pafan wrote:CBB_Fan wrote:It also helps that Wichita State has been moving up the other statistical rankings. They are up to #6 in KenPom, #1 in Nolan's Power Index, #1 in Sagarin's Elo Ratings, and of course #2 in the polls.
Sagarin ELO is based heavily on winning. i.e. the team with the most wins will be #1, and that is Wichita. I doubt anyone seriously looks at ELO as a guage for comparing two teams, its at least as bad and probably worse than RPI.
It was used mostly as an example of a high rating rather than as an argument in and of itself. Alternatively, these are the rankings where Wichita State is ranked #1 or #2 (currently #5 in the Massey Composite, which combines 60 different rankings):
#2 Massey (separate from composite)
#2 Team Rankings
#1 Baker-Bradley Terry
#1 Donchess Inference
#1 Wolfe
#1 Krache
#2 Rothman
#2 Wilson
#2 Sports Ratings
#2 RT Power
#1 Sagarin Elo
#1 Snapper's World
#1 Marchpool
#1 Nolan
#2 Tulsa World
#2 D1A Sports
#2 AP Poll
#2 Coaches
Our mean was 6.67, but our SD was huge at 5.56 (most of the Top 10 had a SD of 3 or less). The Massey Composite seems to rank teams based solely on their mean from wide variety of rankings with different criteria instead of relying on one statistical measure or subjective rating. However, the SD puts us nearly with 1 SD of the 1.00, which is only true for Wichita State and Arizona.
What this means is that criteria heavily favor us as a #1 or even the #1 overall, while others barely put us in the top 25 (lowest score was 21). The Elo rating is a perfect example of the former. Whereas rankings that put heavy emphasis on margin of victory and strength of schedule (like the BVL least squares or SRS) are an example of the latter. By those standards, Wichita State's average opponent was almost a perfectly average team and they would have needed to win by around 25 PPG to have the top spot (Arizona at 24.76 SRS).
As the year goes by, the objective measures that rely on efficiency and simple win rate and the subject measures that rely on human analysis have begun to rate us fairly consistently as a top-seed level program. That was the point I was trying to make in my previous post. If only one type of rating or analysis favored us we would have an even higher SD and a lower mean.