|
|
|
|
SportsQuant RatingsThe SportsQuant college football rankings are based on ratings1 which are estimated for each team in Division I-A. Our hierarchical model estimates both the probability of a win or loss for each team as well as conditional point totals for both situations. By combining this information, we can predict the probability that one team defeats another as well a point-spread. Take the Chick-Fil-A Bowl, for example. The SportsQuant rating model gives Clemson a 58% chance of beating the Auburn. The average score in a Clemson victory is 23-16. If, on the other hand, Auburn wins (which happens with probability 42%), the average score will be 21-17. Here's how to interpret these scores. If your spouse wouldn't let you watch the game, but your friend told you that Clemson had won, what would you expect the score to be? (Answer: 23-16 in favor of Clemson.) What if he told you that Auburn won? (Answer: 21-17 Auburn.) Since you don't know who will win in advance, the prediction is a weighted average of the two possible outcomes (with weights determined by how likely each outcome is). Therefore, the predicted scores (used to calculate the point spread) are: Clemson's expected points = (23)(0.58) + (17)(0.42) ≈
20 So, we'd expect Clemson to be a 2-point favorite (20-18). 1. A ranking is an ordering of teams; a rating is an objective number assigned to each team. Rankings like the Associated Press poll provide an ordering of teams but do not give a measure of inherent ability. Ratings, however, give an objective measure of each team's skill. Rankings can always be constructed using ratings, but not vice versa. Bowl Predictions
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Send mail to
webmaster@sportsquant.com with
questions or comments about this web site.
|