Report of collision Bedford (Elstow) - 19/06/26

[Class455]

A very sad day for the railway. Thoughts go out to everyone involved in the incident and to everyone at EMR.

 

[Towers]

Excellent point. Even with any camera footage, the worst outcome I can think of for the investigation is if all the data is reviewed and shows no signalling irregularities and the train being controlled appropriately to the conditions, up to the point of collision. Under those circumstances the only person who could explain the incident would be the driver themselves.

With respect, both of those things cannot be simultaneously true - either of them (or both) must have become an issue at some stage before ‘the point of collision’ - but hopefully of course the investigation will clarify.

 

[tram21]

Sadly, this will also make it harder to find out what caused this accident, as the driver of the moving train would otherwise have been a vital witness to the state of the signals and the behaviour of the train.

Without going into too much detail, and I'm sure more info will become available soon about this, this won't be a problem and there will be good evidence here.

 

[BAFRA77]

Condolences to the family and colleagues of the driver who has sadly lost their life.

 

[yorkie]

This thread will shortly be locked overnight.

Also, until further notice, please can we only submit posts that contain actual updates and/or new information.

 

[Recessio]

The Guardian confirming it is one of the drivers who has sadly died, The RMT have issued a statement

https://www.theguardian.com/uk-news/2026/jun/19/two-trains-collide-bedford

A train driver has died and 89 people have been injured after two trains collided in the Bedford area, with emergency services working into the night.

Eddie Dempsey, general secretary of the RMT union, said: “We are devastated to learn that a train driver and former RMT rep has tragically died as a result of today’s crash between Luton and Bedford. The thoughts of RMT are with their family, friends, colleagues and the ASLEF trade union at this awful time.”

While the circumstances remain unclear, the scale and nature of the crash, involving new trains on the upgraded Midland mainline, will be a deep shock for the rail industry.

It is the first fatal collision between two passenger trains on a British mainline this century, with what appears in early reports to involve the highest number of serious injuries in any UK rail accident in two decades.


Accident investigators will be examining why one train had halted and the other failed to stop, given the safety systems put in place in the wake of a series of crashes in the late 1990s and early 2000s.

My thoughts to the family and friends of the driver and everybody injured, and my thanks to the railway staff and emergency services at the scene.

Moderator note: The Guardian article contained an incorrect statement; reference to this have been removed from the thread.

 

[liamf656]

The Guardian confirming it is one of the drivers who has sadly died, in the first fatal collision between two trains on a Mainline in this century. The RMT have issued a statement. https://www.theguardian.com/uk-news/2026/jun/19/two-trains-collide-bedford

RIP to one of our colleagues on this dark day. Heartbreaking

EMR are running a pre planned amended service with an engineering block south of Bedford this weekend. This should run as planned so the incident will cause no further alterations today (20th) and tomorrow (21st), for those wondering

 

[Snow1964]

Whilst this thread was locked overnight British Transport Police issue an update

Officers were called to a collision between two East Midlands Railway trains on the railway line in Elstow at around 5.15pm yesterday (19 June).

Tragically, one person was pronounced dead at the scene. He is the driver of one of the trains involved and his family have been informed. Our thoughts are with his loved ones and with our colleagues at East Midlands Trains.

33 people were taken to hospital with 11 of those in a serious condition. 56 other people were treated for injuries.

We are being supported by colleagues from Bedfordshire Police, the local Fire and Rescue and Ambulance Services. We are also receiving invaluable support from colleagues in the National Police Air Service, Network Rail, East Midlands Railway and the local authority.

We know how concerning this incident will be and can assure people that officers will do everything we they can to establish.

We are working with Network Rail and East Midlands Railway colleagues to keep disruption to a minimum as much as we possibly can, and we are grateful for people’s patience.

https://www.btp.police.uk/news/btp/news/england/update-following-train-crash---bedford/

 

[westcoaster]

RIP to my diving colleague from EMR.
And continued thoughts to all of those injured and traumatised by this tragic accident.

A dark day on the railway, and we will all feel this. Just seems unreal in 2026 and especially over the route I drive multiple times daily.

 

[Skimpot flyer]

Condolences to the family of the driver and my best wishes to those injured.

 

[TrainSailing]

It should be noted that the now obsolete GWML ATP system is far easier to overlay on an existing signalling system than TPWS is for plain line signals like automatic signals.

I believe the ATP on the GWML is still being used by the class 80Xs.

On another note. Signallers, when a train SPADs a signal is it automatically sent a GSMR Emergency stop message or is this sent manually by the signaller to the train concerned?

 

[bengley]

I believe the ATP on the GWML is still being used by the class 80Xs.

On another note. Signallers, when a train SPADs a signal is it automatically sent a GSMR Emergency stop message or is this sent manually by the signaller to the train concerned?

Railway Emergency Calls are not triggered automatically in any circumstances

 

[SWTurbo]

I believe the ATP on the GWML is still being used by the class 80Xs.

On another note. Signallers, when a train SPADs a signal is it automatically sent a GSMR Emergency stop message or is this sent manually by the signaller to the train concerned?

In many locations there’s little to no indication a SPAD has even taken place. Modern VDU units have SPAD alarms, or TC occupation alarms, but I’ve had a SPAR before due to a TPWS fault (signal reverted), and there was no indication on the panel at all. First I knew of it was the driver calling me up as I’d been dealing with an issue at the other end of the panel

 

[TrainSailing]

In many locations there’s little to no indication a SPAD has even taken place. Modern VDU units have SPAD alarms, or TC occupation alarms, but I’ve had a SPAR before due to a TPWS fault (signal reverted), and there was no indication on the panel at all. First I knew of it was the driver calling me up as I’d been dealing with an issue at the other end of the panel

Wow thank you for the insight I did not know that. I had assumed all SPADs (at least on major routes) were picked up almost instantly by the signaller with a plethora of bells and whistles as I witnessed many years ago on a visit to Liverpool Street IECC.

 

[Dan G]

So it's likely that the 360 passed a signal at danger, right? An IET doesn't stop on a dime so the signal protecting it should have been red for some time, and the one before it at caution. Anyone have an idea of how up the line the 810 was from the last signal?

The track layout is such that the 360 saw the caution? At what speed would a driver normally proceed at on that route at that time? The 360 clearly was far below line speed (damage suggests 20-30 mph relative impact speeds; I assume the 810 was stationary), but I haven't read much eye witnesses talk of the train being under heavy braking (unless I've missed it).

Obviously such an event like this, which should never happen, raises many questions. Those are mine right now. (And yes I will wait for the RAIB report, I've been reading the Cambrian line crash report this week.)

 

[yorkie]

So it's likely that the 360 passed a signal at danger, right? An IET doesn't stop on a dime so the signal protecting it should have been red for some time, and the one before it at caution. Anyone have an idea of how up the line the 810 was from the last signal?

The track layout is such that the 360 saw the caution? At what speed would a driver normally proceed at on that route at that time? The 360 clearly was far below line speed (damage suggests 20-30 mph relative impact speeds; I assume the 810 was stationary), but I haven't read much eye witnesses talk of the train being under heavy braking (unless I've missed it).

Obviously such an event like this, which should never happen, raises many questions. Those are mine right now. (And yes I will wait for the RAIB report, I've been reading the Cambrian line crash report this week.)

Yes, let's not go into any unanswered questions right now.

But @Poggs did post the known element of the sequence of events here:

Peter Hicks (@poggs) on X

Some context on the events at Elstow/Bedford this evening. There is a lot you can tell from train describer data on @OpenTrainTimes, however there is a lot of information that you cannot determine. The purple trace is the Class 360's movements, the green is the 810's.

x.com

Some context on the events at Elstow/Bedford this evening. There is a lot you can tell from train describer data on @OpenTrainTimes , however there is a lot of information that you cannot determine. The purple trace is the Class 360's movements, the green is the 810's.

My understanding is that what happened is as described below by @800301:

From the signalling replay I've seen on an internal system (Berth Maps), the 360 was approaching a red at WH356, which stepped up on approach to a single yellow; WH154 was held at danger as the section ahead was occupied by the 810 and WH154 was passed at danger.

(and if anyone is wondering, this information has been confirmed as correct by multiple people in the industry)

Note 154 is an automatic signal and is not fitted with TPWS, however it appears that there has been some suggestion from Nigel harris that the signal was fitted with TPWS which was debunked here:

Bengley (@Bengley) on X

@railnigel Hi Nigel. You seemed to suggest that TPWS stops trains passing red signals on Sky News just now. Just to make you aware that the signal protecting the EMR 810 (the front train) is not fitted with TPWS as it's an automatic signal (most of which are not fitted).

x.com

@railnigel Hi Nigel. You seemed to suggest that TPWS stops trains passing red signals on Sky News just now. Just to make you aware that the signal protecting the EMR 810 (the front train) is not fitted with TPWS as it's an automatic signal (most of which are not fitted).

(And yes, if anyone is wondering, the statement by bengley correcting Nigel Harris is correct, and confirmed by multiple sources)

Yes that leaves unanswered questions, but can I please ask that people don't speculate. I am sure that the unanswered questions will be answered in due course.

 

[Annetts key]

I believe the ATP on the GWML is still being used by the class 80Xs.

Yes, the GWML ATP is still operational, both the track mounted and track side equipment and the class 80X trains. I use the term obsolete because it's not realistic to fit it anywhere else (the technology is out of date and it would have to be re-engineered using modern semiconductor parts), the better solution is to fit ETCS/ERTMS.

On another note. Signallers, when a train SPADs a signal is it automatically sent a GSMR Emergency stop message or is this sent manually by the signaller to the train concerned?

SPAD alarm equipment is a relativity new thing. Unless there has been a determination of a requirement to provide it because a signal has been identified as high risk, there will be no such equipment or systems fitted to conventional interlocking systems (relay / route relay / geographical etc.).

On these signalling systems, the only indication to the signaller that a train has passed a signal at red are noticing that either the track circuit(s) / axle counter section(s) beyond a red signal has/have become occupied when it/they should not have, or the berth track circuit / axle counter (that is on approach and leading up to the signal) has gone clear when it should not have.

Sometimes the Train Describer can help, but it depends on how it gets it's stepping information.

 

[800301]

So it's likely that the 360 passed a signal at danger, right? An IET doesn't stop on a dime so the signal protecting it should have been red for some time, and the one before it at caution. Anyone have an idea of how up the line the 810 was from the last signal?

The track layout is such that the 360 saw the caution? At what speed would a driver normally proceed at on that route at that time? The 360 clearly was far below line speed (damage suggests 20-30 mph relative impact speeds; I assume the 810 was stationary), but I haven't read much eye witnesses talk of the train being under heavy braking (unless I've missed it).

Obviously such an event like this, which should never happen, raises many questions. Those are mine right now. (And yes I will wait for the RAIB report, I've been reading the Cambrian line crash report this week.)

From the signalling replay I've seen on an internal system (Berth Maps), the 360 was approaching a red at WH356, which stepped up on approach to a single yellow; WH154 was held at danger as the section ahead was occupied by the 810 and WH154 was passed at danger.

A few drivers have said WH154 is a signal that 'normally' clears pretty quickly, but I'm not going to speculate.

 

[capital12]

Whilst it’s probably still true as a statistic that we’ve one of the safest railways I can’t help but feel there has been a significant/worrying increase in these bigger incidents and we’ve been on a downward trajectory since 2020 - Carmont, Salisbury, the Welsh one and now Bedford. Whereas I can only really think of the Desiro at Watford in the whole of the 2010s but I may be mistaken?

Luckily total fatalities have been low but in different circumstances these could have been much higher, especially at Carmont had it been in non COVID times.

Are we just in an unlucky phase at the moment or are there other reasons for this unwelcome increase in major incidents?

The other thing I find worrying is how the Hitachi trains react to a shunt - the low speed one at Neville Hill caused that one to derail and this seems to have done the same, but to a larger extent (obviously with the larger forces involved) and as has been mentioned by someone earlier in the thread, it was just lucky nothing was coming the other way as the derailed 810 coach could have been fowling another line.

I know it’s hard to say for sure but would, for example a 222 or HST have reacted differently and more likely stayed on the track?

Thoughts to those involved and hopefully all make a speedy recovery.

 

[bramling]

So it's likely that the 360 passed a signal at danger, right?

There is an automatic signal in the rear of the 810, so yes it’s fair to say that providing everything was functioning correctly then the 360 would have passed a signal at danger. It’s also of note that from what is visible in open data, the train description for the Corby service doesn’t appear to have stepped forwards into the next berth, which is consistent with the above.

An IET doesn't stop on a dime so the signal protecting it should have been red for some time, and the one before it at caution. Anyone have an idea of how up the line the 810 was from the last signal?

The 810 certainly finished up some way into the signal section, it wasn’t immediately beyond the end of the overlap. We don’t know to what extent it may have been pushed along the track by the impact.

The track layout is such that the 360 saw the caution? At what speed would a driver normally proceed at on that route at that time? The 360 clearly was far below line speed (damage suggests 20-30 mph relative impact speeds; I assume the 810 was stationary), but I haven't read much eye witnesses talk of the train being under heavy braking (unless I've missed it).

50 mph over the crossovers, then pretty much free acceleration if under clear signals. Clearly it wasn’t doing 100-110mph at the time of impact, but there isn’t enough information at present to be able to say how the 360 was being driven.

 

[Trothy]

Its worth noting that even in locations with SPAD alarms, that if the automatic signal sections consisted if only a single Track Circuit/Axle Counter section (and for auto sections they often do), they wouldn't trigger for the second train, as the section ahead is occupied already (by the first train)

 

[357]

(And yes, if anyone is wondering, the statement by bengley correcting Nigel Harris is correct, and confirmed by multiple sources)

@bengley did post the same info in this thread at the time but I can't seem to find it now?

 

[ExRes]

I'm assuming that both the data recorder and forward facing camera from the 360 have been downloaded overnight and the information is available to the correct people, that information will give the facts of the situation and negate the need for any speculation

 

[Nicholas Lewis]

From the signalling replay I've seen on an internal system (Berth Maps), the 360 was approaching a red at WH356, which stepped up on approach to a single yellow; WH154 was held at danger as the section ahead was occupied by the 810 and WH154 was passed at danger.

A few drivers have said WH154 is a signal that 'normally' clears pretty quickly, but I'm not going to speculate.

WH356 with junction indicator position 4 will have some form of approach control on it given speed differential is >20mph so I wouldn't be surprised to see this open up another can of worms about how junction signalling operates in the UK and its impact on train handling.

 

[357]

An IET doesn't stop on a dime so the signal protecting it should have been red for some time

The amount of time the signal was at red is irrelevant - the signal won't go back to red because the 810 stopped, it goes to red when the 810 passes it and only steps up through the aspects as the train progresses along the line.

== Doublepost prevention - post automatically merged: 20 Jun 2026 ==

The other thing I find worrying is how the Hitachi trains react to a shunt - the low speed one at Neville Hill caused that one to derail and this seems to have done the same, but to a larger extent (obviously with the larger forces involved) and as has been mentioned by someone earlier in the thread, it was just lucky nothing was coming the other way as the derailed 810 coach could have been fowling another line.

I know it’s hard to say for sure but would, for example a 222 or HST have reacted differently and more likely stayed on the track?

I'd far rather be on a Hitachi than a Voyager, and rather be on a Voyager than a HST in any sort of collision.

The Hitachi units are apparently designed to do this to reduce the chances of damage elsewhere that could pose a higher risk to passengers. I'm sure they've calculated the risk of another train coming, and taken into account how quickly trains can be stopped these days via GSMR.

== Doublepost prevention - post automatically merged: 20 Jun 2026 ==

I'm assuming that both the data recorder and forward facing camera from the 360 have been downloaded overnight and the information is available to the correct people, that information will give the facts of the situation and negate the need for any speculation

The data recorder definitely has

 

[Poggs]

From the signalling replay I've seen on an internal system (Berth Maps), the 360 was approaching a red at WH356, which stepped up on approach to a single yellow; WH154 was held at danger as the section ahead was occupied by the 810 and WH154 was passed at danger.

If Berth Maps was showing a single yellow, then unless it has access to real-time data logging either trackside or from the interlocking, it's inferring what it expects the signal aspect to be.

The TD at West Hampstead will only indicate two states for a signal - "most restrictive aspect" or "not most restrictive aspect".

 

[Signal Head]

From the signalling replay I've seen on an internal system (Berth Maps), the 360 was approaching a red at WH356, which stepped up on approach to a single yellow; WH154 was held at danger as the section ahead was occupied by the 810 and WH154 was passed at danger.

A few drivers have said WH154 is a signal that 'normally' clears pretty quickly, but I'm not going to speculate.

Has it been absolutely confirmed that 154 was displaying a Red? Obviously it should have been, but note on the OTT map there is no aspect information.

Since any system displaying this type of 'TD' information ultimately gets its data from inputs from the signalling system, and for auto signals the state of the signal generally wasn't input as it wasn't considered necessary, the aspect may not be 'known' by any of them, internal or external.

 

[Snow1964]

I'm assuming that both the data recorder and forward facing camera from the 360 have been downloaded overnight and the information is available to the correct people, that information will give the facts of the situation and negate the need for any speculation

Correct, but the longer it takes to produce a preliminary or interim statement, speculation will continue (on other media and forums, if not this heavily moderated forum).

These days with rapid information dispersal (and lot of it unverified facts), there is a trade off time wise between stating not possible scenarios and investigating further

 

[yorkie]

Has it been absolutely confirmed that 154 was displaying a Red? Obviously it should have been, but note on the OTT map there is no aspect information.

I understand that this has not yet been confirmed.

The investigation will look into all possibilities.

Edit: just to add, the post you refer to can only be read in the context of what the replay system shows; no other inferences can be made, as the matter needs to be fully investigated.

It is often the case that knowledgeable internal railway sources divulge some relevant information before it's released, but we are keen to avoid speculation, and to stick to what is known. It is a difficult balancing act at times.

 

[Nicholas Lewis]

Correct, but the longer it takes to produce a preliminary or interim statement, speculation will continue (on other media and forums, if not this heavily moderated forum).

These days with rapid information dispersal (and lot of it unverified facts), there is a trade off time wise between stating not possible scenarios and investigating further

RAIB after more recent incidents have been fairly quick recently with a shortform overview of what they've found and intended course of their investigation as they are mindful of the wider social environment in todays world.