Bongle, great work!
Pretty much repeating what everyone else has suggested, but with a methodology attached in case they're unclear
Can you load all the matches into one matrix and solve OPR/RPI from there?
By the way, before the all regional option got in there, I dumped them one by one and threw them into a spreadsheet, then did some other analysis by hand.
http://www.chiefdelphi.com/forums/sh...t=76194&page=7
Interesting statistical points about the top 25 OPR:
Only 12 went on from setting that high OPR to win that regional.
Also, team 71's 63.55 OPR at the Midwest regional stands out beyond the rest, but their performance (which probably did involve quite a lot of scoring) didn't seem to add up to that. This is just an easy example of the inaccuracy possible in this analysis.
__________________
Team 694
2005 Championship - Galileo Semifinalist
2005 New York - Regional Chairmans Award
2005 New York - Semifinalist (Thanks 1257,1340)