Subject: | |
From: | |
Reply To: | |
Date: | Mon, 6 Mar 1995 15:17:51 EST |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Below are listed the YAM2 rankings as of the end of last weekend
(3/6/95). The explanation of the YAM2 method at the end of the ranking
list has been expanded somewhat for those folks who have had questions
regarding the influence of the ranking factors.
YAM2 RPI Norm. YAM2
Rank Rank W-L% Sched. Metric
1 3 Maine 0.811 0.949 0.769
2 2 Michigan 0.803 0.954 0.766
3 1 Boston_University 0.779 0.973 0.758
4 4 New_Hampshire 0.706 0.970 0.684
5 5 Colorado_College 0.724 0.933 0.675
6 7 Clarkson 0.677 0.926 0.627
7 9 Michigan_State 0.662 0.928 0.614
8 12 Bowling_Green 0.677 0.892 0.604
9 10 Denver 0.639 0.941 0.601
10 6 Minnesota 0.597 1.000 0.597
11 8 Wisconsin 0.583 0.977 0.570
12 13 Vermont 0.594 0.923 0.548
13 14 Lake_Superior 0.588 0.922 0.542
14 11 Northeastern 0.544 0.976 0.531
15 15 Brown 0.574 0.921 0.529
16 20 Colgate 0.565 0.891 0.503
17 18 Princeton 0.537 0.924 0.496
18 21 Miami 0.528 0.912 0.481
19 19 Harvard 0.518 0.921 0.477
20 26 RPI 0.532 0.894 0.476
21 16 North_Dakota 0.486 0.969 0.471
22 17 St_Cloud 0.486 0.964 0.469
23 23 Minnesota-Duluth 0.500 0.922 0.461
24 24 Western_Michigan 0.486 0.930 0.452
25 27 Mass_Lowell 0.487 0.916 0.446
26 29 Merrimack 0.470 0.910 0.428
27 22 Michigan_Tech 0.444 0.961 0.427
28 25 St_Lawrence 0.431 0.963 0.415
29 28 Providence 0.441 0.930 0.410
30 31 Cornell 0.440 0.902 0.397
31 32 Ferris_State 0.412 0.907 0.374
32 33 Dartmouth 0.370 0.931 0.345
33 30 Boston_College 0.353 0.957 0.338
34 35 Illinois-Chicago 0.371 0.905 0.336
35T 34 Northern_Michigan 0.347 0.939 0.326
35T 38 Union 0.370 0.879 0.326
37 37 Yale 0.339 0.910 0.309
38 36 Alaska-Anchorage 0.324 0.924 0.299
39 39 Notre_Dame 0.288 0.900 0.259
40 40 Alaska-Fairbanks 0.268 0.906 0.243
41 43 Air_Force 0.262 0.801 0.210
42 41 Ohio_State 0.182 0.913 0.166
43 42 Mass_Amherst 0.172 0.909 0.156
44 44 Army 0.091 0.749 0.068
The YAM2 is an intuitively based simple formula which seeks to
measure accomplishment over the course of the season. It will not
identify the teams which are currently hot, but views the season as
a whole.
YAM2 differs from the Rating Percentage Index primarily in its
relationship between Win% and Strength of Schedule. In YAM2 this is a
multiplicative relationship, whereas in the RPI it is additive.
YAM2 = (Win%) x (Strength of Sched.)
Strength of schedule is quantified the same way as in RPI: namely
2 parts Opp% added to 1 part Opp-Opp%. In this implementation the strength
of schedule is normalized to the value of the stongest schedule (Minnesota
this week). Also, as in the RPI, the head to head games are subtracted
from the records before calculating the Opp% in order to prevent "inverse"
effects on the ranking metric.
YAM2 gives equal *mathematical* weight to Win% and Strength of Schedule.
However, since Strength of Schedule inherently varies less than Win%, SOS will
produce less effect than Win% upon the ranking placements (approximately
five positions maximum from experience).
As a property of the multiplicative relationship between Win%
and Strength of Schedule, YAM2 will not produce "inverse" ranking
effects. For instance, it will not raise a ranking if a team goes
to MN and loses two games (except in *very* unusual circumstances:-).
-- Dick Tuthill
|
|
|