Human predictions follow a reliable formula: What's Going To Happen Next ≈ What Just Happened.
That is how Dr. Matt "Saturday" Hinton started his post yesterday on the odds that one particular betting website is giving on next year's national champion. All five BCS winners made the top 12, and they went three-for-three at the top: Alabama (7-2), Ohio State (13-2), and Boise State (12-1).
Now, I'm not here to tell you in February that these odds are either sane or insane. It's too early for either of those judgments. However, I can tell you that simply looking at the big bowl winners is not necessarily a good plan for picking a champion. Here's all the champs of the BCS era (including the '03 AP champs) and how they fared in their bowl games the previous season:
SEASON | CHAMPION | PREVIOUS BOWL |
2009 | Alabama | L vs Utah, 31-17 |
2008 | Florida | L vs Michigan, 41-35 |
2007 | LSU | W vs ND, 41-14 |
2006 | Florida | W vs Iowa, 31-24 |
2005 | Texas | W vs Michigan, 38-37 |
2004 | USC | W vs Michigan, 28-14 |
2003 | LSU | L to Texas, 35-20 |
2003 | USC | W vs Iowa, 38-17 |
2002 | Ohio St. | L to S. Carolina, 31-28 |
2001 | Miami (FL) | W vs Florida, 37-20 |
2000 | Oklahoma | L to Ole Miss, 27-25 |
1999 | Florida St. | L to Tennessee, 23-16 |
1998 | Tennessee | L to Nebraska, 42-17 |
Of the 13 teams listed, only six actually won their bowl game the previous season. Tennessee and Florida State lost to the national champions of the prior year, so that's understandable I guess. However, this should temper some of the momentum behind Alabama and Ohio State for 2010; after all, only four champions of the BCS era won their bowl by double digits the prior season. That's nearly equal to the three teams that lost their bowls by double digits the previous year.
If you want to read into things beyond reason, there are some patterns. Each champ from 1999-2002 played a bowl game against an SEC team the prior year. From 2003-08 each BCS champ but '03 LSU played a Big Ten team or Notre Dame in its bowl the previous year, and Michigan appeared three times and Iowa twice. That's not terribly useful, but it's interesting.
Seeing how bowl victories and losses correspond to winning in the next year is a much bigger study, but I'd imagine any correlation would be fairly weak. I love watching bowls as much as anyone does, but just keep in mind that they don't have much predictive power in relation to picking a champion.