Official lateness calculation

[Mark J]

Yesterday I caught the 23:29 to Reading from Paddington. It was due to arrive at 00:05.

It left Paddington around six minutes late, crawling along in places around Slough and Twyford areas.

The late arrival time is documented at 00:19, it was actually 00:20 by the time the train was at a halt and the doors were unlocked for alighting.

It appears the lateness was calculated when the nose of the front of the train arrived into Reading, not when the train stopped and the doors were opened.

I find this a bit dishonest to be honest.

Is this part of the new DfT policy to cut down on Delay Repay compensation, to claim a train has arrived before it has actually stopped, and you are able to alight.

 

[AlterEgo]

Yesterday I caught the 23:29 to Reading from Paddington. It was due to arrive at 00:05.

It left Paddington around six minutes late, crawling along in places around Slough and Twyford areas.

The late arrival time is documented at 00:19, it was actually 00:20 by the time the train was at a halt and the doors were unlocked for alighting.

It appears the lateness was calculated when the nose of the front of the train arrived into Reading, not when the train stopped and the doors were opened.

I find this a bit dishonest to be honest.

Is this part of the new DfT policy to cut down on Delay Repay compensation, to claim a train has arrived before it has actually stopped, and you are able to alight.

No, it is just because timing points record the train automatically when it enters a berth. There is no record of when the doors were released linked to the signalling system.

 

[MrJeeves]

Darwin should apply an offset on top of the data from TRUST (which records the time at each timing point). It's not particularly clear what system operators use to calculate lateness, but I assume it's whatever works out to less compensation

 

[Dr Hoo]

‘Berthing offsets’ have been part of the system architecture of the TRUST performance recording system ever since it was invented by BR. They are individually documented and are for all practical intents and purposes ‘fixed’ unless there is re-modelling or re-signaling.
They cannot allow for human factors such as delayed door release.

 

[Belperpete]

Sounds better than the policy that airlines seem to use, wheels down equals arrival time.

 

[AF91]

‘Berthing offsets’ have been part of the system architecture of the TRUST performance recording system ever since it was invented by BR. They are individually documented and are for all practical intents and purposes ‘fixed’ unless there is re-modelling or re-signaling.
They cannot allow for human factors such as delayed door release.

They're audited by NR on a fairly regular basis to ensure they are as accurate as they can be. There's almost no way the TOCs can manipulate the data to avoid paying out.

 

[MrJeeves]

I was under the impression that there is also additional offsets on top of TRUST to account for the difference between entering a signalling berth and arriving at a station. Is that not right?

Got mixed up between TRUST and just generic TD data...

 

[redreni]

The arrangements for compensation in the event of delay are advertised on the websites of TOCs and independent ticket retailers, though.

The independent ticket retailers, in particular, presumably do this not because they're responsible for administering the schemes, but because they hope to influence consumers' decisions on whether to travel by train. Compensation for delays is one of the selling points of rail over road.

This page, for example, does not say that you might be deemed to have arrived before you arrived: https://www.thetrainline.com/trains/great-britain/delay-repay .
It simply doesn't say that the time the industry data says the train arrived is the yardstick even if it can be shown to be wrong.

If the passenger had some other evidence (e.g. a video with a timestamp set by their phone's clock, which in turn is regularly synced with a reliable time server, corroborated by the station clock and/or the clock on the platform display screens as seen in the video footage) all of which demonstrated that the train arrived later than the industry data says, and this affected whether or to what extent delay repay was due, then I'm not at all sure it would be sensible for the TOC to do anything other than pay out on appeal.

 

[Watershed]

Sounds better than the policy that airlines seem to use, wheels down equals arrival time.

I am unsure of which airline you have in mind here, but wheels down is absolutely not the arrival time for EU/UK261 purposes. Indeed I once had a flight where I was stuck onboard for 3 hours after arriving on stand!

The same goes for trains too. I suppose the OP could put in a GDPR subject access request for CCTV footage of themselves on the platform to Network Rail, as the station facilities manager at Reading, if they really wanted to pursue a claim. That would definitively prove the arrival time as being 15+ mins late. Admittedly perhaps overkill given the minimal compensation due for a 15-29 minute delay, unless the OP was travelling from further afield.

 

[Belperpete]

I am unsure of which airline you have in mind here, but wheels down is absolutely not the arrival time for EU/UK261 purposes. Indeed I once had a flight where I was stuck onboard for 3 hours after arriving on stand!

The problem is if the airline claims that the delay between touching down and opening doors was outside its control. So if the problem is down to the airline 's agents not being available to unload you, you are covered. However, when my airplane was held on the tarmac for an hour waiting for a gate, I had compensation denied as the airline stated it was outside its control.

 

[Paul Kelly]

In my experience Elizabeth Line trains usually crawl into Reading really slowly; it probably has an effect on the accuracy of the signalling berth offset calculation.

 

[lfc84]

Sounds better than the policy that airlines seem to use, wheels down equals arrival time.

Airlines use doors open

 

[Trackman]

It's not particularly clear what system operators use to calculate lateness, but I assume it's whatever works out to less compensation

I thought it was uniformed standard across the network.

 

[MrJeeves]

I thought it was uniformed standard across the network.

It might be, but that doesn't mean it's written anywhere publicly.

TOCs can override Darwin data, signallers can override TRUST data (can TOCs too?).

 

[Falcon1200]

signallers can override TRUST data (can TOCs too?).

Could you clarify please? Do you mean their own manual Trust reports, because altering automatic Trust reports is not something normally done (and staff would get into trouble for doing that to hide delays).

 

[MrJeeves]

Do you mean their own manual Trust reports, because altering automatic Trust reports is not something normally done (and staff would get into trouble for doing that to hide delays).

I mean both. Obviously noone is going to do it to hide a delay (you'd hope!) but my point was that it's not documented anywhere which timings are used for delay repay calculations.

There have been examples where overriding either TRUST or Darwin would fix incorrect "delay" information. One such example was where an LNER train (I think?) was shunted out of its origin and back in, but this was recorded in both systems as being something similar to a 20 minute early departure instead of departing on time which it ended up doing in reality.

 

[Haywain]

One such example was where an LNER train (I think?) was shunted out of its origin and back in, but this was recorded in both systems as being something similar to a 20 minute early departure instead of departing on time which it ended up doing in reality.

Such an example isn't a case of altering the times recorded, it's that the system is incapable of recording more than one time. To record them separately a separate schedule would have to be created for the shunt, and that may not have been possible at extremely short notice.

 

[MrJeeves]

Such an example isn't a case of altering the times recorded, it's that the system is incapable of recording more than one time. To record them separately a separate schedule would have to be created for the shunt, and that may not have been possible at extremely short notice.

My point was more that no altering was done in this case. It even resulted in (at least) one person choosing not to travel as they thought the train had been cancelled or removed from the timetable as it had disappeared from every web/app live departure board.

Back to the point at hand... Darwin? TRUST? Or is there no standard?

 

[Haywain]

Back to the point at hand... Darwin? TRUST? Or is there no standard?

I can't say with certainty, but I would expect it to be TRUST.