[/LIST]
You miss my point, the only reason Newbury is further from Southampton by rail, yet closer by car, is because the Newbury - Southampton direct line was closed, creating the dog-leg route via Reading West. GWR have chosen to price tickets from Newbury at a premium (presumably to avoid the risk of Reading passengers spotting a good value ticket), but this makes a Newbury-Southampton journey by rail uncompetitive compared with other travel modes.
I think the issue you bring up here is really whether mileage-based pricing should be on mileage as the crow files or mileage along the shortest rail route. Both possibilities have advantages and disadvantages. In the scheme I proposed, I used 'along the rail route' and from what you're saying that seems to be what GWR have chosen to do as well in that case. The disadvantage is - as your example demonstrates - it makes rail less competitive on price if the rail route is circuitous. However, I think that's still the least bad option. If you price by 'as the crow flies' distance then the system will be full of absurd anomalies (Think 'Newbury to Andover' or 'Alton to Winchester'), which I think would make it unworkable. Maybe you could get some compromise -something like, distance used for the fare calculation = (4/5) * rail distance + (1/5) * 'as the crow flies' distance, to give some nod to the problem caused by gaps in the network. (It'd take some effort to figure out how workable that would be in practice - whether you could do it without too many anomalies. But maybe that'd work).
Sadly, I fear you probably have to accept that gaps in the network mean rail just isn't a sensible option for many journeys. (Although Southampton to Newbury is would be more competitive if fares generally were lower - we can but hope Government support in and investment in the railways might eventually make that possible).
I don't think I missed the point about the Beeching cuts by the way. I didn't comment on it because I don't think the fact that there used to be a Newbury-Southampton line in the distant past has any bearing on what the fare should be today. You calculate the fare today based on the network as it is today, not on the network as it was 60 years ago.
[/LIST]
Several points:-
1. The Farnborough Stations Group was abolished several years ago.
Sorry, I should've been clearer. I wasn't trying to claim that group existed
now. I was saying that, in my mileage-scheme, they would be considered as a group. [Because the lines cross there and in principle interchange can be made by walking between them, even if in practice few would choose to do that]
[/LIST]
2. Most Blackwater to London tickets are priced factoring in the dog-leg nature of the journey. This means prices via Ascot or Guildford are competitive with nearby stations on other lines e.g. Camberley and Farnborough Main, however I doubt a walk between Farnborough North and Farnborough Main is assumed. Taking that route is often more expensive, especially so at peak travel times where the price premium is 63% or, in money terms, £18.90 (I hope no-one is dumb enough to buy such a ticket, but there may be a few!).
Interesting point there. My formula did feature a 'quality factor' allowing you to reduce the price if the line you're travelling on is low quality. I'd assumed it would be applied for lines that are poor quality in absolute terms (very slow, poor rolling stock, infrequent trains - think Middlesbrough to Whitby or Inveress to Wick). However, maybe there is a case for using it in places where lines are good in absolute terms, but poor compared to neighbouring lines - which I think is the case for Camberley-London.
[/LIST]
3. The difference between the mileages to Waterloo is 43/33 (not 36), a 30% difference. I wouldn't be too chuffed if Blackwater - London tickets were to attract a 30% price premium over Farnborough Main or Camberley fares.
My 36 came from adding your figures of 33 miles London to Farnborough and 3 miles Farnborough to Blackwater to get the notional shortest distance from Blackwater to London.