I took the Associated Press Preseason Football Poll and matched it with the Final Poll for the years 1989 – 2009, the years that the A.P. picked 25 teams in their preseason and final polls.   If a team was not ranked in the top 25 at the end of the season, I looked to see if it had received any votes and counted down from 25 to see where it ranked.   If it received no votes, then I assigned it the rank of one more than the number of teams listed.

From 1968 to 1988, the A.P. picked a Top 20 in both their preseason and final polls.   From 1962 to 1967, the Associated Press picked a Top Ten in both Preseason and Final polls. In 1961, the A.P. picked a Top Ten Preseason poll and a Top 20 Final Poll.   From 1950 to 1960, the A.P. picked twenty teams for both the Preseason and Final polls.   Before 1960, the A.P. apparently didn’t pick a Preseason poll.

The correlation coefficient ranges from -1 to 1.   A value of 1 implies that a linear equation describes the relationship between X and Y perfectly, with all data points lying on a line for which Y increases as X increases.   A value of -1 implies that all data points lie on a line for which Y decreases as X increases.   A value of 0 implies that there is no linear correlation between the variables

In the twenty-one years that the Associated Press has been picking twenty-five teams in its Preseason and Final Polls, the best predictions happened in 2001 when the pollsters had a correlation coefficient of 0.75143.   Despite having eleven (11) teams that were ranked in the preseason that were not ranked in the final poll, the top ten teams in the preseason poll all finished in the top 25 at the end of the season.

The worst year for the pollsters was 1994 when the correlation coefficient was just 0.23232.   One of the teams that hurt the rankings that year was Notre Dame, which was picked to finish second, but didn’t even get a vote at the end of the year.