Quote:
Originally Posted by Mikhail's Fortunes
How is it that not one person suggested 0 wins from a team that went 3-9 last year, yet someone suggested 12 wins?
oops... sorry for the trouble in language... no one said 12 wins... I should've mentioned the methodology
-I gather preseason expert polls that rank teams ordinally 1 to 120
-I weight these polls based on previous accuracy to come up with one final ordinal poll
-through magic, this poll is turned into a power ranking---well, not magic, more like the a weighted average of previous computer system rankings for those ordinal spots
-this power ranking can be used with team specific home field advantage to project spreads for each game a team plays
-that spread is turned into odds of winning a game based on a weighted historical database of every college game spread and result since 1993
-from there, it's just adding up the wins and losses
-this year, there is also an inclusion of the likely distribution of forecasting errors among the original weighted preseason power poll based on back-testing of the data
none of this would stand up in a rigorous scientific journal, but it is good enough IMO