Comments on: Why Nate Silver and others predicted the election perfectly

By: Guest

Guest — Thu, 24 Jan 2013 02:22:29 +0000

I predicted the outcome with 100% accuracy myself. Romney wasn’t “supposed” to win, and never intended to win. It’s just one of those things you don’t see because you don’t want to see.

By: Randy

Randy — Fri, 09 Nov 2012 23:01:05 +0000

Ah. The pungent scent of sour grapes. This is not about who won and who lost, but about the nothing-short-of-amazing abilities of modern analytics and those who know its subtleties. *You* get over it.

By: Aswath Rao

Aswath Rao — Fri, 09 Nov 2012 02:14:20 +0000

He was 98% confident of his ND Senate seat. So should I chalk it up as a failure or consider it as part of the 2% event? What about Montana race? How many strikes will be required before reliability can be questioned?

By: suteebu

suteebu — Thu, 08 Nov 2012 18:21:44 +0000

Sure, but he also tells us the probability of his predictions, which allows us to assess if he is over- or under-confident in his own predictive ability. E.g., we never hear a pundit say, I’m 60% confident in my predictions, so 4 out of 10 of my predictions will be wrong.

By: suteebu

suteebu — Thu, 08 Nov 2012 18:19:47 +0000

Sure, but he also reports the probability of his predictions — he *should* get some of his calls wrong because he tells us what his confidence in those predictions are. And then we can judge if he has a good grasp on the reliability of his predictions — is he over- or under-confident in his predictive ability?

By: Aswath Rao

Aswath Rao — Thu, 08 Nov 2012 12:40:30 +0000

It should be noted that Nate Silver also cherry picks his polls. After all he places weights on the polls and he has some other “secret sauces”, whatever they may be. Analysis of such data analytics should also note his wrong calls, such as Senate races in Montana and ND.

By: Doug

Doug — Wed, 07 Nov 2012 22:11:10 +0000

You won. Let it go.