A part of his current support would never have considered voting for a communist a few months ago. I think there's a chance they reconsider. Not talking about his original supporters obviously.
Not so sure about Fillon. Some people dont care that he is a corrupt establishment dinosaur and will vote for him for practical reasons (thinking they'll pay less taxes with him for example), or because they always vote for the same party. His polling numbers never got really low.

I agree about Le Pen. Her only winnable duel is against Fillon, and she needs polls to underestimate him in the first round and underestimate her in the second. I took a look at Nate Silver's tweets and besides his worries about herding he seems to agree too:

I can't see how Le Pen has a chance when Macron is outpolling her in the first round.

Bringing up Nate Silver in any conversation should be a bannable offense.

Guys, pollsters make honest mistakes, and they learn from them, and unicorns.

Wat.

fillon is a lock to 2nd round

I'll guess fillon/melenchon

better than le pen/anything or macron/anything or GOD FORBID le pen/macron

That is duplicitous. The Trump error wasn't small in individual states and even if it had been he himself abandoned the polls and was pulling numbers out of his arse by the time of the election. He had totally lost faith in his own model and was trying to salvage his own reputation.

Moreover he glossed over the biggest ****-up of his career during the UK election where he was out by over ten points.
(and if you want to say that was just bad luck, no it wasn't, I called it within a few %, so did many other people. He just didn't do his homework. ).

Lmao, this is totally false. Hilarious that you're talking about him pulling stuff out of his ass, when you extracted this whole fiction from yours.

kinda interesting. not much movement between le pen and ps or mélenchon. classic social democrats fairly evenly split between the neoliberals and the harder left as in many countries.



https://twitter.com/mathieugallard/s...33665229451264

First time I see that kind of graph. That's very cool. I was expecting more lepen to mélenchon movement.

He abandoned his own model- http://www.huffingtonpost.com/entry/...b0d9ce6fbc6f7f .

He did so because it was obvious his data-based methods were not working. He was right to do so in terms of theactual result, but in the process he made his entire career a joke. Either you believe in statistical forecasting or you don't.

I believe in statistical forecasting. This article was roundly criticized at the time. You can look back for posts itf about this. The author Ryan Grim is not a statistician and shows his ignorance in the article. He apologized to Silver after the election.

http://www.mediaite.com/online/huffp...rumps-chances/

(which btw shows his results orientedness)

But even the HuffPo article does not accuse him of changing methodology midstream. Indeed, he did not do that.

ok then.

This. That article is/was a complete joke and anyone with a basic understanding of statistics/modelling should easily be able to see that the author has no clue what he was talking about.

How? It looks like a toss up between 4 candidates for the first round. The polls are too close.

So, how did he manage to stay true to his mathematical model when the poll averages were predicting a Clinton win? That doesn't make sense. Either he believes in data or he's using subjective assessments. His whole schtick was big data beats subjective assessments. To call that election correctly you needed to know that disaffected white voters weren't showing up in the polls, that requires subjective analysis.

Silver was out 10 points in the 2015 UK election and I stopped listening to him after that. You can't **** up that badly and expect to have any kind of credibility as a commercial forecaster. That's not standard error, that's a major structural problem with your data.

Um, what? Most of the polls predicted a Clinton win.

Are you just making **** up here or what?

youre the dumbest

What an amazing comeback.

To prevent more Silver fanboyism, you can read 538's own analysis of their own gargantuan ****-up of the 2015 Uk election debacle here:

https://fivethirtyeight.com/datalab/...lection-model/

https://fivethirtyeight.com/datalab/...lection-model/

Hopefully no one is going to argue about what 538 says about 538.

maybe just shut up and be dumb about nate silver and british politics in some other thread

Maybe you should do that yourself mate since you have nothing substantive to contribute and are clearly out of your depth intellectually.

Like, it's all just word vomit.

Which part of "Silver gets things wrong" are you having trouble with?

Did you see the link above where he explained why he gets thing wrong?

"voila"