FiveThirtyEight.com is the most interesting election horse race site. It’s run by Nate Silver, the Baseball Prospectus stats guy, who does the most thorough analysis of polling data, sophisticated in a way that only a Sabermetrician can fully appreciate. Silver rejects the “tightening race” narrative that we’ve started to hear, as he looks at state polls and projects the Electoral College outcome:
If the state polls aren’t showing movement toward McCain, then it is probably the case that any perceived movement in the national polls is sampling noise. If anything, in fact, the state polls are showing movement toward Obama on balance, not just in battleground states like Virginia, but also in non-battlegrounds as diverse as New York, Oklahoma, Oregon and Arizona.
Movement in the popular vote in non-battleground states is not significant, so let’s not get distracted. But let’s not forget to vote, either (I’ve already voted, thank you very much.)
How did sabermetrics work out in the last three world series?
Sabermetric predictions outperform other types of baseball predictions in any given time frame, Scott, and as we approach election day it’s become quite clear that Palin/McCain are going to lose quite dramatically. The Republican Party’s War on Reason is going to cause several years of wandering about the desert. At the end of this period of soul-searching we may very well see the irrational part split from the rest of the party.
Yeah, I know. I’m wondering how they did in the WS lately, because I’ve been out of the baseball stats loop for a while. I wasn’t making a stupid electoral point. Jeezus.
They’re pretty good, and don’t have the emotional bias of other forms of prediction.
Emotional bias or not, that prediction kinda sucked. No Cubs, no White Sox.