How is the actual arrival time of a train defined (SWR)?

[Kaftan]

My train operator has reduced the amount of my Delay Repay on a few occasions. When I have looked into it, I have spotted a pattern: according to my accurate, internet-connected phone, the train comes to a stop at 15, 30, or 60 minutes after the published arrival time. When I check on recent train times, the arrival time is recorded as a minute or two earlier. This reduces the amount of compensation payable by 100%, 50% or 50% respectively. As passengers cannot alight from a moving train, I would like to know what the policy is. I've observed the same thing many times, but of course the delays mostly keep within the same time band.

 

[DarloRich]

It doesn't matter what you accurate, internet connected phone says. What matters is what the system clock says

is this a working timetable v public timetable thing?

 

[WelshBluebird]

Sounds more like a difference in what people consider "arrived". For the average passenger, surely the "arrived" time is the moment the train has stopped at the station and the doors have been opened? In my experience, railway systems use the signalling time which often is when a train passes a marker that it sometimes triggered as the train rolls in to the platform. Certainly sites like Real Time Trains have shown trains as "arrived" at a station when I am waiting at the platform and can see the train still hasn't quite made it in yet!

 

[swt_passenger]

The recorded arrival time is from the signalling detecting the train arriving in the track section that includes the platform, there was an explanation of the details in a thread a few weeks ago, I’ll try and find it.

Might have been this:
https://www.railforums.co.uk/thread...live-departures-web-page.166689/#post-3542825

It isn’t an issue specific to SWR, as far as I can see.

 

[kristiang85]

I did complain once to SWR's Twitter that the doors opened 15 mins late but they rejected my claim; it is based on arrival time, and anything up to 1s before the minute counts at the minute before (so a train arriving at 14m59s counts a 14m delayed on a claim).

I'm confused by SWR's arrival times too. e.g. the 0816-0903 BSK to WAT service, usually around 10-15 mins late. Now I know that 0903 is the timetabled arrival time, so if we arrive at 0917 (14m late) I don't get delay repay.

However, early on in the journey the guard tends to say the scheduled arrival is 0901, and on occasion on delays will also say, e.g. sorry the train is 15 minutes delayed, arriving at 0916 (thus you assume its eligible for DR, but gets rejected when I claim).

I thought it was a mistake at first on the guard's part, but its happening most days now when he states the scheduled arrival time is 0901, 2 mins before the actual timetabled arrival time of 0930? Is this an industry buffer they put in? (needless to say, not once has it arrived at 0901 or indeed 0903 since I started keeping track)

 

[swt_passenger]

The guard seems to be quoting the working timetable, I think he should be quoting public times for passengers.

 

[kristiang85]

The guard seems to be quoting the working timetable, I think he should be quoting public times for passengers.

May I ask why there's two different timetables? (sorry if that's a really dumb question! )

 

[swt_passenger]

May I ask why there's two different timetables? (sorry if that's a really dumb question! )

The working timetable allows for more accurate timings, fractions of minutes, and slight variations to what would otherwise be clockface timings all day. So for instance passengers may be given the same memorable time every hour even though it runs a couple of minutes later twice a day to fit in an odd freight service or whatever...

 

[tsr]

The guard seems to be quoting the working timetable, I think he should be quoting public times for passengers.

Much easier said than done.

 

[swt_passenger]

Much easier said than done.

Does their schedule only show the WTT then? If so not surprising.

 

[Wombat]

For the average passenger, surely the "arrived" time is the moment the train has stopped at the station and the doors have been opened?

That's what I thought, but I keep forgetting that the laws of common sense are frequently suspended when it comes to the railway.

 

[TEW]

Does their schedule only show the WTT then? If so not surprising.

Generally on SWR guards schedules only show the public timings, and not the working timings. Occasionally it is the other way round, but never both.

 

[Muzer]

I once had a funny feeling this was going to happen (we were travelling much slower than usual past the last signal into Basingstoke station) so I recorded a video of us arriving on my phone, with the station clock clearly visible. Sure enough on the system it showed as 14 minutes late, so they denied my claim. I then responded with the video and they accepted it

My advice to those who know they're arriving very close to the threshold would therefore be to record a video of the arrival with a station clock visible.

 

[some bloke]

Sure enough on the system it showed as 14 minutes late, so they denied my claim. I then responded with the video and they accepted it

If the discrepancy is a known issue, is denial fraud?

Also, I would have expected them to record actual arrival time - meaning the time when the typical person could step onto the platform - for punctuality figures.
If they do record it, why don't they use it for delay repay; if they don't, how do they produce the figures?

 

[swt_passenger]

If the discrepancy is a known issue, is denial fraud?

Also, I would have expected them to record actual arrival time - meaning the time when the typical person could step onto the platform - for punctuality figures.
If they do record it, why don't they use it for delay repay; if they don't, how do they produce the figures?

It has already been explained the timings are derived from the signalling system, not the train doors.

 

[WelshBluebird]

It has already been explained the timings are derived from the signalling system, not the train doors.

And from a passengers point of view, those timings are not accurate (as they can show a service as arrived when it isn't even in the platform yet).

So I'd also echo some blokes post, questioning why these numbers are the ones used when they are not at all fit for that purpose.

Of course, the answer is that there isn't any other timing data available. In which case, surely ToC's should just give passengers the benefit of the doubt when we are talking about 14 min v 15 min and the like.

 

[swt_passenger]

And from a passengers point of view, those timings are not accurate (as they can show a service as arrived when it isn't even in the platform yet).
So I'd also echo some blokes post, questioning why these numbers are the ones used when they are not at all fit for that purpose.

I agree. But how do you save a central recording of door timings?

 

[some bloke]

The recorded arrival time is from the signalling detecting the train arriving in the track section that includes the platform

It has already been explained the timings are derived from the signalling system, not the train doors.

The first quotation says "the recorded...time". What was not clear to me, as a non-expert, was whether those are the only recorded times - or whether there are others, including for other purposes.

Are companies reporting wrong punctuality figures as well as wrong timings for Delay Repay?

 

[Muzer]

Yes. The timings based on the signalling system are known to be imperfect but they're currently all that's realistically available. Theoretically you could get data downloaded from every train on door opening times but a) that would be an overnight job rather than near-instant as it is now, b) with current systems I imagine it would massively increase workload at depots etc., and c) I'm not sure all trains currently in service would even record door opening times.

The timings based on the signalling system nominally measure from wheel start to wheel stop (this makes a lot more sense when you consider the system was probably invented before central door locking was completely ubiquitous), but of course they don't have that data directly; they only know when a train moves from one block to another block (sometimes they also have track circuit occupancy data but I'm not sure whether or not they'd always use that), and sometimes data is even entered manually. In my particular case it wasn't a known issue at Basingstoke; I think the driver just had bad luck with the timing when the last signal cleared so the train was a lot slower coming into the platform than it would otherwise normally be (normally I tend to find arrival times at Basingstoke reasonably accurate). I noticed the train was going much slower than it normally does which is what prompted me to record the video in this case.

 

[kristiang85]

I agree. But how do you save a central recording of door timings?

Surely its not beyond the wit of tech to install a log in the train of when the driver/guard activities the opening of the doors?

 

[swt_passenger]

Surely its not beyond the wit of tech to install a log in the train of when the driver/guard activities the opening of the doors?

There may well be one on relatively modern stock, but the problem is making it available as a record usable by the delay people some days or weeks later.

 

[some bloke]

Yes. The timings based on the signalling system are known to be imperfect but they're currently all that's realistically available. ...In my particular case it wasn't a known issue at Basingstoke; I think the driver just had bad luck with the timing when the last signal cleared so the train was a lot slower coming into the platform than it would otherwise normally be

Thanks. The OP talks of a pattern of a minute or two - I don't know whether it's enough to affect punctuality figures significantly.

...why these numbers are the ones used when they are not at all fit for that purpose.

...surely ToC's should just give passengers the benefit of the doubt...

I agree. But how do you save a central recording of door timings?

Are door timings the only realistic option?

Could companies instead add on time - at least the minimum expected between the recorded time and door opening - if there are stations where the difference is significant?

 

[swt_passenger]

Could companies instead add on time - at least the minimum expected between the recorded time and door opening - if there are stations where the difference is significant?

Yes they could, it’s mentioned in post 10 of the thread I linked to earlier. But I have absolutely no idea how widespread its use is.

 

[bb21]

Thanks. The OP talks of a pattern of a minute or two - I don't know whether it's enough to affect punctuality figures significantly.

These figures are audited and recalibrated regularly.

Could companies instead add on time - at least the minimum expected between the recorded time and door opening - if there are stations where the difference is significant?

No. The measure needs to be a consistent one nationwide, and there is no facility to determine additional offset on a nationalwide basis.

Note that it is a Network Rail function. TOCs get no say in this matter.

In reality it affects a tiny number of journeys, so the immense additional costs in infrastructure upgrades required to change the system is never going to be a good way to spend limited resources.

 

[some bloke]

These figures are audited and recalibrated regularly.

Just to be clear - the punctuality figures are not based on offset timings (relative to door opening which is what passengers might expect) or if they are, there isn't a significant problem?

 

[bb21]

Arrival times are wheelstop times. This is the agreed metric and is the metric used across board agreed, between the DfT, Network Rail, and all relevant stakeholders, taking into account what can realistically be achieved on a consistent basis. All punctuality data therefore works on this basis. This has worked well since at least the beginning of privatisation so there is no reason to doubt there is a major issue. The most people will be affected will be a few seconds - hardly a big deal.

Unfortunately if you have an issue with that metric then you may have to consider lobbying the DfT and persuade them that upgrading the infrastructure to fix this "problem" is money well spent.

Until then, we can all moan about it. It won't achieve anything.

Personally I am not averse to the idea floated above, but everything has to be seen in the context of what can practically be achieved with existing infrastructure and whether a good business case can be made for changing the way things work. You are talking about public money at stake here, whether through investment in infrastructure or changing the metric so that future franchise bids being based on associated additional cost implications.

 

[yorkie]

May I ask why there's two different timetables? (sorry if that's a really dumb question! )

Various reasons but one of them appears to be to reduce Delay Repay claims.

They have various other tactics for that, too.

 

[LordCreed]

Generally on SWR guards schedules only show the public timings, and not the working timings. Occasionally it is the other way round, but never both.

It's definitely possible to show both on schedule cards, as we do it for our guards.

 

[kristiang85]

Various reasons but one of them appears to be to reduce Delay Repay claims.

They have various other tactics for that, too.

I suspected as much.

 

[bb21]

The GBTT/WTT differentials existed way before Delay Repay was anywhere near conception.

That said, some of the differentials around the country in the current timetable cannot have helped with public perception that the industry were up to no good.