In "The Favorite Bias in College Football Bowl Lines", I took a look at the well-known bowl angle of playing underdogs on the moneyline. Assuming that continues to hold up, there's not much need for any further handicapping during bowl season. However, I also mentioned a few signs that point to this being a weak or even losing year for that system. For that reason, it is best not to put all your handicapping eggs in one basket.
Here are a few more commonly known bowl game situations, and their results according to a database going back to 1983. A Z-factor of greater than 2 means less than a 5% chance the result was due to luck; a Z-factor of 3 means there is less than a 1% chance of the record occurring randomly.
Play Underdogs in December, Favorites in January
Last time, I mentioned the underdog's 57% ATS winning percentage since 1997. That worsens slightly to a little above 54% going back to 1983. Playing the underdogs in the early bowl games is a basic contrarian strategy which aims to take advantage of the increased betting of recreational bettors during bowl season. The December bowls often feature lesser known teams who have received little media coverage and command less public attention, leading to lopsided betting on "name-brand" teams or members of the larger conferences. This angle has some historical backing - December bowl underdogs have covered an impressive 58% of the time. Narrowing it down to only the largest (greater than a touchdown) underdogs returns a 48-25 record for 66% over 25 years.
The games on New Year's Day and beyond are much more prestigious. They carry a higher motivation factor, so favorites are less likely to come out "flat". They also allow much greater preparation time, which might benefit the "better team" (generally the betting favorite). And the relatively close match-ups, combined with plenty of media coverage, might sway the favorite bias back towards neutral, as "square" bettors are influenced towards predicting an upset. Whatever the reason, favorites do slightly overperform in January, covering a little more than half the games, but nonetheless unimpressive overall. Large favorites (greater than a TD) do come in at better than 57% but that is a very small sample of only 40 games, too little to be worth seriously considering.
System: Play bowl underdogs of 7.5 or more points in December
Historical record: 48-25-3 (65.8%) since 1983; Z-factor 2.69
2007-08 plays: Navy +8.5, Southern Mississippi +11, East Carolina +10.5, Central Michigan +8.5
Play Underdogs With The Tougher Season Schedule
One issue with handicapping bowl games is the vast differences in schedule strength between conferences. While increased inter-conference play has eased this burden a bit recently, accurately handicapping small conference schools against large conference schools can still pose problems. One angle is to look at teams which have played the tougher regular season schedule but are still underdogs in their bowl games. Bettors may be overlooking these teams due to their conference affiliation or lack of media attention.
One way to get a handle on the quality-of-opponents issue is to use strength-of-schedule power ratings like those produced by Jeff Sagarin. Going back to 1997, underdogs with higher Sagarin strength of schedule ratings than their opponents went 52-28-1, for 65% against the spread. Of course, that time period was a good one for underdogs in general, with them winning almost 60% across the board, so it is difficult to say whether this is a legitimate angle or just random noise in an underdog subset. Lending some credence is that the strength-of-schedule qualifiers covered their games by an average of 4.5 points more than underdogs as a whole in that span. And the larger the dog, the better the play; underdogs of more than a field goal qualifying in this system went 71%, and those of more than a TD nearly 79%.
System: Play on any underdog with a higher Sagarin strength-of-schedule rating than their opponent
Historical record: 52-28-1 (65%) since 1997; Z-factor 2.68
2007-08 plays: UCLA +6, Eastern Carolina +10.5, Texas A&M +5.5, Mississippi St +3, Colorado +3.5, Oregon +6.5, Florida State +3, Virginia +6, West Virginia +8
Go under in January Bowls
Like the first angle, unders in the January system places extra importance on the prestige of bowl games. These are the best teams in the country, with plenty of time to game plan and plenty of motivation, which should result in hard fought games matching competent teams. In general, these teams should be able to execute on both sides of the ball and should be less likely to make costly mistakes that will put quick points on the board. Totals in bowl games show no overall bias towards either the over or under since 1997. Totals in December tend to favor the over (54%), while totals in January tend to favor the under (56%), though neither is statistically significant. Playing under only the largest totals (higher than 50) in January improves that to about 65% winners. If we narrow that further by eliminating "coin-flip" games and focusing only on games where one team is favored by more than a field goal, the under comes in at 74%. With the recent move of the International and GMAC bowls to January, it may be best to ignore the low-profile January games and restrict this situation to only the traditional January bowls.
System: Play under in any January bowl games with a total above 50 where one team is favored by 3.5 or more
Historical record: 25-9-2 since 1995; Z-factor 2.74
2007-08 plays: Florida-Michigan under 59, Virginia-Texas Tech under 59, Georgia-Hawaii under 69, West Virginia-Oklahoma under 63, Kansas-Virginia Tech under 53
Although these angles look good in hindsight, keep in mind that this bowl season seems slightly aberrant from past years as mentioned in the previous article. That may not bode well for those who try to exploit situations that have performed well historically. Of course, two of the three systems are tied into underdogs, which may underperform this year. Also keep in mind that the sports markets are dynamic and constantly evolving, so those edges that existed in prior years may no longer be relevant.
If you would like to make or read comments about this article, you may do so by visiting the Mess Hall forum at MajorWager.com where a thread has been started. Please click HERE