Neil Williams wrote:
On Sat, 12 Feb 2005 13:28:20 +0000 (UTC), Stephen Osborn
wrote:

Every can of beans / newspaper / magazine / item of clothing / etc you
buy is individually priced and you cope with that don't you.

Yes, but there is enough room in your average shop for all these items
to be out on display. With a probable average of, say, 5 or 6 ticket
types per relation (which would be N(N-1) where N is the number of
stations on the national system, or on LUL, as applicable), that ain't
practical.

True, but I am not sure that this is fully relevant. I doubt that you
compare the price of every can of baked beans every time you go shopping
to see which is the best value, but you might do so now and then.
However if you were to buy some caviar then you probably would check out
the best value.

So, if you want to check out the prices (caviar / London to Edinburgh)
then you can do so. If you are happy with what you usually get (beans /
day return your-local-station to London) then you can do that.

The key thing is that the system needs to reliably gives the appropriate
ticket.

One basic point to bear in mind is that, in general, systems can be fair
or they can be simple.
A zonal system can be simpler but full of anomalies (e.g. four stops
crossing a zonal boundary costing more than 10 stops with a single zone)
and so less fair.
A point to point system can be fairer (charging for the distance
traveled) but will be more complex.

2. AFAIK, the reason, AFAIK, that fares structure takes 7 volumes or
whatever and it takes an age to buy a ticket is that BR had made
thousands of special terminal in the 1970s and these are what are still
being used by counter staff today. The memory capacity of these is very
limited indeed.

The complexity of the fares structure has nothing to do with the
machines which issue it, which as it happens are largely in the
process of being replaced with machines which do "know" the entire
fares structure.

I was not clear. The 7 volumes are only relevant in that station staff
have to look things up in a number of large paper books and often get
them wrong, because there are so many options /discounts / etc. It does
not really matter if there are 7 volumes or 17 volumes if the system
reliably gives the appropriate ticket.

A modern box (probably running Linux and with a cheap 80-120GB hard
drive) could easily cope with all of the data and spit out the cheapest
or quickest option in a fraction of a second. With a decent UI[*] that
is what the passenger accessible machines would have as well.

The cheapest/quickest *single* ticket, yes (where I mean one ticket,
not just a one-way). The number of possible fares *combinations* is
staggering, and because the fares system (if you'd call it that) is so
badly broken, it is necessary to investigate these for best value.

But single tickets (i.e. A to B and back either one day or seasonal)
versus travelcards is what this discussion is about.

If you regularly travel A to B to C to D to A or your journeys are
irregular (home to work to one of many clients to different one of
many clients to work to pub to home) then a travelcard is almost
guaranteed to be better for you.

If all you do is local train station to London to local train station
then a travelcard is almost guaranteed to be worse for you.

BTW, I would totally agree that the number of possible tickets is
unnecessarily wide. I went from London to Edinburgh last year and there
were well over 20 possible fares.

Neil