I like to think that relative sanity from all sides has helped sports fans (particularly web-savvy sports fans) understand and embrace math to a solid degree. Like I said, there’s some grousing. There are plenty of “watch the game!” comments. But it’s nothing like on the political stage. The abuse Nate Silver has received this election cycle for having the temerity to average polls based on a previously successful regression model is astounding. This is pretty straightforward math. It’s actually quite similar to Hollinger’s Power Rankings. Hollinger takes efficiency differential (which is a more precise indication of team win-loss record) and adjusts for schedule strength (including home-road split) and emphasizes recent play. Silver’s Five Thirty Eight model averages the polls, adjusting for house effects and timing. Like I said, basic math.
Hollinger hears it from math-denying fans, but no columnists, analysts or talking heads rip him openly. Silver is getting challenged by one of the highest-rated political commentators on television. And many others. It’s all a bit crazy, especially when Silver plainly says his work produces probabilities of outcomes, not certainties. Are the anti-Silver commentators really that … for lack of a better word, dumb? (Note: this isn’t to say anything of those who seek to discredit the mechanics of Silver’s model, with the arguments that he overemphasizes state polls or that his house effects adjustments are off. This is about the math deniers that have popped up as the math has told a story that doesn’t agree with their partisan sensibilities.)
Good reading from Ziller on the bizarre turn the political punditry has taken against Nate Silver’s probabilities this election cycle.
I avoid politics here, as you’ve likely realized. But if you’ve somehow missed all the backlash to Silver’s analysis this month, it’s worth investigating. Lots of silliness.