Quote:
Originally Posted by transplant99
Just noticed while looking at the BCS standings from top to bottom...with the one exception of Nevada, the whole thing is fewest losses to most losses. I suppose that shouldnt be that much of a surprise, but it seems like that may be the biggest factor in how the computer does things...and not as much as who lost to who, or more correctly who beat who.
http://espn.go.com/college-football/rankings |
Looking at those rankings, I'd say the human polls are as much reason the rankings are basically ordered by record. Look at both polls closely - I can't see a case where there's a team with a final ranking over +/- 2 away from what the human polls say they are.
The problem with the computer rankings is that they stupidly don't include margin of victory. That's not a BCS/playoff sort of thing either (because the computers would still exist in both cases like RPI in basketball) - it's more of a NCAA "sportsmanship" issue. The problem is that margin of victory which essentially encompasses total point differential (points for - points against) is the best way to determine which teams are better and predict future success.
Obviously strength of schedule and other factors have to be considered, but treating a 1 point squeaker and a 40 point blowout against similar teams as if they're the same is outrageous. The humans voting don't treat them the same so why shouldn't the computers?
As I said earlier in this thread, the alleged sportsmanship issue could easily be alleviated in the computers by simply capping the margin of victory to some agreed upon upper limit (say 28 points - 4 touchdowns).
For an idea of how much margin of victory matters in the computers, look at the Saragin Rankings used in the BCS:
http://www.usatoday.com/sports/sagarin/fbt10.htm
There are some HUGE shake-ups - particularly Auburn and LSU each dropping 10+ spots because of their inability to blow weaker teams out (consistently anyways).