Should we trust Twitter sweepstakes and public polls when it comes to the Conservative Party leadership contest?

Results of polling on questions such as: ‘how long will Boris Johnson last as prime minister?’ may not be much use for predicting the future, but they tell us quite a lot about how low expectations are

Boris Johnson has said he would take the UK out of the EU on 31 October ‘do or die’ (EP (EPA)

So much of political commentary is trying to predict the future. That is one reason I am interested in opinion polls and betting markets. They are different things: pollsters try to capture the state of public opinion at the time, whereas bookmakers put together the predictions of large numbers of people about what is going to happen.

They overlap, because polls are often used to predict how people might vote in future – and they often ask people how they might behave if different things happen. And the evidence is that polls are better at predicting votes than betting markets. They were, for example, a better guide to the 2016 EU referendum, saying it was, in effect, 50-50, whereas the bookies gave Remain an 80 per cent chance of winning.

That is why I pay more attention to surveys of Conservative Party members that find 74 per cent of them intend(ed) to vote for Boris Johnson than I do to the bookies, who now give him a 96 per cent chance of becoming leader.