10-24-2007, 03:45 PM | #1 |
Senior Member
Join Date: Jan 2006
Posts: 6,177
|
SOS, power ranking, and losses
A concept that is lost on many is that one only has to be better than the best team on its schedule to go undefeated. I can show you mathematically that if you take a team that is probably the #49 team in the countery and play them 12 games against other teams that are between #50 and #60, that the team will have next to zero chance to run the table. And in order for a team to have high probability of going undefeated against such a schedule, they would have likely have to be a top 10 team.
You can crudely correlate SOS to losses as 25 spots in SOS equals one loss. So the following teams would all have equal strength and be pick em games by Vegas on neutral fields. 12-0 SOS 100 11-1 SOS 75 10-2 SOS 50 9-3 SOS 25 8-4 SOS 1 These teams would be ranked somewhere around #10-#12 on an average year. There is also a feeling you can get over on the human polls with crappy SOS and gawdy record like Hawaii. It's true, Hawaii is higher in human polls than computers right now, but these things have a way of working themselves out by season end. Usually your computer ranking will be within a couple spots of your BCS ranking, and the max difference will be about four spots. However, if you're a non-BCS you have more chance of getting screwed by human polls than computers. 2006 Boise State, 2004 Utah, and 2001 BYU are all examples of teams ranked higher in computers than humans. |
Bookmarks |
|
|