At the beginning of the month, we took a look at what correlation, if any, exists between an AP national title winning football team and its final AP Poll ranking the year before. I said at that time that I wanted to dig up the data for the AP Preseason polls to look at where the eventual national champions were ranked in the preseason of their title-winning year.
Thanks to College Football Resource, whom I owe a giant tip of the hat, we have that data. Let?s have a look:National Champions1
|Year||National Champ||Preseason Rank|
1AP National Champions
For one thing, we can see that the AP voters struggled from 1980-1984. Before we get into further analysis, let me bring you up to speed on what happened in in that five year span:
The Associated Press had a giant swing and miss in 1981. Their top four combined to lose 15 games. Preseason #1 Michigan went 9-3, #2 Oklahoma finished 7-4, #3 Notre Dame (under first year coach Gerry Faust) plummeted to 5-6, while #4 Alabama went 9-2-1.
1982 was a bizarre year, as well. Preseason #3 Alabama, in what would be Bear Bryant?s final year at the helm, scorched eventual national champion Penn State 42-21 early, but stumbled to an 8-4 record down the stretch. Bryant went out on top, of course, beating Illinois in the Liberty Bowl, 21-15. The Nittany Lions? loss to ?Bama would be their only of the year, while AP preseason favorites #1 Pittsburgh and #2 Washington combined for five losses.
Things almost worked out as the AP predicted in 1983. #1 Nebraska cruised through the season undefeated before #5 Miami, unranked in the preseason, met them in the Orange Bowl. The Hurricanes eeked out a 31-30 win in a classic, taking the national title. Texas, for their part, entered the season ranked #3 under Fred Akers, and ran the regular season table, earning a berth in the Cotton Bowl. Ranked #2, the Longhorns squared with #7 Georgia, losing a heartbreaker 10-9. Had they won, the Miami victory over Nebraska would have given Texas its first national title since 1970. All the sweeter for Mack Brown, I suppose.
1984 was the year of the pass attack, with
Steve Young Robbie Bosco leading Brigham Young to the national title and Doug Flutie winning the Heisman Trophy (and creating the most memorable Hail Mary pass in CFB history). The AP was well of at the top of its preseason ballot. Two of its top four teams (Pittsburgh and Clemson) finished unranked. Auburn, the preseason #1 team, dropped four games.
After 1984, though, the AP seemed to do a much better job with their preseason rankings. The average national champion was ranked 6.32 in the preseason between 1985-2005. Not once was the eventual national champion found outside the AP Preseason Top 20. Interestingly, 2 of the only 3 teams to come from outside the Top 10 preseason rankings to win the national title have done so in the BCS era.
There?s a lot to digest in here, but let?s take a stab at some preliminary questions.
*Could the continuing growth of sports media, and ESPN, throughout the 1980s have helped sports writers be more informed about their voting? A cursory glance at the history of ESPN is very telling. In 1982 they began televising bowl games. In 1984, they began covering regular season college football. Coincidence?
Now, if I had the time, I?d enjoy running a standard deviation check on AP preseason rankings (not just for the national champ, but for the entire poll) over the last 25 years and see whether its decreased in any meaningful way as voters are able to digest more and more information on teams. Is big media actually improving the ranking of teams? Or, could there possibly be an opposite effect, where the media?s ability to drive voter opinion stabilizes polls? These are interesting, worthwhile questions.
*Is the BCS actually doing a good job of giving teams slightly off the radar a better chance to earn a trip to the title game? Do the inclusion of the computer rankings help identify teams that human voters may be undervaluing? Is the BCS actually more egalitarian than the previous system? Does this only apply to major conference teams?
Unfortunately, the limited data that I?m working with only allows me to ask some of the right questions, but in the scientific community, that?s the first step. These are the things we should be thinking about, especially in the era when everyone has an opinion about the BCS, playoffs, plus-one scenarios, and all the rest. Has anyone taken the time to study the history of how this has all worked? And where it can be improved? If you look at any one year, you?re bound to find irregularities, but if you looked at the data in aggregate, there are bound to be useful patterns that can shed light on the best type of system for determining a champion.
As an example, one might run a study on whether or not there would be an added benefit to eliminating preseason rankings altogether. Would releasing the first polls in October alter the final rankings in a meaningful way? Why isn?t someone looking at this with a control group of AP (or other) voters?
Fortunately, in this era, anyone and everyone can do this if they have the desire and commitment. Blogs are popping up everywhere, new statistical databases are evolving, and smart, dedicated writers are asking and answering these questions all the time. It?s not unlike any other evolution in sports analysis. Baseball?s been doing this for decades. Is college football next?