What Would a Good System for Working out Fares Look Like?

[najaB]

I have been having a place with some milages from an old timetable and have come to the conclusion that however you price rail fares you are equally likely to end up with anomalies.

Ladies and Gentlemen, we have a winner!

 

[DynamicSpirit]

I have been having a place with some milages from an old timetable and have come to the conclusion that however you price rail fares you are equally likely to end up with anomalies.

I agree. But I'm not sure if that's a bit of a red herring in the context of this discussion, as I don't think anyone is claiming that you could produce a system that has no anomalies anywhere. Rather, the question is, can you produce a system that is workable, and is (a) is fairer than, and (b) has fewer anomalies than, the current system. I and others are suggesting that a system built more closely on mileage (albeit with other factors built in) can do that. So far, I've not seen anyone provide a good reason in principle for doubting that (although I'd accept that you'd need to do more work and more modeling than any of us is probably able to do to prove it one way or the other)..

 

[sheff1]

I have been having a place with some milages from an old timetable and have come to the conclusion that however you price rail fares you are equally likely to end up with anomalies.

Not at all.

Some systems are much more likely to produce anomalies than others. The current system where different TOCs, with very different pricing models, price fares for the same sections of track will produce more anomalies than when one body sets the fares. The notoriously expensive Cross Country walk up fares which can be easily undercut by, for example, a string of Great Western/Northern/East Midlands walk up fares valid on exactly the same trains is a prime example.

 

[Bletchleyite]

Ladies and Gentlemen, we have a winner!

It's not true, though. A pure kilometric/mileage based (same thing) system based on rail miles will by definition not have any anomalies.

It might have undesirable features, e.g. long distance journeys being pricey, no peak/off-peak so peak overcrowding, some routes non-viable, some journeys that are short as the crow flies but long by rail expensive etc - but none of those are anomalies.

I challenge anyone to find an actual anomaly in such a system with prescriptive routeing.
--- old post above --- --- new post below ---

I agree. But I'm not sure if that's a bit of a red herring in the context of this discussion, as I don't think anyone is claiming that you could produce a system that has no anomalies anywhere.

You could, because provided you have prescriptive routeing a rail-mile based system will not have any actual anomalies at all (i.e. situations where splitting or buying long is cheaper). That's even true if it has a taper, or if it uses notional tariff kilometres.

 

[DynamicSpirit]

You could, because provided you have prescriptive routeing a rail-mile based system will not have any actual anomalies at all (i.e. situations where splitting or buying long is cheaper). That's even true if it has a taper, or if it uses notional tariff kilometres.

OK, I'll grant you that with one proviso: You may find that you also cannot have peak/off-peak distinctions, since I've not yet encountered any suggestion for how you could do that without anomalies. If you can do that, and have prescriptive routing, then there would be no anomalies (although I'm fairly sure the disadvantages of such a scheme - not least, the complexity that completely prescriptive routing would cause to passengers) would far outweigh the 'no anomalies' advantage). Perhaps I should rephrase my early comment to something like, 'no one is seriously proposing a scheme that has no anomalies at all as a sensible way of setting fares)

 

[Bletchleyite]

Agreed, there is, by my mind, no peak/off peak differentiator that avoids anomalies *other than* basing it on journey segments which would require all tickets to be Advances so the correct peak/off peak segment was selected for each train.

 

[DynamicSpirit]

Agreed, there is, by my mind, no peak/off peak differentiator that avoids anomalies *other than* basing it on journey segments which would require all tickets to be Advances so the correct peak/off peak segment was selected for each train.

Yep.

I think by the way there is another way you can get close to no-anomalies, but not all the way there:

1. On each route decide on which trains will count as peak trains, and how far out the peak restrictions will go. So to extend my example from earlier in the thread, you might - say - decide that on the West Coast Main Line, peak restrictions apply between Euston and Rugby on all trains scheduled to arrive at Euston between 6.30am and 9am Mon-Fri. So, on any of those trains, you need an anytime ticket to travel south of Rugby. The important thing here is that there is one fixed geographical area that applies to all peak trains.
2. Set peak fares on all journeys between Euston and Rugby to be some simple multiple of the corresponding off-peak fare. I used 1.5 earlier.
3. For journeys that are only partially in the peak zone - for example, Stafford to Milton Keynes, the 'anytime' far is calculated as 'off-peak fare' PLUS the peak adjustment for the portion of the journey that's in the peak zone. So if the peak multiplier is 1.5, then anytime Stafford-Milton Keynes fare = off-peak Stafford-Milton Keynes + 0.5 * (off-peak Rugby-Milton Keynes).

I don't think it's possible for that algorithm to produce anomalies for one peak zone: The calculation means that split-ticketing at the peak/off-peak boundary will never be advantageous, people travelling mainly outside the restricted area won't pay unfairly large fares (as is currently the case for many longer distance passengers), and it won't produce any anomalies of long journeys being cheaper than short journeys.

However, that reasoning breaks down if you have two peak zones. For example London-Birmingham: You probably want peak restrictions around Birmingham as well as around London. So for a London-Birmingham ticket, do you price the anytime markup to take account only of the London peak restrictions, or only of the Birmingham peak restrictions, or do you add the markups from the two sets of restrictions together? Whichever one you do, you'll end up creating anomalies. However, I'm fairly sure this will create far fewer anomalies than any other reasonable system (sorry - I don't consider requiring all tickets to be advances to be 'reasonable' and I suspect you agree and have only put that up as a theoretical example )

(Actually, you could avoid anomalies in this scheme by doing things like issuing separate London-peak and Birmingham-peak tickets depending in which zone you're travelling on peak trains, but I think it's obvious that would be too complicated for passengers to be workable).
--- old post above --- --- new post below ---

********************************************************************************************************************
******** New post (a few days later and unrelated). Sometimes I wish you could switch off the double post prevention system.... )
********************************************************************************************************************

I think the following post from this thread perfectly demonstrates the need for a radical change to the way fares are calculated. It explains the reasoning behing why going via Preston is OK if travelling Leyland-Chorley.

If you change at Leyland then it is valid under the shortest route rule. If you change at Preston then it is valid because:
Wigan North Western is a routeing point itself (technically a member of Wigan Group)
Preston passes the fares check as a destination routeing point for a Wigan to Chorley journey
Preston to Chorley is a valid local journey in itself (through train)
The Virgin train from Wigan to Preston follows a valid mapped route between those two routeing points.

Because the mapped route and the local journey are validated in separate steps, the doubleback is irrelevant.

I'm assuming the above is correct because noone has contradicted it, but ... how could any reasonable passenger possibly be expected to understand and figure that out for himself? Even I don't understand the reasoning or some of the terminology in the post - and I'm pretty sure I would be in the top 2-3% of passengers when it comes to knowing stuff about the railways. Would even most revenue staff have sufficient knowledge to answer this simple routing question without having to search through documentation? Something better is surely urgently needed when it comes to routing guidelines.

 

[najaB]

I'm assuming the above is correct because noone has contradicted it, but ... how could any reasonable passenger possibly be expected to understand and figure that out for himself?

Does the average passenger need to understand and figure that out for himself? The average passenger goes to a ticket office or website and buys the best ticket on offer.

As Neil Williams posited previously - as long as the ticket office or machine is able to work out what is best then it doesn't matter how complex the system behind it is.

 

[DynamicSpirit]

Does the average passenger need to understand and figure that out for himself? The average passenger goes to a ticket office or website and buys the best ticket on offer.

I would say in this case, quite possibly. If you have a ticket to somewhere for which multiple routes look reasonable, you do need to be able to tell fairly easily if the ticket is or isn't going to be valid by any of those routes.

 

[Bletchleyite]

I would say in this case, quite possibly. If you have a ticket to somewhere for which multiple routes look reasonable, you do need to be able to tell fairly easily if the ticket is or isn't going to be valid by any of those routes.

The best way to do that is to list them on the ticket. With barcodes, tickets could easily be A4 sheets from a laser printer (or otherwise larger) so have much more room for this, or you go DB style and base it on 3 character codes. Again makes no odds as to what esoteric method was used to work it out.

With Advances, all you need to know is the booked train. Again, no need to have a clue how the fare is worked out.

 

[PeterC]

Does the average passenger need to understand and figure that out for himself? The average passenger goes to a ticket office or website and buys the best ticket on offer.

Requires the passenger to be clear about their requirement and to be aware that day and/or time of travel can affect the price.

If you then require a complex dialogue to determine the best fare the traveller may simply decide that it is too complex and intrusive and decide to drive instead, not to mention delaying customers waiting behind.

 

[Bletchleyite]

I still think Trainline-style (or DB-style) TVMs are the way to go for this. Select a connection by price.

 

[MikeWh]

If I want to travel from London to Dartford I would not want to be tied to one of the four different routes. If I had to be I'd probably find another way of travelling.

 

[Haywain]

The best way to do that is to list them on the ticket. With barcodes, tickets could easily be A4 sheets from a laser printer (or otherwise larger) so have much more room for this, or you go DB style and base it on 3 character codes. Again makes no odds as to what esoteric method was used to work it out.

At least that would stop people complaining that Paper Roll Tickets are too big!

 

[najaB]

If I want to travel from London to Dartford I would not want to be tied to one of the four different routes. If I had to be I'd probably find another way of travelling.

Remember that, in this world, all tickets are Advances so you would be tied to a given train, not a given route.

 

[Bletchleyite]

If I want to travel from London to Dartford I would not want to be tied to one of the four different routes. If I had to be I'd probably find another way of travelling.

If a return was twice a single, what would be the problem with deciding your outward route and purchasing a single, then deciding your return route when you arrive at the station and purchasing another single? Or, if you changed your mind mid-journey, exchanging your ticket for one for the route you did take and are intending to take? (that could replace actual excesses)

With a system based in some way on mileage, you could have *any route you liked* at the appropriate cost, no need for the Routeing Guide, just buy a ticket for the route you happen to fancy. Bletchley to Milton Keynes Central via London, Edinburgh, Aberdeen, Inverness, Glasgow and Birmingham? Go ahead, if you like, here's the fare. Just like you can do things like that in Switzerland - in a couple of minutes at any TVM! (And I have priced out little "tours" like that in CH plenty of times before).

 

[MikeWh]

Remember that, in this world, all tickets are Advances so you would be tied to a given train, not a given route.

I've possibly jumped into a conversation mid-way through, but are we really saying that commuters would have to book advances for every journey?

If a return was twice a single, what would be the problem with deciding your outward route and purchasing a single, then deciding your return route when you arrive at the station and purchasing another single? Or, if you changed your mind mid-journey, exchanging your ticket for one for the route you did take and are intending to take? (that could replace actual excesses)

Well we're talking about commuter type routes. It often happens that trains arrive out of sequence so while I might intend to get the Sidcup line train to Dartford, the train which actually turns up first is the Blackheath and Woolwich line train. These are all DOO trains so no chance to swap on the train, and if I had to go back to the ticket office at London Bridge I'd probably end up needing a completely different ticket.

 

[najaB]

I've possibly jumped into a conversation mid-way through, but are we really saying that commuters would have to book advances for every journey?

Under Neil William's proposal, effectively yes. All tickets would be train-specific.

 

[sheff1]

I've possibly jumped into a conversation mid-way through, but are we really saying that commuters would have to book advances for every journey?

If such a proposal has been made I have missed it.

The problem is that this thread has got so complicated, with multiple discussions running in parallel, that a suggestion relevant to only one sub-discussion can be taken out of context and then yet another sub-discussion starts up about something that had never been proposed in the first place !

 

[Bletchleyite]

Under Neil William's proposal, effectively yes. All tickets would be train-specific.

Not necessarily. They (the flexible ones at least, and nobody in this thread is addressing off-peak because it is a very difficult problem) would however be route specific, in the manner that Swiss and German tickets are.
--- old post above --- --- new post below ---

Well we're talking about commuter type routes. It often happens that trains arrive out of sequence so while I might intend to get the Sidcup line train to Dartford, the train which actually turns up first is the Blackheath and Woolwich line train. These are all DOO trains so no chance to swap on the train, and if I had to go back to the ticket office at London Bridge I'd probably end up needing a completely different ticket.

This is within Greater London, I guess? Not sure I wouldn't just go for a zonal German-style joint tariff system managed entirely by TfL (and exactly the same as Tube fares) for those services.

Outside of the big cities there are far fewer similar cases.

 

[John @ home]

If I want to travel from London to Dartford I would not want to be tied to one of the four different routes.

I might intend to get the Sidcup line train to Dartford, the train which actually turns up first is the Blackheath and Woolwich line train. ... if I had to go back to the ticket office at London Bridge I'd probably end up needing a completely different ticket.

This is within Greater London, I guess?

Not quite. Of the places mentioned above, Dartford is in Kent and the others in Greater London. Until recently, Dartford was outside the Oyster zones. It has now been included in Oyster zone 8.

Not sure I wouldn't just go for a zonal German-style joint tariff system managed entirely by TfL (and exactly the same as Tube fares) for those services.

A zonal system with a single management is sensible in and around large cities, but giving that job entirely to TfL will cause howls of outrage from towns such as Broxbourne, Dartford, Redhill and Watford. It seems to me that the northern cities have handled this better, forming City Regions which extend beyond the boundaries of the old metropolitan counties to cover the whole travel-to-work area. It can lead to oddities like Barnsley being in both Leeds City Region and Sheffield City Region, but it does seem to encourage co-operation and avoid the democratic deficit which is such a problem on the edge of Greater London.

Outside of the big cities there are far fewer similar cases.

If devolution is to be genuine, it will lead to anomalies at boundaries. If the government's plan to devolve local transport to city regions is implemented in full, there will be a very large number of anomalies at boundaries. But will it give us better local rail services across the north of England? Despite the anomalies, I tend to think it will.

 

[Bletchleyite]

If devolution is to be genuine, it will lead to anomalies at boundaries. If the government's plan to devolve local transport to city regions is implemented in full, there will be a very large number of anomalies at boundaries. But will it give us better local rail services across the north of England? Despite the anomalies, I tend to think it will.

Would it be more likely that devolution would follow the German model, i.e. provision, operation and management devolved but a national fares model maintained? That model seems to have been very, very successful in Germany, with regional services being rather transformed from the very 1980s-BR look of a loco and a couple of rotting Silberlinge on a 2-hourly frequency to modern DMUs and EMUs much more often - and passenger growth to match.

 

[pne]

route specific, in the manner that Swiss and German tickets are.

And note that German routes are, at least as I understand it, not restricted to routes in the narrow sense of the word (simple A-B-C-D style linear routing) but can use "Raumbegrenzungen" (area restrictions?) which describe a kind of polygon à la "A-(B-C/X-Y)-D" which means "A to D via either B and C on the one hand, or X and Y on the other hand, or any route in between".

This could, if desired, provide for routes of the kind "London-(Woolwich/Sidcup)-Dartford" or "York-(Leeds/Doncaster)-Sheffield", the former also being valid via Bexleyheath and the latter also being valid via Pontefract Baghill.

I don't know how the price is calculated for such routings, however (on the longest possible route, some kind of average, or what).

 

[najaB]

And note that German routes are, at least as I understand it, not restricted to routes in the narrow sense of the word (simple A-B-C-D style linear routing) but can use "Raumbegrenzungen" (area restrictions?) which describe a kind of polygon à la "A-(B-C/X-Y)-D" which means "A to D via either B and C on the one hand, or X and Y on the other hand, or any route in between".

So we start getting back into permitted routes - which is a large part of what people find too complex about the current system.

 

[sheff1]

So we start getting back into permitted routes - which is a large part of what people find too complex about the current system.

A major problem with the current GB system is finding out what the permitted routes are and the lack of logic with some of them. Some permitted routes are frankly bizarre and would only be identified by the type of person who is on this forum, whereas other routes are not allowed, yet to the man in the street seem perfectly reasonable (e.g. the shortest journey time between origin and destination).

"Any permitted" on a ticket tells people very little - some people believe it means they can go 'any' way - York to Sheffield via Manchester, say !

With the system described, the routing is clearly printed on the ticket (including options, if appropriate). It helped by the fact that, in Germany, codes are widely used, outside of the railway, in lieu of the full name of towns and cities. This is not the case in the UK, so any routings would need to spell out the name in full (or very close to).

 

[pne]

A major problem with the current GB system is finding out what the permitted routes are

Agreed; I think a specification that's printed on the ticket is a different beast from the vague "any reasonable route" or "Any Permitted".

 

[yorkie]

Agreed; I think a specification that's printed on the ticket is a different beast from the vague "any reasonable route" or "Any Permitted".

What would you print on a Poppleton to Shireoaks ticket?

 

[pne]

What would you print on a Poppleton to Shireoaks ticket?

One possibility might be "Poppleton - (Harrogate-Leeds-Wakefield-Sheffield / York-Doncaster-Retford) - Shireoaks", implying that anything between the polygon implied by H-L-W-S on the one hand and Y-D-R on the other hand would also be acceptable.

 

[najaB]

...implying that anything between ...

Routeing by implication... And I thought this was going to be a better system.

 

[Bletchleyite]

Routeing by implication... And I thought this was going to be a better system.

So add "Any route bounded by" and "without doubling back i.e. passing through a station twice", then. Not implicit then, actually quite explicit.

I'd imagine writing a piece of software to generate that kind of thing from something similar to the current maps would not be an impossible task.