More Support Last Weekend for the Anti-BCSers

In my post a week ago today about the quest for determining a national champion in college football ["The BCS System: Could it be an illegal restraint on trade?" (9/4/05)], I questioned both the fairness and the reliability of the BCS computer selection process, which takes into account, among other things, the rankings in various polls.

Well, here are the results of the college games this past weekend in which the polls just didn't get it right...again:

1. No. 6 Ohio State lost to No. 16 Penn State (who's coach for some reason is not considered "old" this year)

2. No. 10 California lost to No. 20 UCLA

3. No. 14 Wisconsin lost to unranked Northwestern

4. No. 17 Arizona State lost to No. 25 Oregon (by 2 touchdowns)

5. No. 21 Michigan lost to unranked Minnesota.

At first glance, 5 inaccurately predicted games by the polls doesn't appear to be too many. But the flip-side is that the polls accurately predicted only 11 games! Do the math and last weekend the polls only had a 68% accuracy rate. What's even more interesting is that out of all 11 games that the polls got right, only one game (No. 5 Georgia vs. No. 8 Tennessee) involved two ranked teams playing against each other. The other 10 games were situations in which a ranked team beat an unranked team without a very good record -- So in other words, predicting this past weekend for the polls should have been pretty easy. Ironically, the toughest games in which to predict the outcome were those games listed above and the polls were wrong in each case, which arguably undermines the reliability of the BCS system.

Posted in: