April 2018

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30          

« Statewide news | Main | Well hung »

Nov 07, 2012


Feed You can follow this conversation by subscribing to the comment feed for this post.

Andrew Brod

I saw the Engber piece in Slate earlier today and was struck by how clueless he was. It reminded me of the equally clueless tweet by a Politico writer: "Nate Silver admits he's simply averaging public polls and there is no secret sauce." Not only was it hardly a secret that Silver was averaging polls, but there actually is some secret sauce. Silver doesn't reveal how he does the averaging and weighting of the polling data.

On the other hand, I don't think it's quite right to say that Silver predicted 332 electoral votes, per Slate's pundit scorecard. His actual final forecast was 313 electoral votes.

Yes, Silver predicted that Obama would win enough states to generate 332 electoral votes, but not all states were equally likely wins for the president. It looks like Silver weighted his state forecasts by the probability of each state being won by Obama, in the same way that the expected value of a coin flip is 1/2 rather than 1 (heads) or 0 (tails). Florida was so close in the 538 model that the forecast hedged its bet. And the delayed results in Florida bear that out -- it's still very close!

Even so, it's more accurate to say that Silver predicted 313 electoral votes. If Obama wins Florida, Silver will be a little low; if Romney wins Florida, Silver will be a little high.

The comments to this entry are closed.