New passenger numbers and Crowding figures published

[wls1]

An obvious one is GEML outer services. With five major options to transfer to at Stratford westbound, I wonder how load factors east of Stratford look?

Similarly, didn’t C2c have to address some severe loadings between Barking and West Ham?

Have always wondered this too, with West Ham being the primary destination for many travellers across the LTS, especially with the timetable change in December 2015 including West Ham and Limehouse calls into services with only one train missing West Ham and two trains missing Limehouse in the am high peak

 

[Magdalia]

I wasn't familiar with this dataset, apart from the top 10most crowded trains, so I did spend a bit of time yesterday looking at the methodology.

The statistical release has some information starting here:

Data source and coverage​

This publication focuses on passenger numbers during the morning and evening peak hours on Tuesday, Wednesday and Thursday, when rail travel tends to be busiest. The statistics are based on passenger counts carried out at specific station stops between 16 September and 13 December 2025, and represent travel on a typical autumn weekday. As a result, they do not represent all rail journeys.

There is also more detailed notes and definitions documentation which is here:

Rail passenger numbers and crowding statistics: notes and definitions

www.gov.uk

This document provides a range of background information on the Department for Transport’s (DfT) rail passenger numbers and crowding statistics. It also details some of the factors that may affect the accuracy of the statistics.

The data appears to be taken at terminus stations and a few selected stations eg Vauxhall (but how these stations are selected, and others just before interchanges where many alight aren't selected, is not explained). The choice of these extra stations appears to be be historic from years ago and thus does not include busy sections of line not near a city terminus.

The London stations are shown on a map in figure 4.

It is also my impression that what is included is historical, with the major changes being inclusion of various big cities outside London from 2010, plus Brighton, Cambridge and Reading added in 2017. In London, using zone 1 as a boundary looks much more questionable in 2025 than it would have done many years ago.

An obvious one is GEML outer services. With five major options to transfer to at Stratford westbound, I wonder how load factors east of Stratford look?

Similarly, didn’t C2c have to address some severe loadings between Barking and West Ham?

And most Lea Valley peak trains are most crowded north of Tottenham Hale. Looking at the examples, it is quite challenging to come up with a simple rule of what should be in and what should be out.

This isn't just a London issue. Wearing my Fenland hat, I am surprised that, when Cambridge was added, it did not not also include Cambridge North, and I wonder what will be done in autumn 2026 now that Cambridge South is open.

 

[Posa]

How is London Overground never on these top ten stats when people regularly get left behind on the northern end of the Sydenham corridor AM peak? I've never seen crowding as bad as LO, anywhere.

Exactly.

How can Elizabeth and the Overground not even have one train in the top ten when they regularly leave people behind at multiple stations and drivers physically struggle to close the doors yet a train in Wales is.

 

[The exile]

Exactly.

How can Elizabeth and the Overground not even have one train in the top ten when they regularly leave people behind at multiple stations and drivers physically struggle to close the doors yet a train in Wales is.

How often are either short-formed? I assume that PIXC is measured against the booked number of seats so a “full and standing” (but not physically impossible to board) 2 car train that should have been 4 cars will be at a PICC level that is physically impossible to achieve in a 10 car train designed with a lot of standing in mind.
EDIT: something wrong with my logic there!

 

[JamesT]

Exactly.

How can Elizabeth and the Overground not even have one train in the top ten when they regularly leave people behind at multiple stations and drivers physically struggle to close the doors yet a train in Wales is.

The notes and definitions document linked by @Magdalia in the previous post may be quite enlightening. The statistics are based on passengers in _excess_ of capacity. For long-distance services, effectively any standees will count towards that excess. For EL and LO their trains are designed around a certain level of standing. So what may look quite a full train is still within capacity, by the time people can't fit onboard it may only be something like 10-20% over capacity, but the headline services are 46-76% over.

 

[Horizon22]

I do wonder how easy it is to do an accurate passenger account when the train is at 170% of stated capacity.

Yes there’s on board systems that can measure this but it’s a lot of averages.

== Doublepost prevention - post automatically merged: 29 Jul 2026 ==

How often are either short-formed? I assume that PIXC is measured against the booked number of seats so a “full and standing” (but not physically impossible to board) 2 car train that should have been 4 cars will be at a PICC level that is physically impossible to achieve in a 10 car train designed with a lot of standing in mind.
EDIT: something wrong with my logic there!

It would be hard to short-form an Elizabeth line train! Same with most Overground routes.

It’s the frequency that helps here. An Elizabeth line is stated to hold 1,500 but I’ve seen data that suggests more in the past.

Peak loadings are normally just before the terminal when people are getting on but less so getting off, but before the exodus in Zone 1 so Overground routes might be somewhere like New Cross, or Elizabeth line at Ealing Broadway or Maryland.

Routinely short-forming a peak train will impact the PiXC data.

== Doublepost prevention - post automatically merged: 29 Jul 2026 ==

That’ll be because, short-forms aside (which can slip through the net as these figures are based on actual counts), the biggest problems at Bath are not during the weekday peaks. And yes, it’s my local station…

Bath definitely has a weekend capacity problem.

 

[The exile]

It would be hard to short-form an Elizabeth line train! Same with most Overground routes.

That was my point.

Bath definitely has a weekend capacity problem.

The city as well as the trains!

 

[Horizon22]

That was my point.

Ah yes, it was a theoretical question.

As for Bath yes it suffers with some of the same problems as Oxford and a lot of the GWR route has these weekend/leisure capacity issues but has done for 10 years and not adequately resolved.

 

[dk1]

The Bittern line from Norwich-Sheringham would be a good contender today. Not only glorious weather but it was the 150th anniversary of the opening of Gunton Railway station with celebrations at the excellent Suffield Arms opposite which helped the station see a 20% usage increase last year. Standing from front to back on the 755/4 that formed the 11:41 ex-Norwich which had as many or more joining than alighting at all stations enroute to Cromer.

 

[Russel]

The Bittern line from Norwich-Sheringham would be a good contender today. Not only glorious weather but it was the 150th anniversary of the opening of Gunton Railway station with celebrations at the excellent Suffield Arms opposite which helped the station see a 20% usage increase last year. Standing from front to back on the 755/4 that formed the 11:41 ex-Norwich which had as many or more joining than alighting at all stations enroute to Cromer.

Last time I did the Sherringham line was about 20 years ago and I seem to recall it being a single Anglia 153, how times change...

 

[dk1]

Last time I did the Sherringham line was about 20 years ago and I seem to recall it being a single Anglia 153, how times change...

Oh my days mate, I had to drive the blessed heaps. It’s a world away but so are passenger numbers. Unfortunately not every service can be formed of a 4-car such is their demand across the network. GA as a whole saw another table topping 10% growth between Jan/Mar this year too!!

 

[Russel]

Oh my days mate, I had to drive the blessed heaps. It’s a world away but so are passenger numbers. Unfortunately not every service can be formed of a 4-car such is their demand across the network. GA as a whole saw another table topping 10% growth between Jan/Mar this year too!!

To be fair, it's good to see that capacity has at least tried to keep up with demand in GA land, in a lot of areas it has stood still since the early 2000s, XC being a prime example.

 

[800002]

I do wonder how easy it is to do an accurate passenger account when the train is at 170% of stated capacity.

There are two counting methods generally agreed upon.

1. Phyaical counts at the stations with observers counting people on and off.

2. Analysis of gate line data / ticket data / automatic on board loading data.

There can be some manual revisions based on outlying data, but otherwise it's generally what the oberservers note and the correlation of the data collected in point 2.

 

[Krokodil]

To be fair, it's good to see that capacity has at least tried to keep up with demand in GA land

Or has capacity driven the demand? "Build it and they will come" so to speak.

There are two counting methods generally agreed upon.

1. Phyaical counts at the stations with observers counting people on and off.

2. Analysis of gate line data / ticket data / automatic on board loading data.

There can be some manual revisions based on outlying data, but otherwise it's generally what the oberservers note and the correlation of the data collected in point 2.

Don't many units also have automatic counting now? Admittedly of variable reliability.

 

[Russel]

Or has capacity driven the demand? "Build it and they will come" so to speak.

Very true, it could well be.

Either way, it should be replicated across many areas as it clearly works.

 

[800002]

Don't many units also have automatic counting now? Admittedly of variable reliability.

Yes, but it isn't universal across all operators / fleets.
I'm also told retro fitted devices can be quite temperamental/ inaccurate due to design and fitment issues.

 

[Bald Rick]

Exactly.

How can Elizabeth and the Overground not even have one train in the top ten when they regularly leave people behind at multiple stations and drivers physically struggle to close the doors yet a train in Wales is.

PIXC has a 20 minute allowance. Hence the busiest train, 9T81 0713 Bedford to Three Bridges is in, because it is booked 21 mins from the last stop (St Albans) before London. The train after it, 9T13 0716 Bedford to Brighton, which in my experience is far more crowded (to the extent that I always let it go) is not on the list as it has only 12 mins St Albans to West Hamsptead, and a further 10 minutes West Hampstead to St Pancras.

By chance I was on 9T81 today, and there were spare seats approaching St Pancras (albeit more people standing than seats spare, but not many)

 

[Snow1964]

By chance I was on 9T81 today, and there were spare seats approaching St Pancras (albeit more people standing than seats spare, but not many)

Can probably allow about 20-25% reduction during school summer holidays as number of commuters will be away. Thats why the big Shutdowns (Waterloo in 2017, and now Charing Cross) happen this time of year

From memory first 3 weeks of November are traditionally busiest, because hardly anyone takes holiday then. From late November seems many take odd days to use up holiday allocations, or do pre Christmas events before end of year

 

[dk1]

PIXC has a 20 minute allowance. Hence the busiest train, 9T81 0713 Bedford to Three Bridges is in, because it is booked 21 mins from the last stop (St Albans) before London. The train after it, 9T13 0716 Bedford to Brighton, which in my experience is far more crowded (to the extent that I always let it go) is not on the list as it has only 12 mins St Albans to West Hamsptead, and a further 10 minutes West Hampstead to St Pancras.

By chance I was on 9T81 today, and there were spare seats approaching St Pancras (albeit more people standing than seats spare, but not many)

I had totally forgotten about PIXC. Thank you for rekindling

 

[Bald Rick]

I should also say, whilst the methodology is clear, it is also a game of averages.

On Thameslink services from the Midland for example, it is a rare morning train in the high peak in the autumn survey period that, on Tuesdays to Thursdays, has fewer people standing than seated arriving at St Pancras. Yet on Friday, it is a rare morning train that has more than 50% of seats occupied (Mondays are usually only a few standees). On average this will give the numbers quoted. However, passengers who only ever travel Tuesdays - Thursdays (and there’s lots of them) will feel it is quite different!

 

[em81]

I presume Manchester's latest figures would be affected by the closure of Stockport station last year.

The year on year comparison next year could be interesting, given it's currently Metrolink lines that are closed, with passengers turning to rail services instead.

== Doublepost prevention - post automatically merged: 30 Jul 2026 ==

To be fair, it's good to see that capacity has at least tried to keep up with demand in GA land, in a lot of areas it has stood still since the early 2000s, XC being a prime example.

I saw the EMR Liverpool to Norwich service pass through Stockport at around 16:55 today. 4 car 158 with standing passengers in every carriage. It's not ideal if you're getting on a regional train with luggage to find there's no seats, and even getting your suitcase along the carriage to the luggage rack isn't possible.

Moderator note: discussion continued at:

Lack of capacity on EMR regional services, such as Skegness and Liverpool-Norwich. What can be done?

Moderator note: Split from https://www.railforums.co.uk/threads/new-passenger-numbers-and-crowding-figures-published.305159/ I saw the EMR Liverpool to Norwich service pass through Stockport at around 16:55 today. 4 car 158 with standing passengers in every carriage. It's not ideal if you're...

www.railforums.co.uk

 

[LLivery]

Thanks all for the responses.

I had totally forgotten about PIXC. Thank you for rekindling

So had I haha.

It's really not a great to state these are the 'most crowded' when most people's idea of crowding isn't based on such rules. I'd take 20 mins on a crowded Thameslink over 10 mins crushed on LO any day.

When this methodology is eventually reviewed, a method using flow data matched with time, capacity provided and weight data might be a better way in years to come, especially with the prevalence of smartcards.

 

[matthenj]

The caption for Figure 2 is wrong: Reading is not a city, and has never been - interesting that it rates massive above a lot of the major cities in the North. I guess a lot of people Drive/Bus into Reading to get to London quickly.

 

[The exile]

The caption for Figure 2 is wrong: Reading is not a city, and has never been - interesting that it rates massive above a lot of the major cities in the North. I guess a lot of people Drive/Bus into Reading to get to London quickly.

Don’t know whether it still applies, but when I was briefly living there in the early 90s, there were said to be more people commuting into Reading by rail than out of it.

 

[crablab]

Don’t know whether it still applies, but when I was briefly living there in the early 90s, there were said to be more people commuting into Reading by rail than out of it.

Interesting question. Anecdotally I would say 50:50; Reading is very much commuter territory and even in the last few years has gained a lot of shiny flats (Station Hill, Riverworks etc.) which seem specifically marketed towards the London commuter.
That said, it is still a regional centre of employment so there's significant inflow too.

or Elizabeth line at Ealing Broadway or Maryland.

I would say more Paddington than Ealing Broadway; not many people get off at Paddington Low Level but trains arriving are usually full & standing with loads more getting on. It helps that Paddington has the turn back so there are empty trains to absorb the crowds!

PIXC has a 20 minute allowance.

This seems somewhat reductive. I think one would find EL between Paddington and Ealing Broadway pretty crush loaded and leaving people behind at some points in the peak, not withstanding the accounting for standing capacity.
Perhaps the methodology needs some adjustment for city metro services?

 

[coppercapped]

Don’t know whether it still applies, but when I was briefly living there in the early 90s, there were said to be more people commuting into Reading by rail than out of it.

There is still certainly a net inflow of people into Reading to work, but I haven't found any specific recent statistics about rail flows. The closest I have come are figures given in the 2023 update to Reading Borough Council's Local Plan which states in Section 1.2.3:

Reading is a major centre of employment, with 117,000 employee jobs available as of 2022. There are more jobs in Reading than workers (e.g. there were 91,400 economically active persons within the borough in 2022), which means that Reading typically imports workers from other local authority areas, placing strain on the transport network and on the housing market. This reflects the economic success of the town, which functions as the centre of the Thames Valley, one of the most economically dynamic regions in the country.

However it would not surprise me if there are more rail travellers to Reading to work than leave it. I suppose it might be possible to work this out using the Origin and Destination Matrix, but I haven't got the energy!

 

[evergreenadam]

Hasn't the 1746 London Euston - Crewe been extended to 10 coaches (previously 8?). Is this including that, or is it still 170% over capacity with that included?

They used to say how many coaches the service was formed of, do they not do that anymore?

 

[dk1]

They used to say how many coaches the service was formed of, do they not do that anymore?

Of course. More so now than ever. This is just old news.

 

[Horizon22]

I would say more Paddington than Ealing Broadway; not many people get off at Paddington Low Level but trains arriving are usually full & standing with loads more getting on. It helps that Paddington has the turn back so there are empty trains to absorb the crowds!

It depends which peak you are talking about; AM peak highest loading is between Ealing Broadway and Paddington thus making the former the “peak” loading. Reverse in the PM peak.

== Doublepost prevention - post automatically merged: 31 Jul 2026 ==

That isn't true though. The 158s and 159s have been concentrated on fewer routes in that time, as 170s, 175s, 196s, 197s and displacement from electrification has allowed that to happen. 34 years ago, 158s worked primarily in 2-car formations on alnost all their routes. Cascades have provided increased capacity.

Indeed such length trains were also common into London Terminals. Although not mentioned here 165/166s routinely worked cars in sometimes max 3 (but often more) car trains, now supplanted by 9-car Elizabeth line services.

 

[Magdalia]

However it would not surprise me if there are more rail travellers to Reading to work than leave it. I suppose it might be possible to work this out using the Origin and Destination Matrix, but I haven't got the energy!

The dataset published last Tuesday is better for this than the origin and destination matrix, because it has the hourly breakdown through the day.

Earlier I devised a way of extracting the hourly arrivals and departures at Cambridge in 2025, to produce figures for net arrivals in each hour, so it was easy to repopulate my spreadsheet with the Reading numbers. And they are a bit strange!

In the evening peak 1600-1859 arrivals and departures at Reading are almost in balance over the 3 hours, but there are net departures 1600-1659 and net arrivals 1700-1859.

In the morning peak 0700-0959 departures exceed arrivals, but that's by a big margin 0700-0759, offset by arrivals slightly higher than departures 0800-0959.

I think that there are probably still more travellers leaving than entering Reading for work, but the gap isn't very big.