Nate Silver has been using pretty sound statistical methodology to forecast the US presidential election (and senatorial races) on his blog. He did so exceedingly, getting the outcome in 50 out of 50 states right:
and therefore also nailing the distribution of votes in the electoral college.
Or as XKCD put it:
Such predictions are a good thing! They allow to take the focus away from the currently dominant “who’s gonna win it” narrative and to talk about what’s the winner going to do politically. It could also be a good moment to talk about changes to the electoral college, for instance, since, once again, the distribution of electoral votes did not mirror the distribution of the popular vote.
Not everybody agrees, however, as chronicled here. Because as long as one can throw up smokescreens of “who’s gonna win it”, the policies are not looked at in detail…
Update: Martin Robbins weighs in, pointing out that UK pundits had as much trouble to come to terms with the fact that the numbers told a different story from their intuition:
The embarrassing truth is that there was nothing remarkable about Nate Silver’s model, as the man himself freely admitted. He understood that the swing states are the ones that matter, he had a good understanding of the reliability of the sophisticated polling operations in those states, and he could see that when the polls were aggregated the public consistently favoured Obama. None of this was rocket science, yet to political commentators and their ‘conventional wisdom’ it was as alien as witchcraft.
Leading the anti-reality charge in the UK were the Telegraph’s Tim Stanley and Janet Daley. On November 5th, Stanley could be found patiently explaining that Mitt had the X-Factor, as polls showed even Florida slipping from the Republican’s grasp. His ‘methodology’ relied on cherry-picking three polls out of the hundreds available, all put in the field by Republican supporters, while misunderstanding a Pew study by apparently assuming it failed to account for likely voters. Janey Daley failed to meet even this standard a day later, declaring “I believe Mitt Romney will win tonight” on the basis that “Obama’s campaign does not look or feel like it is winning – and Mitt Romney’s does.” That this is considered an acceptable standard for a broadsheet newspaper columnist speaks volumes.