What's new
By:
USCHO Fan Forum

This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

  • The USCHO Fan Forum has migrated to a new plaform, xenForo. Most of the function of the forum should work in familiar ways. Please note that you can switch between light and dark modes by clicking on the gear icon in the upper right of the main menu bar. We are hoping that this new platform will prove to be faster and more reliable. Please feel free to explore its features.

2009-2010 DI Rutter Computer Rankings Thread

Re: 2009-2010 DI Rutter Computer Rankings Thread

LakersFan said:
For game played through October 25, 2009...
Click to expand...
I think that teams like Harvard, Dartmouth, and Minnesota State have discovered a good way to climb the rankings ladder -- don't play. Let others expose their flaws, while any that you may have remain hidden.

BU is a enigma to me. The computer and the pollsters seem to like them. But if they are truly a top 10 team, shouldn't they find somebody that they can beat in regulation once in a while?:confused:
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

LakersFan said:
Wisconsin is 15th in my rankings, 19th in RPI.
Click to expand...
How is Vermont so far ahead of Clarkson in RPI? Given that VMU split with Princeton (a team with the 28th best RPI) and swept Union (one that is 30th), I don't see how they got so far ahead of a team with good results whom they split H2H. I don't require an answer, I'm just thinking out loud about how the computer sees things different than my biased mind.:o

EDIT: Clarkson has the 4th-best RPI at 0.6059, while Vermont is 3rd with 0.6559.
 
Last edited:
Re: 2009-2010 DI Rutter Computer Rankings Thread

ARM said:
How is Vermont so far ahead of Clarkson in RPI? Given that VMU split with Princeton (a team with the 28th best RPI) and swept Union (one that is 30th), I don't see how they got so far ahead of a team with good results whom they split H2H. I don't require an answer, I'm just thinking out loud about how the computer sees things different than my biased mind.:o

EDIT: Clarkson has the 4th-best RPI at 0.6059, while Vermont is 3rd with 0.6559.
Click to expand...

RPI doesn't mean squat until enough games have been played by all teams. That's one of the reasons the Rutter rankings are better.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

Sorry. Here are the true updated ratings. My server lost its internet connection sometime over the weekend, so I missed some games. Now through October 25, 2009.

Code: 
   	Team  	Rating  	
1	Minnesota	2.5297	 
2	Mercyhurst	1.8325	
3	Clarkson	0.8904	
4	UMD		0.8238	
5	New Hampshire	0.7920	
6	Harvard		0.6739	
7	Dartmouth	0.5568	
8	Minnesota State	0.4076	
9	St. Lawrence	0.3636	
10	Boston Univ.	0.2927
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

ARM,

The reason why Vermont has the 5th best RPI (again, sorry for not noticing the lack of internet on the server) is because the NCAA has strange weights for opponent's winning percentage. Opponent's opponent's winning percentage counts almost twice as much (46%) as does opponent's winning percentage (24%).

And now Minnesota State has fallen a little bit. The worst thing is not to play and have all the teams you have play lose.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

LakersFan said:
The reason why Vermont has the 5th best RPI (again, sorry for not noticing the lack of internet on the server) is because the NCAA has strange weights for opponent's winning percentage.
Click to expand...
That makes more sense. I realize that the weightings play tricks but I was having trouble reconciling how they could be ahead of Clarkson, and by quite a bit. The update makes sense.

Thanks for the fix and for all that you do to maintain the rankings (which are really all we have to go by until the PWR debuts in January.):)
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

I'm a huge Vermont fan (hell, I may be the ONLY one, on this board anyway:) , but even I take the rankings with a grain of salt. I know Vermont is not one of the top ten teams in the nation (at least not YET), but I think what the rankings show is that the Catamounts have played well enough in the early season that they are, for the first time, getting some recognition nationally. The fact that they are even in the discussion on this thread shows how far they've come. And for that, I'm proud, and excited to see where they go from here.
 
  • Like
Reactions: D2D
Re: 2009-2010 DI Rutter Computer Rankings Thread

For games played through 11/1/09

Code: 
   	Team  	Rating  	
1	Mercyhurst	1.6053	
2	Minnesota	1.5928	
3	Clarkson	1.0433	
4	Wisconsin	1.0181	
5	New Hampshire	0.9498	
6	UMD		0.9171	
7	Northeastern	0.5830	
8	Ohio State	0.5788	
9	Boston Univ.	0.5498
10	St. Lawrence	0.4577

At this point, we have clusters. Hurst and UM are tied at 1 and 1a. 3-6 are lumped together, then 7-9. 13-22 are separated by less than .25 rating points. We need more games to create more separation.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

How is there so much separation between Mercyhurst/Minny and Clarkson? Clarkson has defeated UNH, SLU, BU, and BC, all of which were ranked in the top 10 at the time they beat them.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

Clarkson lost to Vermont. That brings them down a bit.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

FireKnight said:
How is there so much separation between Mercyhurst/Minny and Clarkson? Clarkson has defeated UNH, SLU, BU, and BC, all of which were ranked in the top 10 at the time they beat them.
Click to expand...

In terms of my model, where the team was ranked when they game occured has no effect. All games take place at the same time, in terms of the model. Even if Clarkson didn't play a game the rest of the season, their ranking would change based on the quality of the teams they did beat. They BC win is not impressive (right now) and the loss to Vermont does hurt. Hurst has one loss to UMD (6), Clarkson has a loss to Vermont (22) and a tie with BC (18). That is your seperation.

How much separation is there? If Mercyhurst played Clarkson today (assuming no ties possible), the model predicts the Lakers would win 71% of the time. We will test that prediction Jan 15/16.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

Interesting theory I guess. However, there are so many variables that all games being played without respect to time really makes for a useless number.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

FireKnight said:
Interesting theory I guess. However, there are so many variables that all games being played without respect to time really makes for a useless number.
Click to expand...

It is possible to construct a model that weights more recent games more heavily. Teams with big non-conference wins early in the season would dislike that model, while teams with big conference wins late in the season would be in favour of it. The NCAA's RPI does not include date of game and I do the same.

As soon as you start playing the game "team A beat team B back when team B's goaltender was on a hot streak and playing really well," you start crossing the line from well designed algorithm to an opinion poll. It may be possible to create such an "expert system" that takes into account hot players, injuries, international tournaments, etc. It would require a tremendous amount of data collection as well as data points in terms of results. I don't think college hockey teams play enough games to tease out anything more than home ice advantage or effect of back to back games.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

LakersFan said:
The NCAA's RPI does not include date of game and I do the same.

As soon as you start playing the game "team A beat team B back when team B's goaltender was on a hot streak and playing really well," you start crossing the line from well designed algorithm to an opinion poll.
Click to expand...
And we all know how great a system opinion polls are. For reference, see the BCS rankings. :p
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

LakersFan said:
In terms of my model, where the team was ranked when they game occured has no effect. All games take place at the same time, in terms of the model. Even if Clarkson didn't play a game the rest of the season, their ranking would change based on the quality of the teams they did beat. They BC win is not impressive (right now) and the loss to Vermont does hurt. Hurst has one loss to UMD (6), Clarkson has a loss to Vermont (22) and a tie with BC (18). That is your seperation.

How much separation is there? If Mercyhurst played Clarkson today (assuming no ties possible), the model predicts the Lakers would win 71% of the time. We will test that prediction Jan 15/16.
Click to expand...

I love it. We get a a free Statistical analysis lessons while following the college game we love. :cool:
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

OnMAA said:
I love it. We get a a free Statistical analysis lessons while following the college game we love. :cool:
Click to expand...

And some say that the math you learn in school is never used in real life.:D
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

Rutter Rankings for games played through November 8th

Code: 
	Team		Rating
1	Mercyhurst	1.6955	 
2	Minnesota	1.4723	
3	Clarkson	1.1740	
4	New Hampshire	0.9099	
5	UMD		0.7872	
6	Wisconsin	0.7468	
7	Cornell		0.7232	
8	Northeastern	0.4711	
9	Harvard		0.4412	
10	Ohio State	0.3368

I have added a cool feature (I think) to the web site, available only through this link at this point (not ready to add to main page).

http://math.bd.psu.edu/faculty/rutter/D1_history/season_history.html

As shown in the picture below, you can see the ratings as a function of time over the season (I posted a picture before), but now you can click on a team and see their rating highlighted over time. On slower connections, it may take some time to load each image. Not recommended for those on iPhones and similar devices. If it is not working for you, please let me know. This is a D-I feature only (at least for this season).

Screenshot.png
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

Rutter Rankings for games played through November 15, 2009

Please note the web server that hosts my page was down most of the weekend and the site was not updating. The latest rankings are now up.

http://math.bd.psu.edu/faculty/rutter/WomensRankings.html

Code: 
   	Team  	Rating  	
1	Mercyhurst	1.7110	 
2	Minnesota	1.2610
3	Clarkson	0.9942	
4	New Hampshire	0.9102
5	UMD		0.9001	
6	Cornell		0.6664	
7	Wisconsin	0.6389		
8	Harvard		0.3102		
9	Boston Univ.	0.2460		
10	Northeastern	0.2272
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

Rutter Rankings for games played through November 22, 2009

Code: 
   	Team  	Rating  	
1	Mercyhurst	1.4374
2	Minnesota	1.3267
3	Clarkson	1.0078	
4	New Hampshire	0.9489	
5	UMD		0.7619	
6	Cornell		0.6443
7	Wisconsin	0.5482	
8	Harvard		0.5125	
9	Boston Univ.	0.3052	
10	Northeastern	0.2779

If you check out the season long graphical representation of the rankings (http://math.bd.psu.edu/faculty/rutter/D1_history/season_history.html), we are starting to see the top 8 separate from the pack. Not sure how long that will last.
 
Re: 2009-2010 DI Rutter Computer Rankings Thread

Rutter Rankings for games played through November 29, 2009

Code: 
   	Team  	Rating  	
1	Minnesota	1.5037	
2	Mercyhurst	1.4965	
3	New Hampshire	0.9980	
4	UMD		0.9077	
5	Clarkson	0.8206	
6	Harvard		0.5873	
7	Wisconsin	0.5554	
8	Cornell		0.5462	
9	Northeastern	0.3001	
10	Niagara		0.2142

Virtual tie for first. I think it is safe to say we have 1 and 1a in Hurst and Minn. Is Cornell for real? Will they maintain a top 8 position and make the NCAAs?
 
You must log in or register to reply here.
Back
Top