ON election night, the BBC had a digital readout counting down to 10pm, when they would reveal the results of the exit poll commissioned jointly by the major broadcasters.

With about 20 seconds to go, one of my sons asked me for a prediction, and with no time to add any ifs and buts, I said: “Tory majority of 100”.

It’s a good job I hadn’t been to the bookmaker’s with that one.

I take more interest in politics than most people and I’ve been speaking to politicians almost daily for about 25 years, so people sometimes expect me to know something about these things. But like most observers, I didn’t see this result coming.

Incidentally, I thought the public would warm to Ed Miliband by the 2015 election. And I initially expected Tim Farron’s tactic of targeting Remain voters would gain traction, although it was clear weeks before polling day that it wasn’t working.

In my defence, I did think last year that the polls were too close for the Remain side to be confident about the EU referendum. And I could see Donald Trump becoming president, although that was probably more through some sort of weary fatalism rather than a reading of the evidence.

There’s a maxim I’ve heard quoted a lot in recent times: “Nobody knows anything.”

It’s normally attributed to the great screenwriter William Goldman, who used it in his 1982 memoir Adventures in the Screen Trade.

Goldman was talking about the inability of even highly-paid Hollywood experts to predict what would be a hit film. Back then, Chariots of Fire had recently made a lot of money, far outstripping expectations, and been named Best Picture.

Four years before that, 20th Century-Fox had been telling some cinemas that they wouldn’t be allowed to show its sure fire 1977 hit, The Other Side of Midnight, unless they agreed to book another film that the studio was much less confident about – Star Wars.

Goldman’s quote is applied to all kinds of subjects now. I wonder whether he might agree that it’s sometimes overused in areas that have nothing to do with the entertainment business.

It’s a good aphorism, and it’s true that informed commentators in public life have been proved to be fallible several times recently. But that doesn’t mean we should be hostile to all expert opinion. It would be stupid to write off man-made climate change as though it were just somebody’s pet hypothesis, for example, but some are doing just that.

Yet when it comes to matters of politics and public opinion, I’m starting to believe that the “nobody knows anything” theory has something going for it.

My guess at the general election result missed by a mile, because I had been conditioned by experience. Every time the opinion polls have been wrong in the past, it’s been because – as in 1992 and 2015 – they overestimated Labour support.

What’s more, the conventional wisdom says general elections are won from the centre, and for the past 20 years the public haven’t given a majority to anyone left of Tony Blair or right of John Major. This time, a further-left Labour Party gained ground, while the centre party was badly squeezed.

We shouldn’t overstate Jeremy Corbyn’s achievement. He has, after all, brought Labour back to roughly the position Gordon Brown left them in seven years ago. But compared with the predicted Labour meltdown, the party pulled off something astonishing.

The evidence of previous elections suggested that young people don’t vote in the numbers they say they will. Yet this time, it looks as though they did.

Experience taught us that election campaigns rarely change minds – that people stick with the judgement they have formed over several years. Yet this campaign seems to have turned things around. In May we saw real votes cast in local elections, and Labour did not do well. By June 8, things were different.

Traditionally, parties whose poll ratings look good for leadership and the economy are on course to win. But Theresa May fell dramatically short.

We’re heading into a political landscape unlike any we’ve seen in modern times, in which a minority government will battle daily to set the agenda.

How will it turn out?

Obviously, I have no idea.