Thanks to Erik for the following information. I was interested in doing
an analysis of what would happen in RPICH given different results involving
closely-matched teams in the rating, who are playing in Hockeyfest the
next few days. Remember that other results from the West and the championship
games, etc. from the East could also have an effect.
Here's what happens in six scenarios. I didn't study the West because
there are still many games left, with the best-of-three series being played
this weekend and the final fours coming next weekend.
These scenarios involve 1) PC & UNH; 2) SLU/Clarkson/Harvard.
Note: the other teams all remain the same as they currently are in RPICH.
Rankings of teams around the ones considered may change (i.e. below, Wis
from 8th to 9th) but their actual rating changes insignificantly.
PC-UNH: current ranking has PC 9th (.5378), UNH 10th (.5375). What happens
in the two scenarios involving their semifinal game against each other?
1)PC beats UNH Opp OppOpp
G W L T Win% Win% Win% RPICH
8 Providence 36 22-12- 2 .6389 .5205 .5152 .5421
9 Wisconsin 34 21-11- 2 .6471 .5176 .5124 .5414
10 New Hampshire 35 21-12- 2 .6286 .5234 .4991 .5347
----------
2)UNH beats PC Opp OppOpp
G W L T Win% Win% Win% RPICH
8 Wisconsin 34 21-11- 2 .6471 .5176 .5121 .5413
9 New Hampshire 35 22-11- 2 .6571 .5234 .4996 .5406
10 Providence 36 21-13- 2 .6111 .5205 .5147 .5363
----------
The winner appears to have a clear advantage here, as we would expect.
PC's 3-0-0 regular season record could play a part. (BU remains #6 in
all scenarios and had a 2-2-0 record vs UNH, 1-1-1 vs PC.)
SLU-Clarkson-Harvard: current ranking has SLU 12th (.5270), WMU 13th (.5240),
Harvard 14th (.5227), and Clarkson 15th (.5200). SLU plays RPI in one
ECAC semifinal and Clarkson plays Cornell in the other. Harvard has no
more games before bids are handed out. (As explained, WMU isn't being
studied but will be included here for continuity.) What happens in the
four possible scenarios involving these two ECAC semifinal matchups?
1)SLU, Clarkson both win Opp OppOpp
G W L T Win% Win% Win% RPICH
12 St Lawrence 30 21- 8- 1 .7167 .4726 .4925 .5294
13 Western Michigan 34 16-12- 6 .5588 .5122 .5183 .5240
14 Clarkson 30 21- 8- 1 .7167 .4597 .4909 .5236
15 Harvard 27 14- 7- 6 .6296 .5035 .4874 .5223
----------
2)SLU wins, Clarkson loses Opp OppOpp
G W L T Win% Win% Win% RPICH
12 St Lawrence 30 21- 8- 1 .7167 .4728 .4925 .5295
13 Western Michigan 34 16-12- 6 .5588 .5122 .5182 .5239
14 Harvard 27 14- 7- 6 .6296 .5037 .4875 .5224
15 Clarkson 30 20- 9- 1 .6833 .4597 .4909 .5169
----------
3)Clarkson wins, SLU loses Opp OppOpp
G W L T Win% Win% Win% RPICH
12 Western Michigan 34 16-12- 6 .5588 .5122 .5183 .5240
13 Clarkson 30 21- 8- 1 .7167 .4596 .4908 .5235
14 Harvard 27 14- 7- 6 .6296 .5046 .4873 .5227
15 St Lawrence 30 20- 9- 1 .6833 .4726 .4922 .5226
----------
4)SLU, Clarkson both lose Opp OppOpp
G W L T Win% Win% Win% RPICH
12 Western Michigan 34 16-12- 6 .5588 .5122 .5182 .5239
13 Harvard 27 14- 7- 6 .6296 .5048 .4874 .5228
14 St Lawrence 30 20- 9- 1 .6833 .4728 .4923 .5227
15 Clarkson 30 20- 9- 1 .6833 .4596 .4907 .5168
----------
Whichever one wins appears to be in. Harvard's best chance may be for Clarkson
to lose its semifinal and SLU to win the final. If both SLU and Clarkson
lose their semis, we will be guaranteed an underdog wins the ECAC (and gets
one of the bids), and Harvard then might have a shot at the other bid
assuming only two get picked from the ECAC. If both win their semis, Harvard
appears to be finished.
Here's a table showing how each of the scenarios affected each of the
five teams' (PC, UNH, SLU, Clarkson, Harvard) ratings:
========================================
BEFORE PC W, NH L NH W, PC L |
PC .5378 #9 | +.0043 #8 | -.0015 #10 |
UNH .5375 #10 | -.0028 #10 | +.0031 #9 |
|
SLU .5270 #12 | -.0004 #12 | +.0001 #12 |
Har .5227 #14 | -.0001 #14 | ------ #14 |
Cla .5200 #15 | -.0003 #15 | +.0001 #15 |
------------------------------------------------------------------
BEFORE | SLU W, C W | SLU W, C L | SLU L, C W | SLU L, C L |
PC .5378 #9 | -.0002 #9 | +.0001 #9 | -.0002 #9 | +.0001 #9 |
UNH .5375 #10 | +.0001 #10 | -.0003 #10 | ------ #10 | -.0004 #10 |
|
SLU .5270 #12 | +.0024 #12 | +.0025 #12 | -.0044 #15 | -.0043 #14 |
Har .5227 #14 | -.0004 #15 | -.0003 #14 | ------ #14 | +.0001 #13 |
Cla .5200 #15 | +.0036 #14 | -.0031 #15 | +.0035 #13 | -.0032 #15 |
==================================================================
By the way, if Clarkson loses, then BC at #16 comes within .0015-.0016
of Clarkson (#15). That's before BC's game with Maine.
So what does all of this say to me? 1) the loser of PC-UNH could be in
danger of not getting a bid, especially if it is UNH; 2) SLU/Clarkson
can likely each earn a bid with a semifinal win; 3) Harvard is already in
trouble but it may still have an outside chance if at least one of SLU/
Clarkson drops its semifinal.
All of that for something we already knew. :-) If anyone else interprets
this differently and can add anything more, please do...
BTW, on how the results of PC-UNH affect H/SLU/CU, it looks like:
* PC wants at least Clarkson to lose, preferably both
* UNH wants at least Clarkson to win, preferably both
* SLU & Clarkson want UNH to win
* Harvard wants PC to win (and SLU & Clarkson to lose; likely also BC to
win since Harvard beat BC in the Beanpot)
but again, the effects are not very significant as far as these 1-2 games
go.
Now you ECAC/HE fans attending Hockeyfest have a reason to be interested
in the other league's games :-), if you weren't already.
Disclaimer: since these scenarios were run separately, running them at the
same time may produce slightly different results since more new games are
affecting a team's rating, but the above should give a reasonably close
approximation of how each outcome would affect a rating on its own. And
don't forget the Western games and the HE/ECAC championship games (no
consolations again this year).
- mike
|