The end analytics in diagrams, evening of June 7th – as you can see, the Tories & Labour are very close according to the system and it pretty much got it.
So the UK election results are now known, Tories have 318 seats but needed 326 for a majority of 1. The Tory eventual vote share in 2017 was 43, Labour was 40%, a 7% difference and puts the Tories 8 seats below. In 2015 it was Tories 36.9%, Labour 30%, a 23% difference.
Our algorithms had it, predicting a hung Parliament, with the upper bound a slightly smaller Tory majority. The algorithms were saying the ratio between the 2 parties vote shares were 7% low, 15% median, 21% top (rounded). It is very risky to translate % difference into seats, given the way the UK voting system works but we calculated we were looking at 15% = just under zero majority, 7% was about 8 too low and 21% was about 5 seats over
So we got it, right – ok, it was on our lower bound but that is a very tight bound, about 1% either way, pollsters usually allow a c 2-3% margin.
Well, yes, the algorithms behaved very well, but the humans (mainly me) didn’t. The reason was the worry over social media biassing too liberal/labour, as it did in 2010 and 2105. Now we knew it wasn’t as biassed as 2015, given the larger demographic now on it, but we thought there would be a bias, which we put at somewhere between 3 and 8%, so our adjusted eventual range was 10% to 30%.
In short – fire the human, promote the algorithms!
(To be fair, this was still a range from hung to “slightly better than last time”, with a median of 17.5%, ie a Tory win but just, and it was a damn sight better than nearly all the polls.)