NFL prediction time is here again with writers from various outlets providing their views. Here are SI's Preview predictions. Betting markets such as Vegas Insider supply futures bets on AFC, NFC, and Super Bowl Champions from which relative rankings of teams can be inferred.
Last year at this time, I collected the predictions from senior SI writers Peter King and Dr. Z along with rankings implied by Vegas futures. For conference rankings, Peter King's accuracy (measured by R-square from regressions with actual rankings) was only 11 percent. Dr. Z came in second at 21 percent with Vegas topping the group at 26 percent (Yahoo didn't provide comparisons). At the divisional level rankings, Dr. Z came in at 8 percent, King at 14 percent and Vegas at 15 percent.
A couple of observations come to mind. First, the NFL future is an uncertain place regardless of whose predictions one wants to use. Second, few, if any, economists would be surprised by the outcome. In the end, the writer picks are just one guy's view. The Vegas line represents an initial estimate by a stat-based person or group that then gets adjusted up or down by the opinions of hundreds or thousands of bettors. In fact, I expected the Vegas advantage to be a bit higher than it actually was. In looking at the big misses, they tend to be the same whether writer- Vegas-based. Nobody had the Saints doing anything close to what they did. The Redskins tended to be overvalued, and so on. Third, most of the differences lie down in the middle and bottom teams. Vegas is picking a New England-Chicago Super Bowl with Indianapolis and the Saints as runner-ups. That likely reflects what many writers would pick.