Have ticket restrictions become unusable for the general public?

[rob.rjt]

When you look up a train ticket on National Rail (on desktop at least) you can view the restrictions on a ticket by clicking the ticket type before clicking "Buy Now". Other suppliers may vary - I tried a journey booking on GWR and no link to the restrictions came up.

These restrictions however do not seem to be helpful to the public - for a Super Off Peak return from Swindon to Wellingborough (as an example) the restriction code is CJ. The restrictions themselves are intelligible (up to a point) but then you get into the notes where additional restrictions are applied depending on the origin.

When the origin of the outward portion of the ticket is related to Westbury, Castle Cary, Pewsey or Swindon the return travel restrictions above also apply.

This seems to suggest that if the outward portion is not from those stations, the return restrictions do not apply (otherwise, why specify that they do?)

In addition:

• Journeys from London Kings Cross and Stevenage please use restriction 9D.
• Journeys from London Euston please use restriction 9I.
• If travel is not to/via the stations listed above or is via Peterborough, see restriction code 2V.
• Journeys from London Paddington please use restrictions 5R, YN, YP, YQ, YU, YW, YX & XC.
• Travel from Tring is allowed after 09:30.

For some places, there is one additional restriction code, for Paddington there are 8! This is without going into the unpublished restrictions for the ticket (https://www.brfares.com/!faredetail?orig=SWI&dest=WEL&rte=200&ldn=1&tkt=SSR) - in this case one not to depart Swindon before 1029. It may be the case that these unpublished restrictions relate to the additional restriction codes, but this is not clear.

This is further complicated for the traveller as one of the notes (in relation to Peterborough) applies the additional restriction if travel is to or via Peterborough, but Paddington restrictions only apply to that station. A traveller on this ticket could well expect that the additional 8 restrictions do not apply as they are traveling via not to Paddington.

Is this not a problem because:

• most people book their travel in advance so know what trains they will be on? (in which case what about changing the return trip - this ticket is for a month and I don't know anywhere that allows you to check a journey based on a ticket you already hold)
• the complicated unpublished restrictions are on the outward journey which is much more likely to be planned ahead of time?

I would have thought that there should be a single page for journey restrictions for any ticket, rather than having to look in 9 and work out if your journey is affected by any of them, but this may be because there are a maximum of 1296 restriction codes.

Is there a better way of doing things?

 

[Sonic1234]

The complexity of restrictions and the ticketing system, for most customers, is less of a problem now because of the move to apps and online purchasing. People rely on the journey planners to tell them what ticket is valid, and I expect a lot of customers treat their ticket like an Advance (selecting the train they plan to use in the journey planner), even if it isn't.

As you say, the issue arises when plans change. The system is so complex now most staff don't know how it works, so you either get people being denied travel when they have a valid ticket (this happens often when people have an "off peak" which is valid at traditionally peak times) or just get waved through. Or tickets don't get checked.

 

[Tetchytyke]

People rely on the journey planners to tell them what ticket is valid, and I expect a lot of customers treat their ticket like an Advance (selecting the train they plan to use in the journey planner), even if it isn't.

The disputes forum is full of people who get caught out by this, selecting a train then getting PF’d or worse when they decide to travel earlier or later when different restrictions apply. People with railcards in particular.

The whole system is a mess and you’ll be threatened with prosecution if you fail to understand the nine pages of conflicting restrictions.

 

[Snow1964]

I think the question is wrong, regarding if too complicated for public.

If some staff get it wrong occasionally, or can't understand it, then its too complicated, Full stop.

The most common problem, is if things get changed (cancellations, missed connections, even customers plans) then there is no effective communication of what alternatives apply, do need upgrade, replacement, or can travel anyway getting a next best train. Far too much inconsistency on how this is applied because the staff can't work it out either.

 

[Sonic1234]

The most common problem, is if things get changed (cancellations, missed connections, even customers plans) then there is no effective communication of what alternatives apply, do need upgrade, replacement, or can travel anyway getting a next best train. Far too much inconsistency on how this is applied because the staff can't work it out either.

Yes, exactly. Using ticket acceptance is always a nervous experience, because you never know the updates given to and the understanding of staff on the ground.

 

[yorkie]

In the past, before people bought online, restrictions were often relatively simple and people generally relied on the human readable text.

But over time, online sales have massively increased and now the assumption now is that people buy online and rely on a computer to work it out (or, if you ask at a ticket office, the ticket clerk would similarly be using a computer, likely in journey planning mode)

Hence these days a reliance by all parties on the data, with the assumption that people will simply look up what's valid, which is pretty much the same way that air fares work.

Unfortunately the DfT are very keen for prices to increase, not decrease and therefore if any simplifications are ever deemed to be required, then while there might be a few winners, overall fares are likely to increase (this has happened on various occasions; see LNER's single leg pricing!)

We often get people confused because the journey planners won't do something the text suggests they should; those of us who understand the data can then see just how different the electronic data is compared to the human readable text.

I don't think the system necessarily needs to fundamentally change but I would like to see some straightforward principles applied, to protect customers; however the chances of that happening are virtually zero!

I fear that what could happen instead is once things get to breaking point, we end up with just one or two restrictions which are "simple" but overly restrictive, thus increasing fares by stealth (if that does happen it would often be possible to circumvent these by "splitting")

Some of the 'complexity' is due to 'easements' but without them it means higher fares and/or more need to 'split' to get the brat deal.

Ticket acceptance needs to be much more automatic than it is now; I've known situations where staff refuse to guarantee a ticket will be accepted in the event of disruption and you are advised to approach staff at each stage and beg/hope for acceptance, but I would say that's a different topic to restriction codes.

A small quick win would be for the Penalty Fare Guidelines to make it clear that PFs do not apply to being on the wrong route or travelling at an invalid time; this is merely reiterating the current position and not a material change as the NRCoT makes it clear that an excess applies. But without this clarification some poorly trained revenue staff attempt to change PFs incorrectly, and some customers wrongly fear that they could be subject to a PF if they travel at the wrong time. Of course I'd like to see far more reforms in this area but that's covered by a separate thread.

We can't trust the DfT to do any reforms as any changes will, overall , be negative towards customers.

Yes, exactly. Using ticket acceptance is always a nervous experience, because you never know the updates given to and the understanding of staff on the ground.

Indeed but I suggest a new thread is created for this; I'd have a lot to say but will hold off until someone does this. If no-one does, I'll try to do it this evening if I get round to it (and if I do, I will link to it from here)

 

[rob.rjt]

In the past, before people bought online, restrictions were often relatively simple and people generally relied on the human readable text.

But over time, online sales have massively increased and now the assumption now is that people buy online and rely on a computer to work it out (or, if you ask at a ticket office, the ticket clerk would similarly be using a computer, likely in journey planning mode)

Hence these days a reliance by all parties on the data, with the assumption that people will simply look up what's valid, which is pretty much the same way that air fares work.

In most cases, an air fare is point to point - you get a few odd things like US air fares where you can attempt to save money by getting off at a change point and not getting on the connecting flight to the official final destination. The rules around re-routing in disruption are generally better understood.

It is probably a good thing that restrictions can be made more specific (rather than a general "off peak valid after 0930" on all services), but, given the prevalence of e-tickets, is there any reason that the precise restrictions applying to a ticket could not be linked from the ticket in a human readable format? You could show the valid routes and precisely list any restrictions that could apply to that ticket? E-tickets also have space to state the existing human readable restriction text for any restrictions that apply to the ticket.

This would reduce the risk of arguments on trains if the ticket shows a map with the routes, and can also help with another of my bugbears - the difference on an advance routed "and connections" between trains you must travel on and trains that are suggested in the itinerary, especially on routes with regular services.

 

[py_megapixel]

What does "When the origin of the outward portion of the ticket is related to Westbury, Castle Cary, Pewsey or Swindon" even mean? How does one determine whether their station is "related to Castle Cary"??

I assume it means "If the origin of the outward portion of the ticket is Westbury, Castle Cary, Pewsey or Swindon", but then why not just say that?

 

[Snow1964]

In the past, before people bought online, restrictions were often relatively simple and people generally relied on the human readable text.

We often get people confused because the journey planners won't do something the text suggests they should; those of us who understand the data can then see just how different the electronic data is compared to the human readable text.

That difference between data and text is more of a poorly administered admin problem.

When a restriction is invented, should not write text that doesn't match what the data says. Should always convert the definitive data into text, not try and vaguely bodge text into a dataset.

The text and data should always match for a restriction, and if they don't then clearly text should be corrected.

If the text becomes unreadable and full of exceptions and cross references to other restrictions then clearly the restriction isn't fit for purpose.

 

[AngusH]

Yes, these restrictions have become far too complicated.

I agree that each ticket should be offering specific restrictions.



One obvious plan would be to forbid any incorporating by reference in the public text
In short, no restriction may reference any other restriction.

This doesn't prevent the railway company from having this in the planning stage,
merely that the final restriction doesn't use it. Any incorporated sections being
combined together automatically before the final publishing.


so:

working document:

"Includes 9x,9y,9z"

9x: via Birmingham only
9y: no luggage allowed
9z: via bath only

in the final public text becomes:

"Restrictions (via Birmingham only) and (no luggage allowed) and (via Bristol only)"


The other idea, agreeing with the above posters completely, being that all restrictions must be specific to the journey, none of these conditional statements that are used in different tickets.
Again it can be used by the railway company internally and then be combined together automatically at the publishing stage.


This type of combining and conditional publishing is trivial in most structured document systems.



I've also vaguely wondered if restriction codes should be chargeable to the railway companies.
Each code might cost them £10,000 per year, or 100,000 or whatever.

It might reduce the number of different codes, but would have to be combined with the rules above or they'd have very complicated ones.


edit: clarify on conditionals for entirely different tickets (forbidden) vs options for different possible uses of the same ticket (allowed)

 

[CyrusWuff]

What does "When the origin of the outward portion of the ticket is related to Westbury, Castle Cary, Pewsey or Swindon" even mean? How does one determine whether their station is "related to Castle Cary"??

To avoid having to set individual fares for every possible combination of stations, some stations are "related to" another station for pricing purposes.

In the old paper fares manuals this was indicated by saying "Unless specific fares are shown use prices to/from <Station>" against the relevant station.

On BRFares, it's indicated by the use of Clusters.

For example, fares between Westbury and Brighton are priced between a Westbury cluster (which also includes Avoncliff, Bradford on Avon, Dilton Marsh, Freshford, Frome, Trowbridge and Warminster) and a Brighton cluster (which also includes Aldrington, Falmer, Hove, London Road Brighton, Moulsecoomb and Preston Park.)

So in that example, Avoncliff would be said to be related to Westbury, and Hove would be related to Brighton.

 

[AngusH]

In these cases there should be a split between the railway working document which can use these groupings for convenience and the public information which should expand out all of the specific details and skip anything not needed.

The whole thing should be done by a computer somewhere.

 

[nickswift99]

To avoid having to set individual fares for every possible combination of stations, some stations are "related to" another station for pricing purposes.

In the old paper fares manuals this was indicated by saying "Unless specific fares are shown use prices to/from <Station>" against the relevant station.

On BRFares, it's indicated by the use of Clusters.

For example, fares between Westbury and Brighton are priced between a Westbury cluster (which also includes Avoncliff, Bradford on Avon, Dilton Marsh, Freshford, Frome, Trowbridge and Warminster) and a Brighton cluster (which also includes Aldrington, Falmer, Hove, London Road Brighton, Moulsecoomb and Preston Park.)

So in that example, Avoncliff would be said to be related to Westbury, and Hove would be related to Brighton.

That's great, but how is a member of the public supposed to know this using the information in front of them from authoritative sources? (BRFares is great but it is not authoritative).

 

[yorksrob]

Just trying to find out what constitutes off-peak, where can be extremely difficult.

 

[CyrusWuff]

That's great, but how is a member of the public supposed to know this using the information in front of them from authoritative sources? (BRFares is great but it is not authoritative).

Given the demise of the paper fares manuals some time ago, BRFares is the only option for most people. Even retail staff wouldn't be able to identify fares priced by cluster without looking there or by asking their TOC's Pricing Team, though they could infer it from the fares being the same for different station pairs.

Yes, these restrictions have become far too complicated.


I've also vaguely wondered if restriction codes should be chargeable to the railway companies.
Each code might cost them £10,000 per year, or 100,000 or whatever.

It might reduce the number of different codes, but would have to be combined with the rules above or they'd have very complicated ones.

I've just taken a look at the March 2023 fares data (latest version available from RSP without needing a licence), and that contains 728 restriction codes, roughly a third of which appear to be set by GWR. Quite why they require so many codes I have no idea!

 

[Dai Corner]

The silver lining of all this complexity is that those in the know, such as readers of this forum, can often pay less than passengers who buy the ticket offered by the TOCs. Some have even been able to monetise their knowledge by selling tickets and taking a cut of the savings.

I wonder how many of the restrictions have been introduced to try to stop this happening and, arguably, be fairer as well as maximise revenue for the TOC/Treasury?

 

[crablab]

Some have even been able to monetise their knowledge by selling tickets and taking a cut of the savings.

There have been other debates about the transferability of tickets, but reselling them in that manner would invalidate them.

I wonder how many of the restrictions have been introduced to try to stop this happening and, arguably, be fairer as well as maximise revenue for the TOC/Treasury?

My guess: Zero. People using interesting routeings or 'interpretations' of the restriction codes make up a rounding error in the fares income.

In any case, if they really wanted to they'd address that by making everyone buy an Advance where there is no ambiguity over validity or routeing.

I imagine they'd be more interested in revenue 'lost' through split tickets, which are more mainstream now Trainline is doing them.

 

[Haywain]

There have been other debates about the transferability of tickets, but reselling them in that manner would invalidate them.

I think it is a reference to Trainsplit, and other legitimate retailers.

 

[mangyiscute]

Ultimately if you want to simplify restrictions, I think either they would drop peak time and make all fares be like 30% higher than the previous off peak amount (a massive fare rise that they'll try to justify by loss of peak revenue) or they'd just enforce a blanket "off peak means nothing before 09:30 and not leaving london on anything from 4-7, super off peak means nothing before 11am", which would also cause more restriction on a lot of current tickets.

 

[fandroid]

The OP mentions the impossibility of finding out the validity of a potential altered return journey. Some idea can be gained by putting in a dummy journey which occurs on the appropriate days of the week - that is the same day for the original outward portion and the revised day for the return portion. Then see if the same type of ticket was offered. All a bit hit and miss but you'd be unlucky to straddle a date when the restrictions were changed generally.

 

[rob.rjt]

Ultimately if you want to simplify restrictions, I think either they would drop peak time and make all fares be like 30% higher than the previous off peak amount (a massive fare rise that they'll try to justify by loss of peak revenue) or they'd just enforce a blanket "off peak means nothing before 09:30 and not leaving london on anything from 4-7, super off peak means nothing before 11am", which would also cause more restriction on a lot of current tickets.

I and other people are not necessarily asking for this, more that all restrictions are easily visible and intelligible to the customer. For the example journey I picked (which was just one I found that used the particular restriction) - I should see the restrictions on the routes available between Swindon and Wellingborough, along with words such as to, from and via having their normal meaning.

The OP mentions the impossibility of finding out the validity of a potential altered return journey. Some idea can be gained by putting in a dummy journey which occurs on the appropriate days of the week - that is the same day for the original outward portion and the revised day for the return portion. Then see if the same type of ticket was offered. All a bit hit and miss but you'd be unlucky to straddle a date when the restrictions were changed generally.

After posting I also considered the possibility of delay on the outward journey. One would hope that staff would show discretion, but if, as a passenger, you are off route, or on the wrong train, you end up paying.

 

[AngusH]

I think one useful tool (which I'm not aware exists) would be a sort of verifiable reasoning tool for ticket rules.
You'd type in the ticket details and the train and it would tell you if the combination was valid or invalid and explain why.

e.g

customer has ticket X can they use it on the 0915.

Which would then show its reasoning by applying the validity rules:

ticket X is not valid on the 0915 because
- it is an off-peak ticket that is only valid for departures from this station after 0930 and the 0915 departs earlier.


or different case:

ticket Y, a ticket from Edinburgh to London is not valid on the 1122 because:
- the 1122 is a LNER train and the ticket is only valid on Lumo trains.


It could also be deployed to barrier staff so they could reliably test any ticket and train combination and then show a definitive proof to the customer who could also run the test themselves on the same system and get the same result.

 

[HurdyGurdy]

I think one useful tool (which I'm not aware exists) would be a sort of verifiable reasoning tool for ticket rules.

If there was such a tool, presumably it would be based on the unpublished restriction data used by booking engines. Currently that data does not reliably align with the human readable text versions of the restrictions.

 

[AngusH]

If there was such a tool, presumably it would be based on the unpublished restriction data used by booking engines. Currently that data does not reliably align with the human readable text versions of the restrictions.

I guess it would have to, unless the human readable rules could be parsed into machine processable rules and used instead.

 

[crablab]

I think one useful tool (which I'm not aware exists) would be a sort of verifiable reasoning tool for ticket rules.

I thought Trainline did that? You could check which services the ticket you'd bought was valid on, against the actual running data.

Currently that data does not reliably align with the human readable text versions of the restrictions.

Indeed. It's unclear which takes precedence.

I wouldn't usually bother doing the 'reverse fare search' to check whether my ticket was valid, unless it was to validate a complex routeing. I'd just check the written restriction and rely on that.

I think it is a reference to Trainsplit, and other legitimate retailers.

Ah, yes I see. I read it as being individuals rather than accredited retailers.

 

[Dai Corner]

I think it is a reference to Trainsplit, and other legitimate retailers.

It is, but having said that

Ah, yes I see. I read it as being individuals rather than accredited retailers.

If a friend saves you a significant amount on a train fare it wouldn't be unreasonable to buy them a coffee, beer or whatever as a thank you.

 

[Haywain]

If a friend saves you a significant amount on a train fare it wouldn't be unreasonable to buy them a coffee, beer or whatever as a thank you.

Maybe, but that's not really "taking a cut of the savings". More a voluntary booking fee.

 

[HurdyGurdy]

I thought Trainline did that?

Yes. But like other booking engines Trainline observes the unpublished restrictions data. Which means that it will often offer an Anytime ticket for an itinerary where the human readable text says an Off Peak ticket is valid. One example in this recent thread, but there are many, many others.

 

[etr221]

Yes. But like other booking engines Trainline observes the unpublished restrictions data. Which means that it will often offer an Anytime ticket for an itinerary where the human readable text says an Off Peak ticket is valid. One example in this recent thread, but there are many, many others.

So restrictions have reached the point of becoming unusable for the railway's system(s) ... so what chance does the public have?

 

[Adam Williams]

We were doomed as soon as there were separate machine-readable restrictions and human-readable text.

It was a stupid idea, and the two were bound to diverge.

NRE should be displaying text built from the electronic restrictions data. If the schema isn't flexible enough for TOCs to define the restrictions they want: tough. You don't get to have those restrictions in place without investing in improving the fares data model.

Never give your users a free-text field if you want any sort of data quality.