The target was New York Times analyst Nate Silver, who was attacked for the crime of using state-level polling data to find that the election was actually pretty stable and Mr. Obama was always more likely than not to win.
Mr. Silver's quasi-academic position derived from the polls ran against the dominant media paradigms at key points in the race. Mr. Silver put forth that Mr. Obama had a better than 61% chance of winning even when headlines were screaming that Mr. Romney was taking control after the first debate. When headlines were calling the race "too close to call" Mr. Silver was saying President Obama had a better than 90% chance of winning.
There was a conservative tint to the criticism, as Mr. Romney supporters refused to believe they were not winning.
What is more, it was remarkably similar to the movie Moneyball, and the rage hurled at quantitative analysts who refused to see the importance of gut and experience in selecting ball players.
The reality is the political analysis the public sees is too often dominated by old chestnuts and insider spin and – pretty much – nonsense.
The fact that no Republican has won the Presidency without Ohio is an interesting piece of trivia but this rear-facing fact renders little help in predicting the future outcome of the election.
The fact that state-level polls show Ohio is consistently backing Mr. Obama over Mr. Romney, and that enough other swing states remain consistently Democrat leaning to get to 270 electoral votes, is crucial in predicting the election.
In the end, Mr. Silver projected every one of the 50 states accurately, and made accurate predictions on most if not all the Senate races. Some will characterize this as a modern re-enactment of the Scopes Monkey Trial, with science victorious over mythology. But I think it proves only that Mr. Silver has the right idea, a good model, and got lucky in places like Florida where he only assigned Mr. Obama a 50.3% probability of winning.
Getting them all right isn't the point because probability isn't about getting it right 100% of the time. It's more about the shape of the race, and the likelihood of certain scenarios.
The role of analytics driven state-level poll aggregation is informative in U.S. Presidential elections precisely because it more accurately follows the process of selecting Presidents than other methods.
Presidents are not picked by Gallup polls, or national vote, or pundit analysis, or media appearances. They win by winning 270 electoral college votes through a coalition of state-level victories. Analysis that focuses on state-level dynamics and aggregates that to the national level with rigour has value beyond the horse race, to better track what is happening and consider the possible outcomes.
This race was not "too close to call." Actually, it was pretty likely Mr. Obama was going to win throughout, and for reasons we will get to now.
4. The new Democrat majority
The term "realignment" gets thrown around a lot.
V.O. Key originally defined the term as a single "critical election" that dramatically shifts the dominance of the political system from one party to another. 1932 is a prime example, when the old Republican dominant coalition collapsed and the New Deal Democrat coalition held the Presidency with few exceptions until 1968.
Realignment since 1932 is harder to detect. Elections from 1964 to 1980 were part of a long process that saw the South move from Democrat bedrock to Republican bedrock. Certainly, realignment took place but it was not packed into a single election.
Bill Clinton picked the Republican lock on the Presidency in 1992 and 1996, reliably adding states like Arkansas, Tennessee, Louisiana and Missouri to a growing Democrat coalition anchored by California. But all those states slipped back to the Republicans in 2000.
Mr. Obama may have created a more durable coalition.
Cast your mind forward to 2016.Report Typo/Error
Follow us on Twitter: