We criticise often enough so perhaps applauding something sensible should also be done:

Train tickets in Britain could be priced like airline seats under a demand-based system being trialled by the government as part of a wider rail shake-up.

The transport secretary, Mark Harper, announced on Tuesday evening that fares on some long-distance trains run by LNER on the East Coast line will fluctuate according to availability.

He said the state-run train company would also move towards scrapping return tickets across its network, as a test for whether the idea could eventually be rolled out across the wider railway. The trial follows a successful pilot of selling only single-leg tickets on some longer intercity journeys such as London-Edinburgh since 2020.

Airlines face the same basic problem as trains. Near all of the costs (near for airlines, all for trains) are the fixed costs of running the routes. The marginal cost of the extra passenger is near (or is) nothing. Thus the game is to maximise total revenue from the seats. That does mean variable pricing. This is then compounded by the manner in which an empty seat on a train/plane 10 minutes before departure has some value or other, that same seat 15 minutes later is worth nothing.

So, given that airlines have indeed found the solution to this problem over the past 30 to 40 years why not the same solution on trains? After all, solving an economic problem is indeed solving an economic problem. This also being an interesting proof that economics - at the level of microeconomics at least - is indeed a science. Generalisable principles extracted from observation and experiment, that’s pretty scientific.

Which does leave us with just the one question really. Why has it taken 30 to 40 years for a solution already known to be applied to trains? Possibly because whatever it is we’re calling the Department for Transport these days has retained control of ticketing methods?