2009′s Final Computer Poll vs AP Poll
At the end of the 2008 season we devised a formula which plugged in the stats from the just completed season and then factored in the opponents and the site the game was played at and came up with an individual team power rating for each game. I was very pleased that the formula did indeed assign a true power rating for each game and it was highly accurate.
In the weekend blog, I gave you the top 15 individual “game grade” performances from 2009. Today I will give you the teams with the highest game grade averages from last year, which make up my computer’s final Top Ten and compare it to the AP Final Top Ten.
First here is the final AP Top Ten from last year.
- Boise St
- Ohio St
- Penn St
- Virginia Tech
Now here is my Computer’s Final Top Ten from last year.
- Alabama 96.2 avg
- Florida 96.0
- Texas 95.7
- TCU 92.4
- Oklahoma 91.2
- Virginia Tech 88.3
- Ohio St 86.0
- Oregon 85.2
- Texas Tech 84.8
- Nebraska 84.5
As you can see there are some differences between the polls. Naturally, Alabama was at the top of both polls. My computer had Florida barely ahead of Texas because the Gators schedule was slightly tougher and they outgained opponents by 205 ypg while Texas outgained their foes by 169 ypg. Despite the loss in the Fiesta Bowl, TCU comes in at #4 thanks to their dominating 12-0 regular season that saw the Horned Frogs outgain foes by 217 ypg! Oklahoma is another surprise here coming in at #5 despite their 5 losses. The Sooners lost 4 games by less than a TD and their +133 ypg in conference play was actually the best in the Big 12. Oregon finished just outside the AP Top Ten at #11 while my computer had them at #8 so there was not much of a difference there. Also Texas Tech and Nebraska appear in my computer’s Top Ten while AP #8 Cincinnati (#23 in my computer poll) and AP #9 Penn St (#14) were left out.
Now before Boise, Cincinnati and Iowa fans start to get riled up with these rankings remember this is my computer’s poll and not the way I would have voted the Final AP poll if I was indeed an AP voter. However, due to past research on comparing my computer poll with the AP poll there is a significant trend with how teams do the next year. Teams that are in the AP Final Top Ten but not in my computers are generally overrated the next year while teams that are in my computer’s Top Ten but not in the AP are underrated.
Let’s take a look at several examples.
In 2004, Iowa finished 10-2 and were ranked #8 in the final AP poll but were just #21 in my computer’s final grades. In 2005, Iowa opened up #11 in the AP poll and they went 7-5 and finished out of the rankings. On the opposite end in 2004, LSU did not finish in the AP Top Ten but my computer had them #8. The next year these “underrated” Tigers finished #6 in the country after their thrashing of #9 Miami Florida 40-3 in the bowl.
In 2005, Alabama was 10-2 and finished with a bowl win over Texas Tech and were #8 in the AP poll. They took on my #55 rated schedule and only averaged 21.9 ppg on offense. My computer did not even have them in the top 15. The next year Alabama finished 6-7 and fired HC Shula.
Also in 2005, Louisville was #19 in the AP poll (#8 in my computer) but were actually the most dominant team in the Big East and would have been in the Sugar Bowl if not for a blown lead at West Virginia. The next year Louisville went 12-1 and their only loss was at Rutgers, a game UL actually led 25-7 but lost on a last second FG. Change one play in that game and Louisville would have played Ohio St in the national title game in 2006! The other team not in the AP Top Ten was Michigan (computer #9), which finished 2005 at just 7-5. Michigan took on the 2nd toughest schedule in the country in 2005 and had losses by 7, 3, 3, 4 and 4! The next year (2006) Michigan was 11-0 and #2 in the country when they played #1 Ohio State in Columbus and after losing that game by just 3, many felt they should have played Ohio State in the title game rather than Florida!
In 2006, despite Florida winning the National Title and playing the best game of the year in the biggest game, they were not the best team over the entire course of the season (Computer #4). LSU had a 22-14 FD edge vs them but the Gators won thanks to being +3 in turnovers. Alabama had a couple of late TO’s and lost a late lead. Florida blocked 2 FG’s and a crucial xp to get past South Carolina by 1 point! I got beat up by Gator fans when I had them ranked out of the Top 10 (#14) in the 2007 preseason poll (lower than everyone else) and the Gators finished 2007 at #13. My computer also did not have Boise St in the Top 10 despite their upset of Oklahoma in the Fiesta Bowl and in 2007 Boise finished unranked.
On the other side, Oklahoma was not in the AP Top 10 at the end of the year but was #8 in my computer’s ranking. In ’07, they won the Big 12 and were #3 at the end of the regular season. My Computer’s #1 team in 2006 was LSU (which was only #3 in the AP poll). How did they fare in 2007? They would “only” go on to win the National Title!
Oregon, Florida and Texas Tech were three 4-loss teams in 2007 that my computer said were Top Ten teams despite their record. How did the “top ten party crashers” (on my computers Top Ten but not in the AP poll) do in 2008? Florida won the national title, Texas Tech was 10-0 and #2 in the country at one point and Oregon finished the year #10 in the final polls.
There were three teams in the AP Top Ten at the end of 2007 that my computer did not have as Top Ten squads. Georgia finished the 2007 season #2 in the final polls but my computer only had them #15 at the end of the year with a 84.7 average game. Only once in 2007 did they even play at over a 100 level and that was a 102.85 effort vs Auburn. They had a 61.05 game in their 35-14 loss to Tennessee and a 78.65 game the next week barely getting past Vanderbilt 20-17 (thanks to a late Vandy fumble). Based on the final record and final game performance they were #2 in the AP in 2007 and #1 in the AP poll in the preseason of 2008. Georgia proved they were overrated in 2008 by finishing #16.
Will those trends continue in 2010? If they do, teams that my computer did not think were Top Ten last year like Boise St, Iowa, Cincinnati and Penn St could disappoint while teams in my computer’s Top Ten like Oklahoma, Oregon, Texas Tech and Nebraska could surprise and this year finish in the Top Ten.
I hope you enjoy the game grade averages as they give you another way of evaluating how well or how poorly your team did last year and they also have proven to be a precursor for the following year.
If you like the game grades averages make sure to check them out on the game-by-game stats for each team in this year’s magazine. At the bottom of the stats page above the year results look for the total game grade average and the ranking below it is my computer’s final ranking.