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.
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.