[xux]

Whenever I see a prediction from academics, I'm reminded of this: http://www.msnbc.com/the-last-word/watch/scientist-predicts-...

"Princeton professor predicts 99% chance of a Hillary Clinton win"

[asafira]

This is a little unfair, it's not clear to me why academics would always have worse predications than any other body of people. This especially goes for the presidential outcomes predictions.

[azinman2]

Especially when most predictions were for her.

Also, if that summary is true, that meant 1% likelihood for Trump. Not zero. That's how stats work.

[paulddraper]

> it's not clear to me why academics would always have worse predications than any other body of people

You're right; it's a problem with predictions _generally_.

It's just that academics make predictions that carry more credibility.

[nl]

Whenever I comments like this I like to ask what they think the chances of losing the popular vote by 2.8 million (or whatever it ended up as) and winning the electoral vote actually are?

Maybe she did have a 99% chance of winning? I don't think so - but I haven't actually worked it out. I'd be interested to see the conditional probabilities.

538 put Trumps chances at ~20%, which seemed about right overall.

[xanderstrike]

"Assuming a perfectly spherical voter..."