Scotrail Delay-Repay & Performance Data Issues

[RodMcMillan]

Hello All,

I’m a journalist investigating issues with ScotRail's Delay Repay scheme and the performance data it relies on. I’ve come across several concerning incidents:

• Payments were made for delays on certain journeys, even though the train was not officially recorded as delayed.
• Payments were refused, appealed, and then the appeals were also rejected because the system could not identify any delay. ( later escalated with proof the train was delayed)
• ScotRail has stated that Network Rail human error leads to incorrect performance data, meaning some delays were not recorded accurately. Impacting both performance measures and delay-repay approvals.
• In some cases, claims for the same delayed journey were approved for some passengers but refused for others.

First – When the "Darwin" system is updated with live data from the "TRUST" system, errors are less likely in areas where track circuits or axle counters provide automatic reporting. However, in regions relying on traditional signalling methods—such as north of Perth and Dundee, the West Highland Line, Ayr to Stranraer, and Barrhead to Gretna Green—the performance and delay data are collected manually. This data is not cross-referenced with technological checks, increasing the risk of inaccuracies.

Second – Human error in the processing of Delay Repay claims by customer services. Mistakes in approving or denying claims suggest inconsistencies in how staff assess and handle applications, further complicating the process for passengers.

I’m keen to hear from anyone who has experienced similar issues or inconsistencies with Delay Repay, or who can shed any light on the systems and processes. Do you have any stories, insights, or thoughts on w these problems? Your input could help shed light on what’s going on and ultimately inform the story.

Thank you in advance for sharing your experiences!

 

[yorkie]

Hello All,

I’m a journalist investigating issues with ScotRail's Delay Repay scheme and the performance data it relies on. I’ve come across several concerning incidents:

• Payments were made for delays on certain journeys, even though the train was not officially recorded as delayed.

Can you elaborate on this?

• Payments were refused, appealed, and then the appeals were also rejected because the system could not identify any delay. ( later escalated with proof the train was delayed)

This is absolutely rife among many train companies. When I put in a claim, I almost expect it to be rejected.

• ScotRail has stated that Network Rail human error leads to incorrect performance data, meaning some delays were not recorded accurately. Impacting both performance measures and delay-repay approvals.

Do you have any examples of this?

• In some cases, claims for the same delayed journey were approved for some passengers but refused for others.

Again, this is absolutely rife and normal.

First – When the "Darwin" system is updated with live data from the "TRUST" system, errors are less likely in areas where track circuits or axle counters provide automatic reporting. However, in regions relying on traditional signalling methods—such as north of Perth and Dundee, the West Highland Line, Ayr to Stranraer, and Barrhead to Gretna Green—the performance and delay data are collected manually. This data is not cross-referenced with technological checks, increasing the risk of inaccuracies.

I've not come across any examples of Delay Repay claims being rejected for this reason, however a sensible approach would be for Scotrail to have a sensible margin of error for such journeys if the arrival times are not being recorded accurately.

Second – Human error in the processing of Delay Repay claims by customer services. Mistakes in approving or denying claims suggest inconsistencies in how staff assess and handle applications, further complicating the process for passengers.

You will never get 100% consistency, but many train companies are in the habit of rejecting significant numbers of claims. It is common for train companies to make things up, such as "you didn't travel, request a refund instead", though I am not sure if Scotrail specifically do this. In cases where there was more than one operator's train involved in the journey, it's not uncommon for two companies to argue that the other company is to blame.

I’m keen to hear from anyone who has experienced similar issues or inconsistencies with Delay Repay, or who can shed any light on the systems and processes. Do you have any stories, insights, or thoughts on w these problems? Your input could help shed light on what’s going on and ultimately inform the story.

Thank you in advance for sharing your experiences!

Have you read the many threads on this forum?

I made a thread for this purpose:

Delay Repay claims rejected

I am hearing a lot of reports of Delay Repay claims being rejected for spurious reasons. Some TOCs have a policy of rejecting claims until you have taken another photograph of the ticket showing the ticket has been "spoiled'' to "prevent you using it again" even when the relevant ticket/portion...

www.railforums.co.uk

But many people have also set up their own threads for their specific issues.

Only a tiny proportion of rejected claims would be posted here, so this is the tip of the iceberg.

 

[Hadders]

One of the issues is train companies use automated systems to decide delay repay claims. This works well for simple journeys but claims often get rejected where there is anything slightly complicated. If an appeal is made it gets looked at by a human and will often be paid.

The problem is I suspect most people don’t actually bother to appeal.

 

[RodMcMillan]

Can you elaborate on this?

This is absolutely rife among many train companies. When I put in a claim, I almost expect it to be rejected.

Do you have any examples of this?

Again, this is absolutely rife and normal.

I've not come across any examples of Delay Repay claims being rejected for this reason, however a sensible approach would be for Scotrail to have a sensible margin of error for such journeys if the arrival times are not being recorded accurately.

You will never get 100% consistency, but many train companies are in the habit of rejecting significant numbers of claims. It is common for train companies to make things up, such as "you didn't travel, request a refund instead", though I am not sure if Scotrail specifically do this. In cases where there was more than one operator's train involved in the journey, it's not uncommon for two companies to argue that the other company is to blame.

Have you read the many threads on this forum?

I made a thread for this purpose:

Delay Repay claims rejected

I am hearing a lot of reports of Delay Repay claims being rejected for spurious reasons. Some TOCs have a policy of rejecting claims until you have taken another photograph of the ticket showing the ticket has been "spoiled'' to "prevent you using it again" even when the relevant ticket/portion...

www.railforums.co.uk

But many people have also set up their own threads for their specific issues.

Only a tiny proportion of rejected claims would be posted here, so this is the tip of the iceberg.

Thank You, and thank you for the threads, will dig into them.
Yes, and I'm giving this explanation at face value: it's come out over several stages and Occam's razor would suggest something simpler.

Glasgow to Mallaig, leaving 12:21
10th September 2024
Arrived Mallaig over 30 minutes late.

5 delay repay claims were made.
4 were approved and paid.
1 was rejected, appealed, and rejected again saying there was no record that the train was delayed.

Here's where it gets convoluted:

Scotrail advised that, although technically the train was delayed, the claim that was rejected was the one that was handled correctly - because Network Rail human error had given them the wrong arrival time, so they had not recorded a delay.

So when it then came out that the first four had been approved, ScotRail say that was their Human error and should not have been approved because their system did not show a delay.

 

[RodMcMillan]

One of the issues is train companies use automated systems to decide delay repay claims. This works well for simple journeys but claims often get rejected where there is anything slightly complicated. If an appeal is made it gets looked at by a human and will often be paid.

The problem is I suspect most people don’t actually bother to appeal.

Thanks, I have 2 years of data I'm working through as part of the investigation, on rejections, the reasons, appeals and successful appeals. It's looking like around 46% of appeals are successful.

 

[WesternLancer]

One of the issues is train companies use automated systems to decide delay repay claims. This works well for simple journeys but claims often get rejected where there is anything slightly complicated. If an appeal is made it gets looked at by a human and will often be paid.

The problem is I suspect most people don’t actually bother to appeal.

Making it easy to appeal does help. I don’t know how easy ScotRail d-r appeals are, but of those I regularly claim for EMR appeals are straightforward forward to do. By contrast Thameslink appeals are hard to do and very little information given by them on how to make an Appeal. I suspect thameslink thus get disproportionately lower numbers of Appeals submitted.

 

[SuspectUsual]

Glasgow to Mallaig, leaving 12:21
10th September 2024
Arrived Mallaig over 30 minutes late.

5 delay repay claims were made.
4 were approved and paid.
1 was rejected, appealed, and rejected again saying there was no record that the train was delayed.

Here's where it gets convoluted:

Scotrail advised that, although technically the train was delayed, the claim that was rejected was the one that was handled correctly - because Network Rail human error had given them the wrong arrival time, so they had not recorded a delay.

So when it then came out that the first four had been approved, ScotRail say that was their Human error and should not have been approved because their system did not show a delay.

Service Information - Recent Train Times

Here is the data from Recent Train Times - looks like it got delayed but caught up time and arrived OK. Are you saying this is totally wrong?

 

[RodMcMillan]

Service Information - Recent Train Times

Here is the data from Recent Train Times - looks like it got delayed but caught up time and arrived OK. Are you saying this is totally wrong?

Yep. 100% wrong.

== Doublepost prevention - post automatically merged: 11 Dec 2024 ==

There are two parts of the investigation / Story - 1 - on Delay-Repay claims being wrongly refused, that's more likely to get the initial coverage, but actually I think the bigger story is wrong performance data being recorded.

== Doublepost prevention - post automatically merged: 11 Dec 2024 ==

Yep. 100% wrong.

== Doublepost prevention - post automatically merged: 11 Dec 2024 ==

There are two parts of the investigation / Story - 1 - on Delay-Repay claims being wrongly refused, that's more likely to get the initial coverage, but actually I think the bigger story is wrong performance data being recorded.

That service was stopped at Arisaig to let a steamtrain, I believe the Jacobite, pass - it didn't leave Arisaig until after 18:15

== Doublepost prevention - post automatically merged: 11 Dec 2024 ==

Yep. 100% wrong.

== Doublepost prevention - post automatically merged: 11 Dec 2024 ==

There are two parts of the investigation / Story - 1 - on Delay-Repay claims being wrongly refused, that's more likely to get the initial coverage, but actually I think the bigger story is wrong performance data being recorded.

== Doublepost prevention - post automatically merged: 11 Dec 2024 ==


That service was stopped at Arisaig to let a steamtrain, I believe the Jacobite, pass - it didn't leave Arisaig until after 18:15

So the other thing I am trying to find out, and this will take digging, is instances where a train was delayed, people put in a claim, the claim was approved... but the official performance data says the train arrived on time! Huge implications.

 

[David M]

I'm guessing you simply want to have a politically motivated go at Scotrail.

 

[RodMcMillan]

There are issues with ScotRail, Currently the source of the performance data error seems to sit with human error at Network Rail on routes where the arrival and departure tracking isn't automated. It's possible, or easier, to cross-reference that with ScotRail data which is subject to greater transparency than private companies, but if there's a systemic fault in how performance data is logged by NR that could impact UK wide.

 

[AlterEgo]

I am not certain we have seen many issues here about incorrect timing recording causing DR claims to fail. What evidence do you have of the example train being wrongly recorded? The recording doesn't *seem* suspicious on the face of it, but I'd be interested to know why you feel sure it's wrong.

 

[RodMcMillan]

I am not certain we have seen many issues here about incorrect timing recording causing DR claims to fail. What evidence do you have of the example train being wrongly recorded? The recording doesn't *seem* suspicious on the face of it, but I'd be interested to know why you feel sure it's wrong.

Sure - it's that specific train journey that started the investigation, and I'm 100% sure the arrival time is recorded incorrectly because I was on that train. I was recording on the train for a travel piece on "Scotland's most scenic journey", and I had 40+ RAW photos and video, with embedded GPS location data recording the delay. Our delay claim was refused, we appealed, and it was refused again - Scotrail said they had no record of the delay - When I advised Scotrail I was a journalist, I had digital proof the train was delayed, I had four eyewitnesses and submitted an FOI for the technical data - they approved the claim saying it was a human error from Network Rail that gave them the wrong performance data and they hoped that settled the matter and would drop the FOI. That's when I started digging - because it's unlikely that would only happen once. FOI then uncovered that were 5 claims made on that journey, and 4 of them were approved - only ours was denied. So - if they are paying delay claims but recording the train as 'on time' in the performance data, that's an issue. If they deny claims because they have wrongly recorded the train as one time - that's another issue. I want to see how big or small an issue is by identifying how many claims / Journey times are impacted.

 

[AlterEgo]

Sure - it's that specific train journey that started the investigation, and I'm 100% sure the arrival time is recorded incorrectly because I was on that train. I was recording on the train for a travel piece on "Scotland's most scenic journey", and I had 40+ RAW photos and video, with embedded GPS location data recording the delay. Our delay claim was refused, we appealed, and it was refused again - Scotrail said they had no record of the delay - When I advised Scotrail I was a journalist, I had digital proof the train was delayed, I had four eyewitnesses and submitted an FOI for the technical data - they approved the claim saying it was a human error from Network Rail that gave them the wrong performance data and they hoped that settled the matter and would drop the FOI. That's when I started digging - because it's unlikely that would only happen once. FOI then uncovered that were 5 claims made on that journey, and 4 of them were approved - only ours was denied. So - if they are paying delay claims but recording the train as 'on time' in the performance data, that's an issue. If they deny claims because they have wrongly recorded the train as one time - that's another issue. I want to see how big or small an issue is by identifying how many claims / Journey times are impacted.

Out of interest, what caused the delay, and where did it occur? As whoever has recorded those timings has at least made it look "natural".

 

[RodMcMillan]

Out of interest, what caused the delay, and where did it occur? As whoever has recorded those timings has at least made it look "natural".

The delay was at Arisaig, the train was stopped at the station for around 30 minutes while we waited for another train to pass - a Steamtrain, I believe it was the 'Jacobite'. We left Arisaig around 18:15,

 

[The_Train]

After a recent 4 day tour of Scotland using a rover ticket, I raised 5 separate delay repay claims. 4 of them were approved without issue within a few days and 1 was rejected because I'd loaded some incorrect information. After I corrected the incorrect information, the claim was processed and paid out within 48 hours.

The service I received from the ScotRail social media team in assisting me in raising these and the actual process of completing the claims and receiving my payments was better than most other operators I have claimed from in the past.

Of course, the OP is a journalist seeking to slam ScotRail into the ground and won't want to be printing anything that might be deemed as remotely positive

 

[RodMcMillan]

After a recent 4 day tour of Scotland using a rover ticket, I raised 5 separate delay repay claims. 4 of them were approved without issue within a few days and 1 was rejected because I'd loaded some incorrect information. After I corrected the incorrect information, the claim was processed and paid out within 48 hours.

The service I received from the ScotRail social media team in assisting me in raising these and the actual process of completing the claims and receiving my payments was better than most other operators I have claimed from in the past.

Of course, the OP is a journalist seeking to slam ScotRail into the ground and won't want to be printing anything that might be deemed as remotely positive

I'm glad you feel you had a positive experience. Objectively I'm not sure that having to make five delay compensation claims in just four days is the ringing endorsement of their efficiency you think it is.

 

[WesternLancer]

I'm glad you feel you had a positive experience. Objectively I'm not sure that having to make five delay compensation claims in just four days is the ringing endorsement of their efficiency you think it is.

The delay you mention of course seems to have been caused by a private open access operator (the Jacobite). It would be worth you asking if ScotRail received any compensation payment from the operator of the Jacobite, since that would be the basis on which ScotRail passengers delay repay compo should be paid with if that caused the inconvenience.

To my mind that is part of the story worth writing about as a journalist.

 

[AlterEgo]

The delay you mention of course seems to have been caused by a private open access operator (the Jacobite). It would be worth you asking if ScotRail received any compensation payment from the operator of the Jacobite, since that would be the basis on which ScotRail passengers delay repay compo should be paid with if that caused the inconvenience.

To my mind that is part of the story worth writing about as a journalist.

The delay might not be attributable to the Jacobite, however. We don’t know what caused that train to be 30 minutes late.

 

[furlong]

Further evidence is that the return working didn't leave till 19:00 (instead of 18:15).

 

[Indigo Soup]

Of course, the OP is a journalist seeking to slam ScotRail into the ground and won't want to be printing anything that might be deemed as remotely positive

I'm not sure that's the case at all here. It does look like OP has experienced something odd and started digging.

The fate of one particular train to Mallaig isn't the story - the story is in the potential systematic shortcomings of the performance data used to determine compensation.

It's possible, or easier, to cross-reference that with ScotRail data which is subject to greater transparency than private companies, but if there's a systemic fault in how performance data is logged by NR that could impact UK wide.

If the problem is only arising on certain routes that don't have modern signalling technology, I wonder how many routes are actually affected.

 

[Haywain]

Objectively I'm not sure that having to make five delay compensation claims in just four days is the ringing endorsement of their efficiency you think it is.

Aren't you seeking information on the handling of Delay Repay claims? Or will you ignore the positive because it doesn't suit your agenda?

 

[SuspectUsual]

There are issues with ScotRail, Currently the source of the performance data error seems to sit with human error at Network Rail on routes where the arrival and departure tracking isn't automated. It's possible, or easier, to cross-reference that with ScotRail data which is subject to greater transparency than private companies, but if there's a systemic fault in how performance data is logged by NR that could impact UK wide.

To put this in perspective, if there are issues like you describe, they’ll be limited to the routes which don’t have more modern technology that automatically reports times, and they’re invariably those with fewer services and fewer passengers.

And if mistakes are being made, they won’t be made on every service on the impacted routes unless you’re suggesting it’s deliberate.

So obviously far from ideal, but the impact of the problems will be a tiny proportion of the trains that operate on the network.

 

[RodMcMillan]

Further evidence is that the return working didn't leave till 19:00 (instead of 18:15).

Thanks: I also have it in writing from ScotRail, I FOI'd various things including communications about the delay, updates to Arrival / Departure boards, and details of driver logs. I'd need to check I think they admitted arrived at 18:33, Still waiting on Network rail FOI on tokens / signals on the single track.

== Doublepost prevention - post automatically merged: 11 Dec 2024 ==

To put this in perspective, if there are issues like you describe, they’ll be limited to the routes which don’t have more modern technology that automatically reports times, and they’re invariably those with fewer services and fewer passengers.

And if mistakes are being made, they won’t be made on every service on the impacted routes unless you’re suggesting it’s deliberate.

So obviously far from ideal, but the impact of the problems will be a tiny proportion of the trains that operate on the network.

Yes, it might be minute. I'm looking to find out the scope. If there are not a large number of people who have had delay-repay claims refused then let's find out how many people have had delay claims approved - but the delay isn't shown in the PPM. If sone late trains aren't being shown in the Performance data, and they aren't, that's worth digging into.

== Doublepost prevention - post automatically merged: 11 Dec 2024 ==

Out of interest, what caused the delay, and where did it occur? As whoever has recorded those timings has at least made it look "natural".

Looking at that data - the times in italics are just estimates. If only the arrival time at Mallaig was manually entered, then the italics are all estimates from between then and the last non italicised departure time.

 

[Falcon1200]

So the other thing I am trying to find out, and this will take digging, is instances where a train was delayed, people put in a claim, the claim was approved... but the official performance data says the train arrived on time!

they approved the claim saying it was a human error from Network Rail that gave them the wrong performance data and they hoped that settled the matter and would drop the FOI. That's when I started digging

Have any other such wildly incorrect arrival times actually been discovered? Clearly in your case a mistake was made, but in my experience such things are incredibly rare, requiring whoever input the wrong time to have been completely unaware that a delay had occurred and to have assumed an on-time arrival, and for no-one else to have noticed this.

 

[eoff]

Although not Scotrail I did a journey in 2022 where my train arrived over 30 minutes late (time taken as train stopped moving using my watch which was never more than 2 seconds fast or slow). The online system (real time trains?) said the train arrived 29 minutes late so I do wonder if the timing point or system used is skewed to have an earlier arrival time than what a normal person would use at some locations.

 

[Haywain]

I do wonder if the timing point or system used is skewed to have an earlier arrival time than what a normal person would use at some locations.

That's nothing to do with it, the times at Mallaig are entered manually. The example that the OP is referring to would appear to be a rather large error, but from my experience of other parts of the network this is very unusual.

 

[eoff]

That's nothing to do with it, the times at Mallaig are entered manually. The example that the OP is referring to would appear to be a rather large error, but from my experience of other parts of the network this is very unusual.

I explicitly said my example was not related to ScotRail.

I'm replying to the gerneral request...

I’m keen to hear from anyone who has experienced similar issues or inconsistencies with Delay Repay, or who can shed any light on the systems and processes. Do you have any stories, insights, or thoughts on w these problems? Your input could help shed light on what’s going on and ultimately inform the story.

 

[RodMcMillan]

Have any other such wildly incorrect arrival times actually been discovered? Clearly in your case a mistake was made, but in my experience such things are incredibly rare, requiring whoever input the wrong time to have been completely unaware that a delay had occurred and to have assumed an on-time arrival, and for no-one else to have noticed this.

One other so far.

== Doublepost prevention - post automatically merged: 14 Dec 2024 ==

Although not Scotrail I did a journey in 2022 where my train arrived over 30 minutes late (time taken as train stopped moving using my watch which was never more than 2 seconds fast or slow). The online system (real time trains?) said the train arrived 29 minutes late so I do wonder if the timing point or system used is skewed to have an earlier arrival time than what a normal person would use at some locations.

That's interesting, thanks.

 

[Falcon1200]

One other so far.

Thanks. Can you advise how many times were checked, where that one occurred, and by how much the time input was wrong?

 

[RodMcMillan]

Thanks. Can you advise how many times were checked, where that one occurred, and by how much the time input was wrong?

Network Rail advised me
"Errors like this are incredibly rare, to our knowledge there has only been one instance of this occurring anywhere on the network in the period referred to in your request. This
occurred on 31 October 2024, the Loch Eil outbound service from Glasgow Queen Street (GLQ) to Mallaig station (MLG) was reported as arriving at 17:00 when in fact it arrived at 17:43. While there has been no formal investigation of the matter, the records we hold indicate that the mistake was the result of human error.

We recognise that it’s possible albeit unlikely that there are further instances of this type
have occurred but since we take plans to rapidly identify and correct any errors in our
data, there are no records that would confirm the existence (or non-existence) of other
cases."

So couple of interesting things from that - first, it's at the same journey / station within a month, but their search didn't show up the initial journey I noticed; I've flagged this to them, as they are still showing it arrived on time. It appears from this that they know human error can happen, but if they apparently fix them there's no record kept of the mistake, and there's no cross referencing of human input against available technical data to check it. Also + shows that when ScotRail are aware of errors - admitting a train was delayed and then paying out - they are not taking steps to correct the recorded performance data. I'm not saying mistakes like this aren't rare, but it's bad form if they are not being checked or tracked when they are clearly possible. Unless there is something unique about Mallaig where it's happened twice there within a month, but never anywhere else since April 2022?

== Doublepost prevention - post automatically merged: 15 Dec 2024 ==

Network Rail advised me
"Errors like this are incredibly rare, to our knowledge there has only been one instance of this occurring anywhere on the network in the period referred to in your request. This
occurred on 31 October 2024, the Loch Eil outbound service from Glasgow Queen Street (GLQ) to Mallaig station (MLG) was reported as arriving at 17:00 when in fact it arrived at 17:43. While there has been no formal investigation of the matter, the records we hold indicate that the mistake was the result of human error.

We recognise that it’s possible albeit unlikely that there are further instances of this type
have occurred but since we take plans to rapidly identify and correct any errors in our
data, there are no records that would confirm the existence (or non-existence) of other
cases."

So couple of interesting things from that - first, it's at the same journey / station within a month, but their search didn't show up the initial journey I noticed; I've flagged this to them, as they are still showing it arrived on time. It appears from this that they know human error can happen, but if they apparently fix them there's no record kept of the mistake, and there's no cross referencing of human input against available technical data to check it. Also + shows that when ScotRail are aware of errors - admitting a train was delayed and then paying out - they are not taking steps to correct the recorded performance data. I'm not saying mistakes like this aren't rare, but it's bad form if they are not being checked or tracked when they are clearly possible. Unless there is something unique about Mallaig where it's happened twice there within a month, but never anywhere else since April 2022?

*Within a month and a half roughly.