I think the typical NFL fan had two surprises this weekend; people who have read my blog or Brian Burke’s only had one, which just happened. The expectation in the divisional round is that all the home teams will win; they’re the top seeds in the playoffs and thus are typically better than their opponents on top of the benefit they get from having home field. So you wouldn’t be too wrong by going chalk and taking the four home teams in this round. In that case you’d be surprised by the Atlanta and New England losses this weekend. But with a little more information, you would have been more likely to dismiss (or at least discount) the Falcons. That leaves the Patriots’ upset, which was surprising for most, I would think. It’s a good reminder that game outcomes are indeed probabilistic; any team can win on any given Sunday, they just differ in how likely they are to win.
How did the models do? Mario and Luigi both expected Chicago/Seattle to go under and the three other games to go over; all four games ended up hitting the over. So in the playoffs Mario is 5-3 and Luigi is 4-4. My models include playoff games so there’s no particular reason to separate their playoff predictions from their regular season predictions, so I would expect both models to go about even. In terms of the moneyline Mario would have taken Seattle and the Jets and thus be 1-2 but ahead; Luigi would have taken Pittsburgh as well and be 2-2 and a bit more ahead. The Jets’ victory alone is enough to put the models ahead since it would have paid $325 on a $100 bet; picking favorites is far less lucrative (Pittsburgh would have only paid about $60) but obviously better than being wrong about them. Luigi split against the spread, only picking Seattle and the Jets to cover. Mario would have also taken Atlanta and been wrong. Finally, Bill Simmons went 1-3 this weekend, only correctly picking Pittsburgh’s cover. Mario went 2-2 and Luigi went 3-1. So in the season-long competition, I picked up another two games on Bill, putting me ahead by four overall.