Most pre-season prognosticators had the Georgia Bulldogs atop the polls to start the year. A strong finish to the 2007 season coupled with incumbent quarterback Matthew Stafford, star running back Knowshon Moreno, and freshman phenom A.J. Green led many to believe the Bulldogs had the inside track at the national championship.

Obviously, the crystal football will not reside in Athens at the end of the season. A closer pre-season examination of Georgia would have shown a lack of experience along the offensive line. As anyone in the coaching business will tell you, exceptional talent at the skill positions is irrelevant without good line play.

Many media pundits either ignored or underestimated the impact of Georgia’s inexperienced offensive line. To be certain, this isn’t the only reason the Bulldog’s didn’t live up to their billing, nor are they the only team who didn’t, but the fact that many so-called experts were wrong casts doubts regarding the validity of pre-season polls. Moreover, pre-season top 10 predictions of Auburn, West Virginia, Clemson, and LSU were even more wrong. Currently none of the aforementioned teams resides in the top 25.

This naturally begs a question, are pre-season polls accurate? Additionally, if pre-season polls are not accurate, does it matter?

The former is answered by comparing the pre-season and year-end polls. The latter can be measured by comparing the year-end polls with an objective measure of team performance, i.e. a computer ranking.

Are Pre-season Rankings Accurate?

The obvious, simple answer is no. But perhaps more important is to what degree inaccuracies in pre-season polls exist.

Only four of the pre-season top 10 teams in the AP Poll currently maintain a top 10 ranking. Four (mentioned above) are no longer ranked in the top 25 and the other two sit at 16 (Georgia) and 25 (Missouri). In other words, 60 percent of the AP pre-season top 10 teams are no longer deemed to be top 10 caliber.

Furthermore, of the remaining top 25 teams in the pre-season AP Poll seven are no longer ranked. All together, only 14 of the top 25 AP pre-season teams are currently ranked.

The USA Today Coaches Poll shows little difference. Three of the pre-season top 10 teams in the Coaches Poll are no longer ranked. Two others now sit at 17 (Georgia) and 23 (Missouri). While slightly better than the AP, only 50 percent of the Coaches Poll pre-season top 10 remain.

Moreover, only five of the remaining top 25 pre-season Coaches Poll teams are still ranked. In summary, only 12 of the pre-season top 25 Coaches Poll teams still hold a ranking.

To better quantify the inaccuracy of the pre-season rankings a direct comparison to the year-end rankings was performed. Any team not ranked in the top 25 was assigned a rank of 26, giving a benefit to the polls by assigning the closest value outside the top 25.

The AP Poll has an average ranking difference of nearly 10 spots when the year-end and pre-season top 25 are compared. The results for the USA Today Poll rankings are nearly identical. Both suggest fairly gross errors in the predictions of team performance prior to the season.

Does It Matter?

If the year-end rankings don’t mimic those of an impartial, objective ranking system, the subjectivity of pre-season rankings have contributed to a bias in the final polls.

For this exercise the AV Ranking will be used as the objective ranking system. Comparing the AP and Coaches Polls to the AV Ranking results in an average rank difference of approximately four spots. This is relatively positive as the difference is fairly small and certainly much less than the differences in the pre-season and year-end polls.

The results are even better when teams with at least one loss and weak strength of schedule rankings are removed. Oregon, BYU, and Northwestern have the 82nd, 114th, and 108th ranked strength of schedules. These three teams have a much lower AV Ranking than either year-end poll. With these teams removed, the average difference between the AV Ranking and final AP and Coaches Polls dips below three.

In Summary…

Pre-season rankings give unfair advantages. Teams that continue to win rise in the rankings so long as the teams ahead of them lose. Similarly, teams rarely fall unless they lose. The subjectivity of the initial pecking order can be biased.

This contributes to a phenomenon where losing early in the season is preferred to losing late. Surprisingly, however, by the end of the year the preference towards the pre-season ranked teams is largely erased. In other words, while the pre-season rankings this year were quite inaccurate, it didn’t make a significant difference in the final AP and Coaches Polls.

The exceptions are teams that played weak schedules and lost at least one game. These teams are largely untested and unproven, i.e. their fate lies more with where they began the season ranked as they have little opportunity to beat a higher ranked opponent.

Instead, they rely on teams above them to lose in order to move up. This is the reason for the large disparity between the year-end AP and Coaches Poll, and the AV Ranking. An objective ranking system does not subjectively start one team ahead of another.

Granted, this analysis only used one year of data. But the large differences in the pre-season and year-end AP and Coaches Polls suggest that even gross errors in pre-season predictions wash out by the end of the season.