Misleading quotes for rail tickets on the internet

Yesterday after choir Tricia happened to remark on the fact that she'd been on the internet to find out the cost of going to Norwich at half term for a day with young Lizzie, for a shopping spree. On the internet the lowest price quoted was something like £30:00 (Cambridge to Norwich return) and they almost decided against going, or scrapping the train and going by car.

But in a fit of extravagance they decided to go anyway, and were pleasantly surprised to discover that the tickets cost £13:00 total for both of them (in fact I suspect that ticket included the price of the bus to the city centre in Norwich as well as the train journey, though they probably didn't know that).

I've just checked The Trainline.com to see what they quote for an imaginary day-return journey tomorrow on that route, and they think it will cost £17:40 for a "saver return" for one adult, £22:30 for a "standard open return", or two single tickets (which for some reason it advises "might be cheaper") at a minimum of £14:10 each way (i.e. £28:20 for the round trip).

This is absurd because the ticket you actually want to buy is an Anglia Day Ranger which will take you not just to Norwich and back but to Sheringham or Yarmouth or anywhere else you want to go in East Anglia, as many journeys as you like in a day, and it costs £10:00.

Today I had a similar experience. I asked The Trainline.com to quote for my day-return journey, Cambridge to Bury St Edmunds. It offered a "standard day return" at £10:80 or two singles at a minimum of £7:50 each (again advising in its stupid way that "two singles might be cheaper"). When I got to the station I was served without hesitation with a "Cheap Day Return" ticket costing £7:90. A "Cheap Day Return" (which is an off-peak day ticket) is, after all, what I would have expected to be looking for.

Now what exactly is the problem here? There seem to be several hypotheses that spring to mind.

In the case of the Norwich trip it seems that the computer is perhaps programmed to give only ticket prices that are to the specific destination, and in this case the cheapest ticket is (unusually) not the return ticket to Norwich but the day ranger ticket for the whole of East Anglia, which is extraordinarily good value.

In the case of the Bury St Edmunds trip it seems that the computer is unaware of the Cheap Day ticket. Why might this be? One option is that the key factor is that the ticket I bought was to "Bury St Edmunds BUS". The crucial thing is that "BUS". The ticket includes the bus fares in the town you're going to. It's not a ticket to Bury St Edmunds station, it's a ticket to the city and is valid on the buses. I've encountered this before (it's the same if you go from Norwich to Yarmouth: you have to buy a ticket to Yarmouth BUS, because that's cheaper than a ticket to Yarmouth.)

If you go to the station you speak to a human person behind the window, and he knows that a ticket with the bus included is cheaper, and that's what you will want. He doesn't even ask (because who would want to buy a ticket without the bus if it costs more?). If you ask the computer the computer quotes you the price without the bus, because you didn't ask it for the bus (and why should you? It doesn't offer you that option).

That's the difference between a machine and a human being. That's why it will never be possible to make artificial intelligence.

As for why the ticket without the bus costs more—that remains a mystery (but it is the privilege of humans, unlike machines, to decide to do things in an irrational way if they so choose).