2005-2006 PWR - playoffs edition!

[ursusminor]

[quote Ken '70]
UMD's RPI will stay about where it is with another result like friday. Since they're at .500 they are at the exact point where a win or a loss has the highest impact on their own record, which is 25% of RPI.  Losing again tonight drops their win % to .486 which is enough to almost offset adding another instance of MN's record to their Opp record column.
[/quote] Could you explain what you mean by that? When a team is at 1.000, a loss has its greatest impact on its record. When a team is at .000, a win has its greatest impact on its record.


I am reading all of these PWR threads so that I will understand it in the one year in the distant future when it will make a difference to RPI. ::rolleyes:: Probably they will have changed the system seven times before that, or I'll be dead ... or both.

[redGrinch]

the short answer.  I'll have others provide the full mathematical analysis

RPI = 25% of the team's record, 50% of opponent's records, 25% of opponents' opponent's records.

Opponent's records exclude your games aginst them - similarly for OO records.
So it's possible that even if you beat a team (e.g. Brown; Northeastern, etc), your RPI will go down because of the weight on opponent's records.

[Ken \'70]

Team A has a 19-1 record, or .9500.  They win their next game to go to 20-1, their win % increases to .9524, an increase of .0024, (25% of which flows to RPI).

Team B has a 10-10 record, or .5000.  They win their next game to go to 11-10. Their new winning % is .5238. an increase of .238, about 100 times larger  increase in winning % than team A.

The delta as to the absolute effect of an incremental win or loss is greater as you get closer to .5000, from either direction.  The greatest single gain, or loss, a team can experience is at it moves from .5000.

Therefore the further you are from .5000 the greater the effect of your next opponent's winning % on your RPI relative to the gain/loss you get from winning/losing the game.

[Ken \'70]

Sorry, the above is true only as you move away from .5000, not toward it. UMD will be moving away from .5000 tonight, hopefully south. Apologize for the confusion. ::nut::

[ursusminor]

[quote Ken '70]Sorry, the above is true only as you move away from .5000, not toward it. UMD will be moving away from .5000 tonight, hopefully south. Apologize for the confusion. ::nut::[/quote] Exactly. In your previous message, ".238" should be ".0238", thus about 10 times bigger than .0024, not 100, but it doesn't change your point.

[Ack]

[quote Ken '70]
The delta as to the absolute effect of an incremental win or loss is greater as you get closer to .5000, from either direction.  The greatest single gain, or loss, a team can experience is at it moves from .5000.

Therefore the further you are from .5000 the greater the effect of your next opponent's winning % on your RPI relative to the gain/loss you get from winning/losing the game.[/quote]

If you have a perfect record, a loss will drop your RPI further than a loss would if you were at 0.5000 prior to the loss.  A win then helps you not at all, other than adding to a buffer if you were to eventually lose (where the RPI drop would be less severe).

As we consider teams with RPI of 0.5000 more than teams with a Brown-like RPI as the "bottom rung," a win certainly does help the 0.5000 team more than the 1.0000 in terms of RPI change, but a loss hurts the 1.0000 team (Minnesota) more than the 0.5000 team (Clarkson).

[DeltaOne81]

[quote billhoward]Or is it like democracy ... worst form of government except everything tried before?[/quote]

Wow, Bill. I think one of your phrases finally hit the nail on the head. PWR is far from perfect, but if you compare it to what football, basketball, or lacrosse do, you should be very greatful for it.

I suppose you're a fan of subjective decisions you could like basketball, but everyone's entitled to be wrong . Football is just messed up, and lacrosse is ust dreadful (love the sport, hate the selection process).

[jy3]

in other sports, cornell is 217 in rpi in mens hoops

und won
uno is winning

[Brad Crispell]

Niagara is very close to becoming a TUC - showing as .4989 on Hockey@SiouxSports.com - with the .003 bonus for its away quality win. I'm not sure what win they have constitutes quality (Alabama-Huntsville?)so perhaps this website isn't correct.  Anyway I believe they are getting close.

It would seem that the Big Red's pairwise rating would jump significantly if Niagara makes it to the "upper level".  Can someone decipher if the Purple Eagles can become a TUC and the impact that will have to Cornell's PWR?

[Al DeFlorio]

[quote Brad Crispell]Can someone decipher if the Purple Eagles can become a TUC and the impact that will have to Cornell's PWR?[/quote]
They could assure TUC status by winning the CHA tournament.

[jtwcornell91]

[quote Brad Crispell]Niagara is very close to becoming a TUC - showing as .4989 on Hockey@SiouxSports.com - with the .003 bonus for its away quality win. I'm not sure what win they have constitutes quality[/quote]

December 30, 2005: Niagara 6, New Hampshire 2. http://www.collegehockeystats.net/0506/boxes/mniaunh1.d30

[DeltaOne81]

[quote Brad Crispell]I'm not sure what win they have constitutes quality (Alabama-Huntsville?)/quote]

You can't get a quality win in a conference game.

[jy3]

or a nonconference game against a conference foe

[Dpperk29]

still 1-0 SLU after 2

[jy3]

1-1 third period, 2:54 in