Preliminary Computer Rankings

[NUProf]

It is early, so early that the interlock teams are not included. To make sure the changes that I made in my program worked, I ran my rankings through games of Thursday. Two things to note: teams with less than four games have been deleted from the list. The fact that the ECAC NE/MASCAC/NE10 teams have basically only played each other means that until more "mixing" occurs in the schedule, the rankings of the top teams in those leagues is likely to be inflated.

The rankings this year are on an SAT-like scale - the mean score is 500, and theoretically a team could get a rating as high as 800 or as low as 200. The lowest ranked team btw is Johnson and Wales at 307

With those warnings, here's the "pre-initial" top 10 teams

1	OSW	672
2	SNB	642
3	ADR	626
4	PLT	625
5	ELM	624
6	WEN	619		
7	MAR	610
8	PLY	603
9	FIT	600
10	WRF	598

[NUProf]

Re: Preliminary Computer Rankings

To update - these rankings include only teams that have played at least 4 games. The following factors are included in the calculations:

Winning Percentage
Winning Percentage of the teams you have defeated (0 is used if you haven't beaten anyone)
OWP
OOWP
OOOWP

1	OSW	678
2	SNB	676
3	ELM	625
4	MAR	612
5	PLT	608
6	FIT	607
7	ADR	607
8	WEN	592
9	UMD	578
10	WSR	575

Surprises - Marian and Fitchburg - in the early season, teams which play in leagues that haven't played many non-league games will exhibit very volatile behavior in these rankings, especially with regard to the winning percentage of the teams that you have defeated. I'm still tweeking the weightings of the factors, right now all are weighted equally,

The scale is the familiar SAT type scale (200-800 with a mean of 500)

After Primelink weekend, when everybody has played at least 4 games, we will publish the whole list.

[AGPennypacker]

Re: Preliminary Computer Rankings

Computer rankings? Smells like the BCS. Oh no, someone stop these rankings! We need to go to a playoff system. Oh wait...phew. Got a little excited there for a second.

[joecct]

Re: Preliminary Computer Rankings

Computer rankings? Smells like the BCS. Oh no, someone stop these rankings! We need to go to a playoff system. Oh wait...phew. Got a little excited there for a second.

Don't calm down too much. The arcane, obscure, and foggy procedures of the NCAA D-III men's ice hockey selection committee are legendary.

[norm1909]

Re: Preliminary Computer Rankings

Don't calm down too much. The arcane, obscure, and foggy procedures of the NCAA D-III men's ice hockey selection committee are legendary.

It's not fog - its smoke - and I have reason to believe it might not come from tobacco.

[AGPennypacker]

Re: Preliminary Computer Rankings

Don't calm down too much. The arcane, obscure, and foggy procedures of the NCAA D-III men's ice hockey selection committee are legendary.

Yeah, but if you win your conference that should get you in, or at least a legitimate shot to get in. Oh wait, there's the Adrian exception (insert Boise State reference here, ad nauseum). I know this has been beat up in other threads, but just couldn't help myself.

[NUProf]

Re: Preliminary Computer Rankings

Yeah, but if you win your conference that should get you in, or at least a legitimate shot to get in. Oh wait, there's the Adrian exception (insert Boise State reference here, ad nauseum). I know this has been beat up in other threads, but just couldn't help myself.

This year the Adrian exception will be called the Salem State exception

[joecct]

Re: Preliminary Computer Rankings

This year the Adrian exception will be called the Salem State exception

It all depends how we do OOC. UMD over Babson offset UMB over Framingham St.

I think the MASCAC champion has about a 5-10% chance of making the tournament. But, as long as there is hope, you gotta play the games.

[cmaulding]

Re: Preliminary Computer Rankings

I honestly think that each division should have an automatic qualifier and make the tourney a little bigger. This would end all the questions on teams that get left out each year. And by the way very nice rankings may i ask wht type of code you are useing ?