Delay Repay system selecting journey from different London station.

[TUC]

Earlier this week I made a Delay Repay application with respect to a journey from London Euston to Halifax, changing at Manchester Piccadilly and Victoria. As with several Delay Repay sites, the Avanti system asks you to select the nearest quarter hour for journeys and then makes suggestions. I entered 1700 and one of the options suggested was for 1713, so I selected this.

It was only on the final page when it showed a breakdown of the journey I realised that, although the site had offered a 1713 option, this actually was a journey which had that as the start time to walk from Euston to Kings Cross, and it actually related to the 1733 from Kings Cross, changing at Leeds.

I was left wondering why, even though I had selected Euston at the start, the system's assumption was to default to a walking fixed link and a Kings Cross journey rather than one from Euston itself? I might have understood had the Kings Cross journey arrived earlier, but it arrives in Halifax at 2044 whereas the Euston option arrives at 2035.

Can anyone explain why the system would take this approach?

 

[30907]

Earlier this week I made a Delay Repay application with respect to a journey from London Euston to Halifax, changing at Manchester Piccadilly and Victoria. As with several Delay Repay sites, the Avanti system asks you to select the nearest quarter hour for journeys and then makes suggestions. I entered 1700 and one of the options suggested was for 1713, so I selected this.

It was only on the final page when it showed a breakdown of the journey I realised that, although the site had offered a 1713 option, this actually was a journey which had that as the start time to walk from Euston to Kings Cross, and it actually related to the 1733 from Kings Cross, changing at Leeds.

I was left wondering why, even though I had selected Euston at the start, the system's assumption was to default to a walking fixed link and a Kings Cross journey rather than one from Euston itself? I might have understood had the Kings Cross journey arrived earlier, but it arrives in Halifax at 2044 whereas the Euston option arrives at 2035.

Can anyone explain why the system would take this approach?

Because it was the first available option (these walking route ones come up frequently in planners!)?

Of course if you had selected it the problem would coincidentally been LNER's not theirs

 

[TUC]

Because it was the first available option (these walking route ones come up frequently in planners!)?

Of course if you had selected it the problem would coincidentally been LNER's not theirs

If you mean it was the first journey to leave, it wasn’t. My own service was also due to depart at 1713 and, faced with two options at the same time (including the walking link time), wouldn’t the logical approach be to prefer the one where the train actually departs from that station?

Remember, up to this point, I hadn’t indicated that my journey was at 1713. It was the system which had come up with that option.

 

[Watershed]

Journey planning systems (which Delay Repay claim sites obviously have to incorporate) generally surface both the journey results with the fewest changes, as well as those which are quickest.

Is it possible the journey you accidentally selected was from Euston to Halifax with a single change at Kings Cross? That might have been slower than changing at Piccadilly and Victoria in certain cases, but it would have been surfaced due to having fewer changes.

I can see where you're coming from with your suggestion that offered journey(s) should always prioritise taking a train as the first leg, but it doesn't really make sense when you look at other situations.

For example, imagine you'd entered London Blackfriars to Halifax as your journey. Surely you'd agree that it wouldn't make sense for the journey planner to prioritise a route which involved taking a train to London Bridge before taking the Underground from there to Euston or Kings Cross?

It may sound a silly example but how can software distinguish between that and your example? Not really, is the answer.

Of course it's possible there's an entirely different reason for what happened, apart from my suggestions above. But I just wanted to walk you through how a piece of journey planning software would approach the issue.

 

[TUC]

Journey planning systems (which Delay Repay claim sites obviously have to incorporate) generally surface both the journey results with the fewest changes, as well as those which are quickest.

Is it possible the journey you accidentally selected was from Euston to Halifax with a single change at Kings Cross? That might have been slower than changing at Piccadilly and Victoria in certain cases, but it would have been surfaced due to having fewer changes.

I can see where you're coming from with your suggestion that offered journey(s) should always prioritise taking a train as the first leg, but it doesn't really make sense when you look at other situations.

For example, imagine you'd entered London Blackfriars to Halifax as your journey. Surely you'd agree that it wouldn't make sense for the journey planner to prioritise a route which involved taking a train to London Bridge before taking the Underground from there to Euston or Kings Cross?

It may sound a silly example but how can software distinguish between that and your example? Not really, is the answer.

Of course it's possible there's an entirely different reason for what happened, apart from my suggestions above. But I just wanted to walk you through how a piece of journey planning software would approach the issue.

I think what is different to the examples you give is that yours involve a cross-London journey before catching a train. I would have understood if it had offered 1713 via Kungs Cross as one option and 1713 from Euston to Manchester as another option. Instead, it didn’t even acknowledge the latter at all.

Remember that most people will not be aware that they are accessing a multi-TOC Delay Repay system which has been labelled as Avanti for its website’s purposes. As far as they are concerned, they travelled from Euston on an Avanti service, and other assumption is that Avanti journeys will be reflected in the options offered to them.

 

[yorkie]

I think what is different to the examples you give is that yours involve a cross-London journey before catching a train...

Your issue also started with a cross London transfer.

Remember that most people will not be aware that they are accessing a multi-TOC Delay Repay system which has been labelled as Avanti for its website’s purposes. As far as they are concerned, they travelled from Euston on an Avanti service, ...

Yes but it has to be multi-TOC by definition as it includes journeys which can go beyond th Avanti network at each end.

and other assumption is that Avanti journeys will be reflected in the options offered to them.

Journey planners don't have a filter to allow any TOC but with the caveat that itineraries must include a particular TOC.

You can stipulate specific TOCs only, or exclude TOCs, but you can't say 'any TOC but hide journeys that don't include at least one leg on Avanti'

 

[Watershed]

I think what is different to the examples you give is that yours involve a cross-London journey before catching a train

There's a 'Tube' link from Euston to Kings Cross in the industry data feed during most of the day and a 'transfer' (make your own way) link overnight. So it's a Tube leg you would have seen in the journey rather than a walking leg. Therefore, in data terms it's a cross-London ticket in exactly the same way as the example I gave.

I would have understood if it had offered 1713 via Kungs Cross as one option and 1713 from Euston to Manchester as another option. Instead, it didn’t even acknowledge the latter at all.

The latter is unlikely to be surfaced as it's typically an overtaken journey with more changes. It generally wouldn't appear on any kind of journey planner (without entering a 'via' point) for the reasons I've described above.

Remember that most people will not be aware that they are accessing a multi-TOC Delay Repay system which has been labelled as Avanti for its website’s purposes.

True, but you were asking why this happened and I've explained the logic as it applies from a software point of view. As @yorkie alludes to, I'm not aware of any journey planning software that has the capability to filter by journeys which have to involve a certain TOC.

I's clearly not an impossible task but it certainly would have an associated development cost, possibly into the 5 figures; you might be surprised how expensive even minor software changes can get, especially in the rail industry where accreditation and regression testing is needed for most changes.

I'm not sure that any Avanti budget available for improving their Delay Repay claim systems would best be spent this way. Allowing people to enter a 'via' point, which is a much simpler change as basically all journey planners support that functionality already, would achieve much the same result at lower cost.

As far as they are concerned, they travelled from Euston on an Avanti service, and other assumption is that Avanti journeys will be reflected in the options offered to them.

Again I can see where you're coming from, but there is absolutely no guarantee they will be travelling out of Euston on Avanti just because they're claiming Delay Repay on Avanti's portal. They could be making a journey along a less common route like yours, that happens to involve Avanti but only further north.

Take Euston to Macclesfield during engineering works that close the southern WCML; a bias towards Avanti might see itineraries involving their replacement buses from Bedford to Milton Keynes being prioritised over journeys involving trains via St Pancras, Derby and Stoke. I'm sure you'd agree that could be counterproductive.

They could also be claiming from the wrong TOC, so prioritising Avanti journeys might result in people selecting the wrong itinerary and thus being underpaid Delay Repay.

 

[TUC]

Your issue also started with a cross London transfer.

Only an invented transfer by the Delay Repay system. In terms of my actual journey, it was Euston to Manchester, cross Manchester transfer, then Manchester to Halifax.

I'm not arguing here for excluding the other journey options, only that in a journey going North from a station whose services from several operators are oriented in that direction, that would be the expected route of most passengers.

Except for where someone has made a cross-London transfer as part of a longer journey (in which case the Delay Repay claim would likely not have started at Euston anyway) if someone had travelled on a tube journey from another London station to Euston, I think it unlikely that most people would call that other station their starting point. They would regard where their rail journey commenced as the starting point. In the same way, I don't think anyone who had walked from Euston to Kings Cross would put Euston on a Delay Repay form. They would put Kings Cross.

The reason why this matters is that someone could unwittingly select what appears to be their 1713 ex-Euston journey, but have the Delay Repay claim rejected because the 1733 from Kings Cross was not late.

== Doublepost prevention - post automatically merged: 13 Jul 2025 ==

The latter is unlikely to be surfaced as it's typically an overtaken journey with more changes. It generally wouldn't appear on any kind of journey planner (without entering a 'via' point) for the reasons I've described above.

The journey is offered on journey planners, with no 'via' point selected. See below.

 

[Watershed]

Only an invented transfer by the Delay Repay system

It's not invented; it's a Tube link that's in the industry data feeds. Should Delay Repay systems ignore established industry data feeds and make things up themselves? I'm sure you can imagine how that would go wrong.

In terms of my actual journey, it was Euston to Manchester, cross Manchester transfer, then Manchester to Halifax.

Yes - but how do you expect the journey planner to know that in the absence of a 'via' field (which I agree would be a sensible addition)? It's not psychic

I'm not arguing here for excluding the other journey options, only that in a journey going North from a station whose services from several operators are oriented in that direction, that would be the expected route of most passengers.

So again, let's come back to my example of starting at London Blackfriars. Most people would be heading south from there by train, so are you saying that Delay Repay systems should be prioritising itineraries involving those trains - even if they're slower than starting with a Tube link to Kings Cross?

Again, that may sound a silly question - but how do you expect the software to know what a 'sensible' person would do? What does 'sensible' even mean? I would suggest travelling via Kings Cross is the obvious option, so to depart from that default you'd have to insert biases into the routing algorithm - and that gets complex and expensive very quickly.

Except for where someone has made a cross-London transfer as part of a longer journey (in which case the Delay Repay claim would likely not have started at Euston anyway) if someone had travelled on a tube journey from another London station to Euston, I think it unlikely that most people would call that other station their starting point. They would regard where their rail journey commenced as the starting point. In the same way, I don't think anyone who had walked from Euston to Kings Cross would put Euston on a Delay Repay form. They would put Kings Cross.

You may think so, but who's to say that everyone else will think the same way? Notably, a lot of TOCs are in the habit of rejecting Delay Repay claims for "not including the whole journey".

People may well intend to include Underground legs from stations like Euston that also have National Rail services - so unless you're advocating a completely separate set of rules for Avanti's portal specifically at Euston (which would encounter the challenges I've outlined previously) the behaviour has to remain as it currently is.

I'm genuinely not trying to play devil's advocate here, I'm just trying to outline the reasons why something you may see as "simple" and "obvious" is actually really quite difficult to do when you look at it in more detail.

 

[TUC]

I recognise that. The puzzle though is why, given this journey arrives earlier than the Kings Cross option and is offered by journey planners (see above), why delay repay systems wouldn’t also identify it?

 

[Watershed]

I recognise that. The puzzle though is why, given this journey arrives earlier than the Kings Cross option and is offered by journey planners (see above), why delay repay systems wouldn’t also identify it?

A fair question. Perhaps there's more to this than I've suggested so far.

 

[Haywain]

A fair question. Perhaps there's more to this than I've suggested so far.

It might be a bit odd but the different journey planners don't always offer the same results. So, in this case, it appears that the Avanti JP is being compared with the (SilverRail) one used by LNER.

 

[yorkie]

I recognise that. The puzzle though is why, given this journey arrives earlier than the Kings Cross option and is offered by journey planners (see above), why delay repay systems wouldn’t also identify it?

This is a different question to what was originally described.

The fastest itineraries are shown here:

Planner - Results - FastJP

www.fastjp.com

The system should come up with the 1713 itinerary via Manchester; if it doesn't, that's a failing of that system.

However, it's absolutely legitimate to come up with itineraries via Kings Cross; as you can see from the above link, this can be quicker.

== Doublepost prevention - post automatically merged: 13 Jul 2025 ==

A good journey planner will have advanced options such as 'avoid tube ', which would give a different set of results, and exclude journeys via KGX:

Planner - Results - FastJP

www.fastjp.com

You could also exclude other types of fixed links, however depending on what you exclude, that may eliminate journeys across Manchester (unless they were by train; for a journey of this nature it would typically involve the use of TPE or alternatively deviating via Salford Crescent)

Another option journey planners may provide is 'no initial fixed links', which would force the timetabled journey to start at the stated origin.

 

[TUC]

Again, that may sound a silly question - but how do you expect the software to know what a 'sensible' person would do? What does 'sensible' even mean? I would suggest travelling via Kings Cross is the obvious option, so to depart from that default you'd have to insert biases into the routing algorithm - and that gets complex and expensive very quickly.

'Sensible' wasn't the best term. What I meant was how most passengers in reality would describe their journey. Of course it is a subjective judgement, but I genuinely think that few passengers who happened to be at Euston and had to walk to Kings Cross to catch their train would describe their journey as having started at Euston.

 

[yorkie]

'Sensible' wasn't the best term. What I meant was how most passengers in reality would describe their journey. Of course it is a subjective judgement, but I genuinely think that few passengers who happened to be at Euston and had to walk to Kings Cross to catch their train would describe their journey as having started at Euston.

If you exclude fixed links from the start of the journey, this would deny people the ability to claim on inclusive London Underground tickets.