Friday, October 03, 2008

Sagarin vs. Vegas, the resurrection

For a couple of seasons now I've run a little experiment because I wanted to see how we would do if we used Jeff Sagarin's ratings to help us make "investments" in the college football market. His track record hasn't been so good. Last season his computer ratings went 26-29 against the spread. If you had wagered $110 per game you'd have been $590 in the hole. If you had gone the other way and picked the exact opposite of what Sagarin's computer suggests you would have been up a measly $40.

Anyway, I've decided to do it again. I will pre-select 3 games each week and place an imaginary wager of $110 (to win $100) on the team that the system suggests.

A few notes on the system:

I am using Sagarin's current "Predictor" ratings. The way it works according to Sagarin is:

To make predictions for upcoming games, simply compare the RATINGS of the teams in question and allow an ADDITIONAL 3 points for the home team. Thus, for example, a HOME team with a rating of 92 would be favored by 5 points over a VISITING team having a rating of 90. Or a VISITING team with a rating of 89 would be favored by 7 points over a HOME team having a rating of 79.

NOTE: Use whatever home advantage is listed in the output below. In the example just above, a home edge of 3 was shown for illustrative purposes. The home edge will vary during the season.
Another thing to keep in mind is that Sagarin's ratings are supposed to become more accurate as the season progresses and a relationship is established between teams relative to Strength of Schedule (SoS).

Remember, the purpose here is to find out what happens when we take Sagarin's advice and try to apply it in Vegas. The point spreads I'm using are from are published by this sportsbook.

There's only 3 possible scenarios for each game we analyze:
  1. The first scenario is the easiest to understand. That's when Sagarin and Vegas disagree on who the favorite is. In this case Sagarin would suggest taking his favorite since that's the Vegas underdog and is getting points.
  2. The second scenario is when Sagarin and Vegas agree on who the favorite is but Sagarin believes the Margin of victory will be more than Vegas. In this case you select the favorite since Vegas will require you to give less points than Sagarin believes the MoV will be.
  3. The third scenario is where Sagarin and Vegas agree on who the favorite is but Sagarin believes the MoV will be smaller than the Vegas spread. Here you would take the underdog to cover the spread that Sagarin believes is too high.
So here's this week's picks:
Florida at Arkansas
Vegas Favorite: Florida -24.5
Sagarin Predictor Rating for Florida: 93.36
Sagarin Predictor Rating for Arkansas: 59.01
Sagarin Spread after accounting for home field (2.87 pts): Florida -31.48

Sagarin suggests: Take the Gators and lay the 24.5 points.

Illinois at Michigan
Vegas Favorite: Michigan -2
Sagarin Predictor Rating for Illinois: 72.89
Sagarin Predictor Rating for Michigan: 72.43
Sagarin Spread after accounting for home field (2.87 pts): Michigan -2.41

Sagarin suggests: Take the Wolverines and lay the 2.

Penn State at Purdue
Vegas Favorite: Penn State -13.5
Sagarin Predictor Rating for Penn State: 92.55
Sagarin Predictor Rating for Purdue: 68.46
Sagarin Spread after accounting for home field (2.87 pts): Penn State -21.22

Sagarin suggests: Take the Nittany Lions and lay the 13.5.
So we're taking three favorites this week. Michigan is more of a lean since the Sagarin's spread is very close to the sportsbook spread.


jj gator said...

Michigan deserves a spanking for the shitty way they played against Illinoise. Why did they even bother to show up for the game? That's the way Michigan played.

I'm surprised Ron Douchebag Zook didn't egg his players on to start a fight after this game like he did after Illinoise beat Michigan State in 2005 in their own house and did the same thing to Ohio State last year. That would have been the ultimate in low class.

PhilipVU94 said...

For this to really work well, at least to simulate how successful sports bettors operate, I'd think you'd need to pick the 3-5 games each week with the biggest delta btw. Sagarin and the Vegas line.

From your listing of Michigan-Illinois, with only 0.41 points delta, it doesn't look like this was a very strong play.

Also, off-topic, is your blogroll based on merit or comprehensive? I write a Vanderbilt-oriented blog with some stats-heavy posts, but perhaps it doesn't make the cut for quality or whatever.