You may have heard that after three weeks there are only three undefeated teams left. It took overtime, but the Lions are one of them. It isn’t too often that you get to say that something has literally never happened in your lifetime, but the Lions are 3-0 for the first time since the year before I was born. The Packers are also undefeated, which I suppose isn’t surprising since they just won the Super Bowl (not that I think that guarantees you a 3-0 start or anything), but the other surprising member of the triumvirate is the Buffalo Bills. It wouldn’t be quite as surprising if you had checked my predictions though, which said Buffalo had a pretty good chance of getting the win. Below are how the predictions turned out overall and my first power rankings of the season.
The over/under continues to be a crap shoot. Luigi went 6-10 this weekend, summing up to 11-20-1 on the season. Why do I say a crap shoot when the model is below chance? Because there’s no reason to think that it’s really going to do worse than chance in the long run. The gambler’s fallacy says that the model will have some above-chance weeks coming up, but overall it’s probably just throwing darts and will end up around 50%. Next off-season will be time for some serious evaluation.
In terms of simply predicting the right winner, Luigi went 9-7 and is 19-13 on the season. For moneyline picks, which also take into account the pay-out odds for picking a winner, Luigi went 6-3 to move to 10-10. Two of the underdog picks came through this week in the Bills and Giants.
Picking against Bodog’s spreads, Luigi was 9-7. If you take out the two games that were too close to call, Luigi was 9-5 (neither Cincy not Chicago covered). So in all games Luigi is 18-13-1 and in ‘pickable’ games Luigi is 16-11-1. Bill Simmons had some different spreads, such as Cleveland -3 against Miami whereas I saw it as Cleveland +1 (probably because Hillis was ruled out by then). Bill went 9-7 against his spreads while Luigi went 10-6. We disagreed on three games and I was right about the Colts and Giants while Bill was right about Seattle. On the season then Bill is 26-20-2 and I am 27-19-2.
It’s probably a little early, but I wanted to go ahead and post the power rankings and season predictions. The power rankings are basically the score that each team gets that goes into the predictions; it assumes that each team would be at home against an average opponent. There’s a line in there for ‘average’ to give you a sense of scale.
Despite their loss, the Patriots are number one. Defense is still obviously their problem, but their offense is moving the ball more or less at will. The Lions have fallen to number 6, but you can see that the values are pretty tightly clustered from 2 through 6. Next up are the season predictions, which take into account the schedule; basically, it adds up the currently predicted probability of winning for each team in each game to get ‘expected’ wins. It also does it for previous games so you can see if a team has been lucky or not so far.
The top team is actually Houston, because they have played well so far and get to play in a weak division. But the top five teams are basically the same. You might also note that the Cardinals are awfully high up there. That’s because the model thinks they’ve been unlucky so far (having ‘earned’ over two wins but only beating Carolina on the field; their two losses are by a combined 4 points) and they also get to play a weak division and have some tough games at home (like the Giants and Steelers) that might otherwise be out of hand. At the other end, the Browns are 2-1 but are only expected to get 5 more wins; their conference games are Baltimore and Pittsburgh and on the road in Cincinnati, and they also play at Houston, Oakland, and the just-mentioned Cardinals. Of course, it’s early so these rankings are going to change quite a bit.