So, this 'debate' then. God, did you really watch those two? I had a lovely soothing evening - and, by the sounds of things, I made the right call. Anyway, on the *face of it*, a split decision and a slight BoJo win. /THREAD 
However, consider these appalling #Corbyn leader ratings from the last time these pollsters reported. He came on and made a cogent case. He held his own. Yeah, he lied through his teeth a bit, but who doesn't? His next set of numbers will be better, perhaps by quite a bit.
Viewers' expectations will have been confounded. If we consult Lord Ashcroft's latest polling dashboard, we find that among all voters, 36% expected Johnson to win, against only 18% thinking Corbz would shade it.
Among the detail of these @YouGov figures are some startling numbers: among those still making their minds up, Corbz won 54-38, and among the 'Not Sures' his victory was an even bigger 67-24.
@YouGov This reminds me strongly of the third US Presidential Debate in 2016. Clinton won by ten points - overall. But lurking in the data was something which worried me a lot at the time: Trump won Independents.
@YouGov This can matter a lot. Take a look at what Labour's glossy 1987 campaign did for Neil Kinnock: from -30 to -13 (Ipsos-Mori). Not bad. Now Corbz is a lot, lot further behind, but this stuff can make a difference.
@YouGov So yes, subsets, unweighted, small numbers, etc. etc. And Jez didn't do well enough to change things around entirely. But together with the free broadband stuff, this will be more modest wind in Labour's sails.
@YouGov They were going to close in anyway, and the counterfactual of last night - Swinson's appearance - shows how badly they now need to squeeze LDs. They're going to get closer, but what we don't know is - how close? /END
