Three days after FiveThirtyEight’s Nate Silver penned a blistering critique of rival election forecaster Sam Wang’s model, Wang has responded in kind, arguing that his model has “matched or outperformed” Silver’s since 2008.
Wang’s rebuttal – which, like Silver’s piece, was published by Taegan Goddard’s Political Wire – contends that Silver’s criticism of Wang’s Princeton Election Consortium relies on “a number of factual and conceptual errors.”
While Silver’s post last week asserted that Wang’s methodology would have generated inaccurate predictions in the past – like a GOP Senate majority after the 2012 elections – Wang defends his model as fundamentally sound. Silver’s post assailed Wang’s model for weighing polls from June, but Wang writes that counting from June at this point in 2012 would have given Democrats a roughly 70 percent shot at maintaining Senate control. Moreover, Wang points out, “on Election Eve in 2012, [the Princeton Election Consortium] called every close Senate race correctly - 10 out of 10.”
Wang also seeks to clear up the misperception that he’s unduly bullish on Democrats’ chances for keeping the Senate this year. Although Wang’s model currently gives Democrats a 51 percent chance of retaining the chamber – a sharp departure from other forecasters, who give the GOP the edge – he writes that the Princeton Election Consortium’s prediction “sits at the end of a very narrow range.” Critics, Wang charges, make the mistake “of mentally rounding up to 100% any probability that is even a hair over 50%.”
The wonk civil war may show no signs of abating anytime soon, but Wang concludes by trying to pooh-pooh its significance. Wang and Silver “aren’t the players” in the midterms match, the Princeton neuroscientist writes; they’re “in the announcers’ booth.”