for a return trip) each day over that period of time with such bouncing around cannot be good for my health.

@Michael; if you’re a non-peak traveller, you’re still very likely to get a seat. Also, I find most people happy enough to give up their seat if requested; these days, I think people are less likely to offer for fear of causing offence, etc, but still willing to assist when they can.

Anecdotally, I think the wider aisles do a lot. I’m normally boarding at an inner city stop so expect to stand. It’s normally easier to board, and hang-on, on the newer layouts than older Comeng and Xtrap designs, specifically because of the wider aisles allowing people to spread out and not be afraid of being ‘stuck in’ once they need to disembark, and having all the extra handholds.

I’d also like to see increased services but support ideas that get the most, for most people, out of the current system.

On the question of what type of network Melbourne wants, the fact is that it needs both. Metro serves two distinct groups with different requirements.

Inner suburban commuters value frequency of service over comfort, while those with a journey greater than 30 minutes understandably don’t want to stand the whole time.

Melbourne’s long-term PT plan should envisage a high-frequency, metro-style service for short trips to the CBD (i.e. essentially what zone 1 is now) along with regular commuter services to the outer suburbs. Commuter services would run express through the inner suburbs and have trains designed to seat large numbers. If commuter trains ran direct to Flinders St this would also help mitigate the City Loop crowding issues.

” would you rather be able to squeeze onto a train with 15% less seats, or be left behind on the platform?”
That last point you make is a very excellent point and a serious question, however, in my personal experience, it seems a large number of people can’t get to this kind of factual conclusion, and reckon that we can just throw on more trains onto the line and keep the seats coming (which of course, can’t happen because there’s many limiting factors).
I’ve personally come to accept that sitting down during rush hour isn’t going to happen (I usually get on at Box Hill, and by the time the train gets to Blackburn (two stations before), it’s standing from thereon), but there are always people at Box Hill who rush on, push through to the front of the line, in hopes of a non-existent seat.
I’m quite happy with the current frequency at Box Hill during rush hour (I can’t say much the same for off peaks, and particularly Sunday mornings), but the trains are getting quite full, and it’s not sustainable in the long term. The new X’Trap’s with the 2×2 seating arrangement is good, and it really does get people moving down the aisle much more compared to other trains (it takes a prod and a shove from the crowd behind to get people moving down the Comengs). But soon, it’s gonna get full, and I guess we’ll either need more frequent services, extra cars (which I doubt will ever be achieved with the finger pointing going on) or we’ll need to remove more seats and make it look like Brisbane. Somehow, I feel the easy option will be taken (the last one).

It’d be interesting to see how New York fits into that table.

In terms of my opinion – I’d much rather have the extreme of no seats on every train as I know that it’d at least allow me to get on my train in the morning instead of waiting for 2-3 to pass me by at Moonee Ponds (and then once I do get on, I have to witness every passeneger at the following stations do the same).

However, avoiding extremes, I’m content with not sitting as I rarely ever get a seat anyway (including morning off-peak from 9 until 11). I just have to agree with your statement:

“provided it includes more handholds along the carriage, and provided itโ€™s accompanied by a service frequency boost”

My biggest issue with having to stand is that I’m very short and I can’t reach the overhead handles. More handle options for short people are required (or even a sign asking tall people to hold onto the ceiling handles, allowing the short people to grab onto the verticle bars!)

Another data point for your table, Tokyo Yamanote line, E231-500 series, 11 cars, (3+7+7+7+3) x2x11=594 seats, 594/(11x20x2.950)=0.91 seats a square meter but there should be an extra column for standard 100% load 2.8 people per square meter 162×11=1782 per set less a bit as the seats are no longer locked up at peak time. Crush load (modern definition 200%) =3564 people, old crush load of 250-300 percent but you need pushers and you’ll get stick from the government and told to increase service frequency of around 5000!

Good analysis.

Get rid of the seats (well, at least a few of them). After living in London for the last 6 years and traveling on rail systems in a fair few of the European systems and coming back to Melbourne I was amazed at how inefficient the seating was in our trains. Like you said – if people want to get on a train they will put up with being squeezed in.

For good data on the Perth A and B sets, Peter Clark’s articles in the January and March 2005 Railway Digests respectively are a good start. The article on the A sets covers the capacity enhancement achieved from turning the seats in the saloon of each car into longitudinal seats. Good article by the way. I think Melbourne is at the tipping point of needing to increase capacity and frequency in preparation for the next great patronage surge.


Has anyone done any modelling as to what injuries would result from emergency braking/derailment/collision with vehicle at crossing etc with full load or overcrowded carriages? Unlikely to happen maybe, but it still could..

Any assessment of increased injury risk (from more standees) would have to factor in a comparison with the risk from alternative trips (ie car journey).

@Alex, thanks. I didn’t go into a comparison of total capacity/crush loads, because I was uncertain if the figures I was looking at were comparable.

@John, I seem to recall some Scandinavian research some years ago, but I doubt the risk to an individual is lower or higher dependent on how crowded the train is, if they don’t have a seat. Even having a seat may make minimal difference in a high-impact accident.

Heavy rail is designed to have its own right-of-way, with signals and other safeguards such as automatic stops to protect from collisions, and absolute priority at road crossings. Of course, this doesn’t always prevent accidents, but even without any kind of seatbelt or other restraint, public transport (be it rail or bus) is at least five times safer than car travel (which of course does mandate seatbelts).

I would just be happy if they’d put something to hold onto in the older trains — even if you’re sitting for most of the journey, it’s quite difficult sometimes to make your way to the doors while the train is moving. I’m tall, so I can grip onto the ceiling a little bit, but it’s no subsitute for a proper rail. It’s a similar problem in the new trams actually.

One factor effecting the seating arraignment in trains is their width. The Comengs, X`traps and Hitachis are wider than the Brisbane trains.

There’s a Jumbo 2000 class consist in Adelaide that had this modification done a few years back – utterly horrible :( For sure there is a massive isle, but the end result is:
1. Less seats
2. Even more people crammed standing at each end of the carriage

Nobody stands down the isle of the carriage – never have never will, at least here anyway, so a total dismal fail. None of the subsequent overhauls/refits have had that change done to them since.

Good analysis Daniel. I was away when you wrote this so I’ll add my 2 cents now for anyone stumbling on the article.

I’d much prefer less seats and a higher capacity for various reasons. Firstly, its a simple, cheap and effective capacity increase, which has a multiplying effect with extra services added. Secondly it should decrease dwell times (getting from the innermost seat in a 6 seat layout, to a door can take a while, especially with a few people in the isle and crowding at doorways), which actually allows for more services to be run. Thirdly less seats and more handles makes for a more comfortable journey for people that do need to stand, which inevitably will continue to be a large portion of passengers.

After catching the Toronto subway for a few months I certainly find their new Rocket trains vastly more comfortable and efficient than any of the rolling stock in Melbourne. I imagine their capacity is also considerably higher, and even when there is significant crowding, there is a lot less of crush in the doorways here than at home.

I spent this morning with my face firmly in the door window from malvern to Melbourne central. The door areas were rammed but the aisles were perfectly able to accommodate more people.

I have to agree with Julian that removal of seats especially for peak hour services has to be a reality for the safety and comfort of peak hour passengers.

Also it is a simple and economical fix which will ,Sam I am more able to catch my train. Years ago on the sandy line I often waited for 2 or 3 trains to go by before I could get on. It’s nonsense that a simple fix can’t be made in the interim to resolve this issue.

Comments are closed.