The thing is, the actions of those pundits worked in Trump's favour. Clinton was seen as a political insider, any one-sided story could be seen by Trump supporters as the establishment trying to protect one of their own. I lost count of the number of YouTube videos with one-sided Trump bashing leaving people in the comments saying it made them more likely to vote for Trump.
Honestly, I'm still a fan. He, unlike most other models, gave Trump a healthy shot and cautioned repeatedly that a large polling miss was possible this year. That's pretty much what happened.
To be fair, he was consistently significantly less favorable towards Clinton than any of the other high profile models. He even took quite a bit of flak for this in the weekend leading up to the election, but he stood his ground. While he was ultimately wrong, he was closer than mostly anyone else running a serious model.
I would hope people on this board would understand statistics. Silver gave Trump about a 1/3 chance of winning, and he said the outcomes could run from a Clinton blowout to a narrow Trump victory, which is what happened.
If people were interpreting his "33%" as "0%", despite all his warnings and explanations about what was possible, that's their problem, not Silver's.
I think he came out of this with his reputation enhanced: he gave Trump a 28% chance of winning the morning of the election which was vastly more than any other model.
I think about polling and forecasting a lot, and I've concluded that the presentation of forecasts is a big problem. People often seem to conflate "30% chance" with "30% of the vote" which is clearly wrong - or something like this anyway.
Silver went though the last month going "25% is a big chance!!" and no one really listened. I'm not sure what a better way to express it is though.
He put Trump's chances at near 30% leading up to the election. How can you ask for better than that? I don't think you can argue that an upset like this would be expected to happen more than thirty percent of the time.
Nate Silver's calls for caution are looking like some of the most reasonable pre-election commentary at this point.
Edit: I know, not EVERY one of them.