Division I Rutter Computer Ranking Top Ten for Monday, November 5th

[LakersFan]

Rutter Division I Women's Ice Hockey Computer Ranking Top Ten

Through Games of Sunday, November 5th

Rating is based on wins, loses, and ties only. Margin of victory and game location are not included in the rating. A team's rating from the previous year is used as an initial weight using a Bayesian approach.


Rank Team Rating Previous Record
1 New Hampshire 1.8334 1 7-1-0
2 St. Lawrence 1.5507 4 6-2-0
3 Connecticut 1.5441 6 9-1-0
4 Mercyhurst 1.5366 3 8-1-1
5 Harvard 1.5312 8 2-0-0
6 Wisconsin 1.4063 2 9-2-1
7 Minnesota Duluth 1.3944 5 7-2-1
8 Minnesota 1.2183 7 6-3-1
9 Dartmouth 1.0266 NR 3-0-1
10 Ohio State 0.8871 10 5-3-2


Dropped from top ten: Boston College

For more information, visit http://math.bd.psu.edu/faculty/rutter/WomensRankings.html

Copyright 2007 Michael A. Rutter

This work is licensed under the Creative Commons Attribution-Share Alike 3.0 License. To view a copy of this license, visit http://creativecommons.org/licenses/by-sa/3.0/us/; or, (b) send a letter to Creative Commons, 171 2nd Street, Suite 300, San Francisco, California, 94105, USA.

[RStarr]

WOW. Look at UConn....!!

Wildcat's on top and Wisco slides downtown Julie Brown.

[dave1381]

I understand the logic of the rankings. Essentially, what's happening is, Minnesota had a slow start and lost to a bunch of weaker teams, and then in turn improved better and beat Wisconsin and UMD. Thus, all the WCHA teams get pulled down to the level of other teams in the conference -- leaving UConn, SLU, and UNH -- who have only lost to each other -- in the top three.

It's a fair, objective system, but I don't agree with it. I think what's actually happened is Minnesota played at much higher level in recent weeks, and the ranking system -- which calculates a performance parameter for a performance function that's normally distributed -- simply doesn't model this all that well. I imagine you'll see those three WCHA teams all in the top five still.

[hab]

The value of this system is that it takes a lot of subjectivity out of the rankings. And the biggest drawback is that it takes a lot of subjectivity out of the rankings:D . One thing that it doesn't take into account, I think, is the margin of victory. So a team like UConn that has put together a number of close wins (including 2 overtime wins this past weekend against unranked teams) comes out looking a lot better than where I would (subjectively) rank them.

[D2D]

The value of this system is that it takes a lot of subjectivity out of the rankings. And the biggest drawback is that it takes a lot of subjectivity out of the rankings:D . One thing that it doesn't take into account, I think, is the margin of victory. So a team like UConn that has put together a number of close wins (including 2 overtime wins this past weekend against unranked teams) comes out looking a lot better than where I would (subjectively) rank them.

Also how the most recent games should count more than earlier games. What kind of weighting does this poll give more recent wins vs. older losses? Come February/March, it's not important how you were playing back in October and early November.

[RStarr]

Also how the most recent games should count more than earlier games. What kind of weighting does this poll give more recent wins vs. older losses? Come February/March, it's not important how you were playing back in October and early November.
I thought it was cumulative through the whole season, meaning every game counts...no matter when it was played. :confused:

[gotice]

The value of this system is that it takes a lot of subjectivity out of the rankings. And the biggest drawback is that it takes a lot of subjectivity out of the rankings:D . One thing that it doesn't take into account, I think, is the margin of victory. So a team like UConn that has put together a number of close wins (including 2 overtime wins this past weekend against unranked teams) comes out looking a lot better than where I would (subjectively) rank them.

There is a good ranking system used to rank boys Midget and DI mens college(among other leagues) that does just what you are talking about. They factor in strength of competition and actually have a number assigned to strength of schedule. So a team that may have more loses could be ranked higher than a team with more wins, based on the strength of the competition. They also have a 6 goal max to keep things even, so a team does not move up simply because they score 11 goals against weak competition.

DI college rankings (http://www.ushsho.com/d1mrank.htm)

midget major rankings (http://www.ushsho.com/midgetrankings.htm)

[LakersFan]

To clarify:

My ranking system does not take into account margin of victory. This is very hard to do for hockey, IMO. A one goal lead turns into a two goal lead, or a two into a three with an empty-net goal. So if you beat a time by four or more goals, is that a big win? Or should it be five? Can a team completely outplay another, yet only win 1-0 because a goalie stands on her head? I have decided to take these type of discussions out of the equation by not using MOV. (Bad math pun, sorry)

My ranking system does not take into account when the game was played. Again, I question how to incorporate this in a fair manor. UMD's win over Merychurst is big, and should not be diminished because it took place in the first month. On the other hand, I am sure Minn. would like to down weight its early games.

If anybody would like to add these ideas to my system, I would be happy to share the code I use to create them with you. The more ranking systems, the better, I say. We have to have something to talk about between games anyway.

Personally, I am still trying to deal with the fact that Dave does not "agree" with my system ;)

[RStarr]

To clarify:

My ranking system does not take into account margin of victory. This is very hard to do for hockey, IMO. A one goal lead turns into a two goal lead, or a two into a three with an empty-net goal. So if you beat a time by four or more goals, is that a big win? Or should it be five? Can a team completely outplay another, yet only win 1-0 because a goalie stands on her head? I have decided to take these type of discussions out of the equation by not using MOV. (Bad math pun, sorry)

My ranking system does not take into account when the game was played. Again, I question how to incorporate this in a fair manor. UMD's win over Merychurst is big, and should not be diminished because it took place in the first month. On the other hand, I am sure Minn. would like to down weight its early games.

If anybody would like to add these ideas to my system, I would be happy to share the code I use to create them with you. The more ranking systems, the better, I say. We have to have something to talk about between games anyway.

Personally, I am still trying to deal with the fact that Dave does not "agree" with my system ;)
I like the way it is, that is, margin of victory not counting and the date of the game played not having mattered.

[dave1381]

Personally, I am still trying to deal with the fact that Dave does not "agree" with my system ;)
Ha... yeah I should have used better language there. I think you get what I was saying. If it's not clear to everyone, what I was trying to say is I have a fairly good understanding of how the system works, but this week's rankings differed from my personal ranking of teams more so than ever before. However, it's impossible to fully "agree" with any system since you need simplifying assumptions to make it tractable, and those assumptions will never always match reality -- thus, my comment was total nonsense.

Of course I still value the rankings as they are, and very much appreciate the effort. I think they become more useful as more games are played because (1) the increased sample size improves the quality of the estimation (2) it becomes more complex for humans to rank teams as games played increases, while the computer works just as well.