(05-09-2016 09:56 AM)DownEastPirate Wrote: (05-09-2016 09:08 AM)pesik Wrote: his power rankings are out
http://www.sbnation.com/college-football...s-rankings
So his blinding amount of stats and calculations say Houston loses two conference games, USF goes undefeated and Navy is #1 in the conference even though they lose two games and USF is 8-0. Then his power rankings say Houston is #1 in the conference, in front of his project undefeated team USF, and 4 spots above his #1 team Navy who sits in 5th. Makes sense.
The fact there is 24 pages arguing about his projections is sort of sad.
1. His projections have USF winning 7.9 games... certainly not undefeated in conference. Even if a team is a favorite in every conference game.. if they are only a slight favorite in multiple, you expect losses to occur.
2. Where was Navy #1 in the conference? Navy was the last team previewed based on being number 1 based on his numbers from last season. His numbers project Navy to be the 4th best team in the AAC (4th best does not necessarily mean a team ends up in 4th, particularly in a conference like this where a number of teams are grouped very closely).
3. His power rankings of the conference are highly influenced by his numbers, but he also uses his personal interpretation and tries to adjust for where he thinks his numbers may be off. For instance with Houston he says the injury to Ward artificially lowered Houston's ranking some. His numbers have Houston projected as the 2nd best team in the league, he thinks they'll be the best.
So basically you made these criticism starting from completely incorrect information and demonstrated a lack of understanding of his evaluation. Not surprising, but kind of hilarious none the less.
No computer system is perfect for evaluating anything (though I much prefer them to the human's who don't watch every game and are heavily influenced by agenda's, overly influenced by overall record without regard to difficulty of games, and subject to personal bias). You can have a rational, reasonable discussion on where you disagree with computer rankings. What's ridiculous with so much of this thread is the number of people basically arguing with positions that the author doesn't have (or his numbers don't show) or misrepresenting what his position actually is (or what his numbers actually say).
I don't know which computer systems are the best. His do very well on a week to week basis as a projections system. What I really like about his numbers is the way they break down each aspect of the game so you can have really good discussions about a team and how it plays. systems like Sagarin and others may do that some but the data is not readily available and right in front of you on the same way. It's the same thing I like about Kenpom in basketball. That doesn't mean his numbers are always correct and don't miss. Seasons aren't long...teams can be different teams at different times during the season, there are many ways where numbers can be off, but I find his numbers to be a great starting point for rational discuss. Of course that means people need to actually get what his numbers are saying... and maybe I overestimate people's ability to do that.
