LakersFan
03-15-2004, 09:45 PM
One more week to go, and here are your current computer rankings:
http://vortex.bd.psu.edu/~rutter/WomensHockey/WomensDIHockeyRankings.html
Mease Top Ten:
1. Minnesota
2. Harvard
3. Dartmouth
4. Wisconsin
5. St. Lawrence
6. Minnesota-Duluth
7. Mercyhurst
8. New Hampshire
9. Princeton
10. Brown
The top three seem pretty well set for the frozen four. Here are the remaining three candidates for the final slot:
Wisconsin: 4 Mease, 4 RKRACH, 6 RPI, 5 PWR
St. Lawrence 5 Mease, 6 RKRACH, 4 RPI, 4 PWR
UMD 6 Mease, 5 RKRACH, 5 RPI, 6 PWR
Based on these rankings, none of the four methods are consistent. The mathematical/statistical based systems favor Wisconsin, while the committee systems favor St. Lawrence. In the PWR, St. Lawrence benefits from record against teams under consideration and common opponents, two criteria Mease and RKRACH ignore. Mease (and RKRACH) is heavily influenced by results against teams with high Mease scores (and low Mease scores, if you lose to them). This is a slightly different criteria then teams with records above .500. Mease in a sense uses common opponents, but doesn't weigh those reuslts more or less then results against non-common opponents. RKRACH gives hope to UMD, ranking them above St. Lawrence. I haven't explored the reason (i.e. which games) cause the slight differences in Mease and KRACH. That sounds like a summer project.
http://vortex.bd.psu.edu/~rutter/WomensHockey/WomensDIHockeyRankings.html
Mease Top Ten:
1. Minnesota
2. Harvard
3. Dartmouth
4. Wisconsin
5. St. Lawrence
6. Minnesota-Duluth
7. Mercyhurst
8. New Hampshire
9. Princeton
10. Brown
The top three seem pretty well set for the frozen four. Here are the remaining three candidates for the final slot:
Wisconsin: 4 Mease, 4 RKRACH, 6 RPI, 5 PWR
St. Lawrence 5 Mease, 6 RKRACH, 4 RPI, 4 PWR
UMD 6 Mease, 5 RKRACH, 5 RPI, 6 PWR
Based on these rankings, none of the four methods are consistent. The mathematical/statistical based systems favor Wisconsin, while the committee systems favor St. Lawrence. In the PWR, St. Lawrence benefits from record against teams under consideration and common opponents, two criteria Mease and RKRACH ignore. Mease (and RKRACH) is heavily influenced by results against teams with high Mease scores (and low Mease scores, if you lose to them). This is a slightly different criteria then teams with records above .500. Mease in a sense uses common opponents, but doesn't weigh those reuslts more or less then results against non-common opponents. RKRACH gives hope to UMD, ranking them above St. Lawrence. I haven't explored the reason (i.e. which games) cause the slight differences in Mease and KRACH. That sounds like a summer project.