Vegas Knows Best: Yesterday Was Once Tomorrow

After taking last week off, this week's Vegas segment will look at preseason expectations and determine how badly "we" misjudged based on against-the-spread numbers.  The conclusion will be that preseason expectations are crap, but you already knew that.

One way to judge a team against prior expectations is to look at their record against the spread. Vegas' lines aren't designed to totally split betting as is sometimes reported, but they do adjust true linesmakers' odds (ie the "true spread") if one particular school is being bet on too heavily (often the result of the general public being very high on a particular team).

First, we should review some preseason expectations in order to see where each team was pegged in August. Next to each preseason rank is the current Big Ten rank based on in-conference record. This is clearly not any type of predictor as to how they will finish because, obviously, there are more games to play.  But it's what we have so far and, honestly, probably not that far off from how things will end.

It's worth nothing that every single preseason ranking (comp chart found here, and linked in the table) had Ohio State winning the thing.  Wisconsin was a popular pick for the two spot and Minnesota found itself sporting the last position in 10 of the 15 polls.  Penn State was put as high as two and as low as seven.

I used a little bit of fuzzy math on the current standings.  I wanted a round number of 66, the sum of the standings, and so several teams got half points.  I also gave teams with more wins/losses more/less credit, which might not be fair but this isn't a research paper or the BCS standings so it shouldn't bother anyone.

So listed in order of their Delta, with two new columns: the current record against the Vegas spread and that record's winning percentage.

Now the Delta figure has some limits, of course. Teams that start close to the bottom, like Indiana and Purdue, simply don't have that far to drop. Teams with higher expectations though, like Michigan and Wisconsin, have a long fall and thus were able to pick up more points. This works in reverse too, of course.

The ATS numbers are a lot more complex. Since the spread for each game is set a week before kickoff and not all at the beginning of the year, you get evolving expectation built into them. Another thing to keep in mind is that more popular teams, namely Penn State, Ohio State and Michigan, are going to generally perform worse against the spread because of action they get for simply carrying the name. Ohio State is always going to get the benefit of the doubt, and besides, betting on Northwestern-Duke just isn't that much fun.

The three best performers ATS: Minnesota, Penn State and Iowa, represent preseason rank 11, T3 and 8, respectively. Penn State is the exception, but generally speaking picking up these pre-season basement dwellers would have been a good investment. It was highly unlikely that Minnesota, a team that returned most of their starters, was going to equal last season's abomination.

The three worst performers ATS: Indiana, Michigan, Ohio State, represent preseason rank 10, 5, and 1, respectively. Ohio State has been underperforming all year against lesser competition; always winning but never looking very impressive. The reputation of their talent, however, leads most people to think they are simply being sloppy rather than conclude that the team just isn't very good. As a result, the spread continues to be higher than the final score. Indiana and Michigan, well those two teams are terrible, there isn't much else going on there.

So the moral of the story: well for starters, the preseason rankings, even when looked at in consensus, are not very accurate. There are just too many variables. Secondly: even when accurate, the lows are not as low as they seem and the highest ranked teams are a lot more vulnerable than we are lead to believe. Ohio State, fresh off of an MNC game and returning just about all of their key players, was demolished against (an also vulnerable) USC team and will be playing the role of underdog against a team that was ranked pre-season in the late teens. While the prior season's performance and current number of returning starters matters, it's hardly a reliable way to predict future performance and appears to be incredibly overvalued by both the AP voters and the betting public.