I use many stats. I use many stats. Let me tell you, you have stats that are far worse than the ones that I use. I use many stats.
Friday, July 21, 2006
Predicting the Winners of College Bowl Games
Ever notice how bowl games seem to be such crap shoots? Take for example the case of the Cal Golden Bears circa 2004. They roll through the regular season with a 10-1 record losing only to eventual national champion USC by 6 points in Los Angeles. Besides USC, the only other close game they played was against Oregon, whom they beat by a single point. They throttled everyone else. Then in the Holiday Bowl against Texas Tech, a mediocre 7-4 team who lost to New Mexico, they get beat by 14 in a game that was not that close. Or how about 2005, when a 9-2 Auburn team whose only setback since Labor Day was a 3 point loss at Death Valley against LSU. In the Capital One Bowl against a 9-3 Wisconsin team that lose to Iowa (7-5) and Northwestern (7-5) and needed a miracle blocked punt to beat Minnesota (7-5) they manage only 10 points and lose 24-10. Is this a common refrain or do bowl games follow some sort of pattern in their results. To find out I examined every college bowl game from the 2000-2001 bowl season to the recently completed 2005-2006 bowl season. I included 5 categories in the study. I looked at how often the team with the better overall record, better point differential, better conference record, better offense (points per game), and better defense (points allowed per game) pre bowl won the game. Here are the results for each season.
2000-01
Better Record: 15-6 (4 bowls matched teams with the same record) .714
Better Point Differential: 14-11 .560
Better Conference Record: 10-6 (9 with same record or Independent teams) .625
Better Offense: 15-10 .600
Better Defense: 13-12 .520
2001-02
Better Record: 13-12 .520
Better Point Differential: 16-9 .640
Better Conference Record: 9-11 (5 with same record or Independent teams) .450
Better Offense: 13-12 .520
Better Defense: 15-10 .600
2002-03
Better Record: 14-11 (3 bowls matched teams with the same record) .560
Better Point Differential: 15-13 .536
Better Conference Record: 9-10 (9 with same record or Independent teams) .474
Better Offense: 13-15 .464
Better Defense: 15-12 (1 bowl matched teams with the same average point allowed) .556
2003-04
Better Record: 14-8 (6 bowls matched teams with the same record) .636
Better Point Differential: 17-11 .607
Better Conference Record: 9-11 (8 with same record or Independent teams) .450
Better Offense: 13-15 .464
Better Defense: 18-10 .643
2004-05
Better Record: 9-7 (12 bowls matched teams with the same record) .563
Better Point Differential: 16-12 .571
Better Conference Record: 10-9 (9 with same record or Independent teams) .526
Better Offense: 14-14 .500
Better Defense: 16-12 .571
2005-06
Better Record: 11-7 (9 bowls matched teams with the same record) .611
Better Point Differential: 14-12 (1 bowl matched teams with same point differential) .538
Better Conference Record: 6-13 (8 with same record or Independent teams) .316
Better Offense: 13-14 .481
Better Defense: 16-11 .593
Total:
Better Record: 76-51 (34 bowls matched teams with the same record) .598
Better Point Differential: 92-68 (1 bowl matched teams with same point differential) .575
Better Conference Record: 53-60 (48 with same record or Independent teams) .469
Better Offense: 81-80 .503
Better Defense: 93-67 (1 bowl matched teams with the same average point allowed) .581
A few things. Its amazing how consistent most of these measures are. With the exception of 2000-01, the better offense has alternated winning slightly more than half and slightly less than half the time. The better record has been either the best predictor or second best predictor every year except for one (2001-02). It appears the bowl upsets mentioned above are relatively rare. The team with the superior record wins almost 60% of the time. These upsets simply stick out in our minds more so than the games that play out like we expect. If you're predicting bowl games, and the teams have the same record, err on the side of the better defense or the team with the better point differential. If you go with the better offense, you might as well flip a coin. If you go by better conference record, then your the guy book makers love. Speaking of which, look for a post next week analyzing the same thing, but this time looking at how bowl games play out against the spread.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment