A season recap, by the numbers (long, as per usual).
- Thread starter daj2576
- Start date
General Jack
Vorschlaghammer
- Joined
- Sep 28, 2004
- Messages
- 13,213
- Likes
- 4,947
governmentmule
as always Go Lady Vols :clapping:
- Joined
- Aug 27, 2007
- Messages
- 28,234
- Likes
- 5,076
Here is something else to chew on. Using the pre-season talent average rankings of all FBS teams, 13 of the top 25 ended up in the projected BCS final top 25. If you look at the final BCS projections, only 7 (28%) were not in the pre-season top 40 of talent rankings.
Of the top 10 in the final projected BCS rankings only 3 teams had a significant net positive performance against the talent on their schedule (Stanford +4, Baylor +3, Mizzou +5). Three teams had a net effect of 0 (meaning in aggregate they won and lost all of the games they were supposed to (FSU, Auburn and Clemson). In total of the top 10 projected BCS teams, 7 performed within one game of talent predictions.
Using my adjustment (discussed above) the top 4 BCS teams should be, in order 1) Bama, 2) FSU, 3)Auburn, and 4)LSU. That means that when adjusting for starting talent, and moving teams for aggregate wins or losses against the talent they compete against, those are the four teams ranked in order of the mathematical likelihood that the higher ranked team would beat the lower ranked team over a series of games, and not simply a one game exception.
Of note, BCS championship games have about a 90% prediction rate using talent averages as a guide. This year will likely pit two teams that are the most closely related in talent using my available data (back to 2005). Talent averages suggest that FSU (6.75) beats Auburn (7.25). The only team that has upset a more talented team was Texas over USC in 2005.
To put a fine point on this: A team must recruit in the top 15 to make it to a championship, and then the higher ranked team will win; if a team recruits in the top 25, there is greater than a 50% chance they'll end up in the top 25; if a team recruits outside of the top 40, there is less than an 8% chance they can perform well enough to end with a top 25 ranking.
Of the top 10 in the final projected BCS rankings only 3 teams had a significant net positive performance against the talent on their schedule (Stanford +4, Baylor +3, Mizzou +5). Three teams had a net effect of 0 (meaning in aggregate they won and lost all of the games they were supposed to (FSU, Auburn and Clemson). In total of the top 10 projected BCS teams, 7 performed within one game of talent predictions.
Using my adjustment (discussed above) the top 4 BCS teams should be, in order 1) Bama, 2) FSU, 3)Auburn, and 4)LSU. That means that when adjusting for starting talent, and moving teams for aggregate wins or losses against the talent they compete against, those are the four teams ranked in order of the mathematical likelihood that the higher ranked team would beat the lower ranked team over a series of games, and not simply a one game exception.
Of note, BCS championship games have about a 90% prediction rate using talent averages as a guide. This year will likely pit two teams that are the most closely related in talent using my available data (back to 2005). Talent averages suggest that FSU (6.75) beats Auburn (7.25). The only team that has upset a more talented team was Texas over USC in 2005.
To put a fine point on this: A team must recruit in the top 15 to make it to a championship, and then the higher ranked team will win; if a team recruits in the top 25, there is greater than a 50% chance they'll end up in the top 25; if a team recruits outside of the top 40, there is less than an 8% chance they can perform well enough to end with a top 25 ranking.
Last edited:
BennisDrewer
Senior Lurker
- Joined
- Jan 1, 2013
- Messages
- 1,447
- Likes
- 780
GiveHimSIX!!!
Must be Nice GUY
- Joined
- Sep 3, 2008
- Messages
- 944
- Likes
- 2
VN Store
