What's new
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.

John t whelan ranking simulator

Re: John t whelan ranking simulator

Does autobids have any impact on that? Alabama Huntsville made the tournament a few years ago with a sub-500 record, but they had the CHA autobid. Also, I seem to remember about 6 or 7 years ago UVM had a team that almost didn't make the Hockey East playoffs but ended making some noise in the national tournament.
No, it doesn't affect autobids, just at-large bids. Wisconsin earned an at-large bid in 07-08 with a record of 15-16-7. They did end up beating Denver 6-2 in their first NCAA tournament game, but lost to North Dakota in overtime the next game. The "Wisconsin rule" was implemented the year after.
 
Re: John t whelan ranking simulator

No, it doesn't affect autobids, just at-large bids. Wisconsin earned an at-large bid in 07-08 with a record of 15-16-7. They did end up beating Denver 6-2 in their first NCAA tournament game, but lost to North Dakota in overtime the next game. The "Wisconsin rule" was implemented the year after.

If that's still the case, New Hampshire would be the replacement. It does show that it's very early in the season, though.
 
Re: John t whelan ranking simulator

exp(ki) is the KRACH rating, ie, ki for team i is a number which needs to exponentiated to be in Whelan's form. You don't need to exponentiate again. h, the home ice advantage, also needs to be exponentiated to make patman's formula pr(home win) = ca/(ca+b) correct. It's generally easier to estimate ki instead of Ki=exp(ki) because the constraint that Ki>0 is a pain. when working with ki, ie ln(Ki) the coefficient is unconstrained.

Note all of the Ks are still only determined up to a multiplicative constant, so you can define the absolute value of the KRACH ratings any way you want. Some common methods are to make the best team 100, to make the sum 1, or, in an understandable if somewhat idiosyncratic view at siouxsports.com, to make North Dakota = 100.

(Neither h nor theta can be scaled. They have to be takes as is.)
 
Last edited:
Re: John t whelan ranking simulator

exp(ki) is the KRACH rating, ie, ki for team i is a number which needs to exponentiated to be in Whelan's form. You don't need to exponentiate again. h, the home ice advantage, also needs to be exponentiated to make patman's formula pr(home win) = ca/(ca+b) correct. It's generally easier to estimate ki instead of Ki=exp(ki) because the constraint that Ki>0 is a pain. when working with ki, ie ln(Ki) the coefficient is unconstrained.

Note all of the Ks are still only determined up to a multiplicative constant, so you can define the absolute value of the KRACH ratings any way you want. Some common methods are to make the best team 100, to make the sum 1, or, in an understandable if somewhat idiosyncratic view at siouxsports.com, to make North Dakota = 100.

(Neither h nor theta can be scaled. They have to be takes as is.)

Thanks Go,

That's what I meant. I appreciate all the explaining.
 
Re: John t whelan ranking simulator

As of games completed 22 November 2013, 11:59 Aleutian Time:

5.00 Quinnipiac
4.75 Minnesota
4.50 Michigan
4.25 Ferris State
4.00 St. Cloud State
3.75 Boston College
3.50 LSSU
3.25 Providence
3.00 Yale
2.75 Cornell
2.50 Wisconsin
2.25 Bowling Green
2.00 Notre Dame
1.75 New Hampshire
1.50 Clarkson
1.25 UMASS Lowell
1.00 Miami
0.75 Northeastern
0.50 Northern Michigan
0.25 Union

And the tournament field:

Minnesota
Quinnipiac
Michigan
Ferris State

Boston College
St. Cloud State
LSSU
Providence

Wisconsin
Yale
New Hampshire
Cornell

Notre Dame
Bowling Green
UMASS Lowell
AHA Champ (36 - Mercyhurst)
 
Re: John t whelan ranking simulator

As of games completed 23 November 2013, 11:59 Aleutian Time:

5.00 Minnesota
4.75 Michigan
4.50 Ferris State
4.25 Providence
4.00 St. Cloud State
3.75 Quinnipiac
3.50 Boston College
3.25 Cornell
3.00 LSSU
2.75 Minnesota State Mankato
2.50 Clarkson
2.25 UMASS Lowell
2.00 Yale
1.75 Bowling Green
1.50 Wisconsin
1.25 Notre Dame
1.00 New Hampshire
0.75 Miami
0.50 Union
0.25 Northeastern

And the tournament field:

Minnesota
Michigan
Providence
Ferris State

Quinnipiac
St. Cloud State
Boston College
Cornell

LSSU
UMASS Lowell
Clarkson
Bowling Green

New Hampshire
Wisconsin
Yale
AHA Champ (37 - Air Force)

NOTE: Minnesota State Mankato was removed from 10th overall due to their sub-500 record.
 
Re: John t whelan ranking simulator

The killer of a loss against Princeton loses 1.25 points. Hopefully a win against Providence Wednesday makes that difference up.
 
Re: John t whelan ranking simulator

The killer of a loss against Princeton loses 1.25 points. Hopefully a win against Providence Wednesday makes that difference up.

That's just Quality Wins Bonus, not your RatingsPI. Your RatingsPI is currently at 59.60 with removals of Bentley, both Holy Cross games, Dartmouth, and the Princeton game from Friday. Beating Providence alone would take you to 60.58, but that may change based upon other happenings today and Tuesday.
 
That's just Quality Wins Bonus, not your RatingsPI. Your RatingsPI is currently at 59.60 with removals of Bentley, both Holy Cross games, Dartmouth, and the Princeton game from Friday. Beating Providence alone would take you to 60.58, but that may change based upon other happenings today and Tuesday.

Still it is a horrible loss either way. Losing quality wins bonus points hurts especially when the loss is to a bottom feeder team.
 
Re: John t whelan ranking simulator

Still it is a horrible loss either way. Losing quality wins bonus points hurts especially when the loss is to a bottom feeder team.

Actually, depending on who beats or ties you, it's actually not all that bad. What I posted is not how many you received, but the rate at which other teams may receive if they play and beat or tie you. Thus far, you have 0.62 Quality Wins Bonus points, and these are from games vs. UMASS Lowell, Cornell, and Yale. Remember that a tie is half a win and half a loss, so you still get some points in there.
 
Re: John t whelan ranking simulator

Still it is a horrible loss either way. Losing quality wins bonus points hurts especially when the loss is to a bottom feeder team.

And, the horrible-ness of the loss is compounded by how much your own RatingsPI suffers from it. Own win% goes down a lot from 12-1-1 to 12-2-1, and Princeton not being a great power makes the effect on RatingsPI even worse.

If I understand this QWB part right, Quinn actually lost 0 QWB points from not beating Princeton. The real problem is as above - lose to a bottom team and your RatingsPI takes a huge hit.
 
Re: John t whelan ranking simulator

And, the horrible-ness of the loss is compounded by how much your own RatingsPI suffers from it. Own win% goes down a lot from 12-1-1 to 12-2-1, and Princeton not being a great power makes the effect on RatingsPI even worse.

If I understand this QWB part right, Quinn actually lost 0 QWB points from not beating Princeton. The real problem is as above - lose to a bottom team and your RatingsPI takes a huge hit.

They lost some QWB points, because remember that the factor I post is multiplied by the weighted win rate, and then total of those is divided by the team's potential number of weighted win% points, but that's going to happen whenever you play a non-quality team. Not to mention, UMASS Lowell took a hit, and that dropped the QWB.
 
Re: John t whelan ranking simulator

They lost some QWB points, because remember that the factor I post is multiplied by the weighted win rate, and then total of those is divided by the team's potential number of weighted win% points, but that's going to happen whenever you play a non-quality team. Not to mention, UMASS Lowell took a hit, and that dropped the QWB.

Understood, FlagDude. I was trying to point out the concept that losing to a bottom team hurts because your RatingsPI takes a huge hit on your own part.

Lost QWB points due to the number of games played is just a fact of scheduling (yes, I know that defeating a very bottom team in totally neutral, because the game essentially doesn't count).

Lost QWB points because some other team you defeated loses is also a fact of life. In some cases, a team you defeated earlier might sweep the #1 team, and then move up, giving you more QWB points that you didn't earn yourself.

Lost Ratings PI points because you lost to a team with a low win% is on yourself, for the most part.
 
Re: John t whelan ranking simulator

These are the types of things that are going to drive fans mad.

Wisconsin doesn't play.
Drops from 8 to 16 in raw RPI.
BU beats and ties North Dakota.
Falls from #18 to #34 because they beat Wisconsin.
 
Re: John t whelan ranking simulator

Might as well make a Sunday games update:

5.00 Minnesota
4.75 Michigan
4.50 Ferris State
4.25 St. Cloud State
4.00 Providence
3.75 Quinnipiac
3.50 Cornell
3.25 Boston College
3.00 LSSU
2.75 Clarkson
2.50 Minnesota Duluth
2.25 Notre Dame
2.00 Minnesota State Mankato
1.75 UMASS Lowell
1.50 Yale
1.25 Wisconsin
1.00 Bowling Green
0.75 New Hampshire
0.50 Miami
0.25 Union

And the tournament field:

Minnesota
Michigan
Providence
Ferris State

St. Cloud State
Quinnipiac
Boston College
Cornell

LSSU
Minnesota Duluth
UMASS Lowell
Notre Dame

Clarkson
Yale
Wisconsin
AHA Champ (37 - Air Force)
 
Re: John t whelan ranking simulator

These are the types of things that are going to drive fans mad.

Wisconsin doesn't play.
Drops from 8 to 16 in raw RPI.
BU beats and ties North Dakota.
Falls from #18 to #34 because they beat Wisconsin.

You mean 28 to 34? There's no way they were at 18.
 
Back
Top