PWR Question

[calgARI '07]

For the PWR experts out there:

What are the chances that Cornell can break into the top 14 in the PWR without winning the postseason ECAC?  It seems like they'll have to pretty much win every game for that to happen but you never know with the PWR.

[DeltaOne81]

The problem is, at this stage, if every team has 10 games left, you have 2^10 possibilities of what can happen to each team, so you have something approaching 2^10^58 possibilities of the way the season can play out. I'd suggest you play with JTW's script and see if you can work out some best case scenarios for Cornell and see where it lands us.

[KenP]

Not to be entirely picky...okay yes I am being picky.  The last exponent should reflect at games, not teams.  If every team has 10 games left, the formula should be 2^10^29 possibilities.

Of course even though that removes 87 zeroes from the number, you still have approximately 1,989,290,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000 possibilities.  I'll get started on my analysis right away....

[ursusminor]

Wouldn't it be (3^10)^29 possibilities including ties? (Of course, in some playoff games there can't be ties.)

(3^10)^29 = 3^290 is about 2.3183 x 10^138.

[Trotsky]

[quote calgARI '07]For the PWR experts out there:

What are the chances that Cornell can break into the top 14 in the PWR without winning the postseason ECAC?  It seems like they'll have to pretty much win every game for that to happen but you never know with the PWR.[/quote]

The "feel" I have from watching PWR over the last few seasons is that Cornell can't climb back into an at-large at this point.  The best possible at-large (non-auto) scenario would be a sweep of the final 6 RS and the ECAC QF, and then a split in Albany, for a stretch record of 9-1.  That would give them an overall record of 20-10-3 .652 going into Selection Sunday.

I don't seem to have the final 2006 PWR on file (hat tip, anyone?), but here are the winning percentages and PWR for ECAC teams for the years I do have, 2002-05.

Code Select Expand


[b]Yr Tm Pct PWR[/b]
03 Cor .864 1
05 Cor .833 5
02 Cor .743 9
03 Hvd .697 12
05 Col .697 14
05 Hvd .682 9
--------------
04 Col .628 15
03 Drt .603 19
05 Drt .600 15
02 RPI .595 18
04 Cor .594 16
04 RPI .590 24
03 Yal .563 24
04 Hvd .557 22
04 Drt .544 20
02 Clk .526 22
02 Drt .516 25
02 Hvd .500 24
04 Clk .500 29
05 SLU .474 27



So, between 2002 and 2005, no ECAC team with a winning percentage lower than .697 placed in the top 14 in PWR.  Throw in the likelihood of an upset or two among the majors, and Cornell's chance at an at large looks to be about zero.

[Chris 02]

I've gone to 3 different websites and have seen 3 different versions of the current pairwise.

http://www.siouxsports.com/hockey/rankings/pwr.php has us tied for 17th (18th).

http://www.uscho.com/rankings/pwr.php has us at 16th.

http://www.rpihockey.net/misc.rank1.shtml has us at 19th.

Edit: http://www.collegehockeynews.com/ratings/ncaapwcr.php has at 16th (same as USCHO)

Can anyone explain the discrepancy?  Seems like there must be some computational differences with RIT or perhaps with the bonus.

[Chris '03]

[quote Chris 02]I've gone to 3 different websites and have seen 3 different versions of the current pairwise.

http://www.siouxsports.com/hockey/rankings/pwr.php has us tied for 17th (18th).

http://www.uscho.com/rankings/pwr.php has us at 16th.

http://www.rpihockey.net/misc.rank1.shtml has us at 19th.

Edit: http://www.collegehockeynews.com/ratings/ncaapwcr.php has at 16th (same as USCHO)

Can anyone explain the discrepancy?  Seems like there must be some computational differences with RIT or perhaps with the bonus.[/quote]

Well the differences are bonus and RIT. SiouxSports seems the most in tune with what is likely to happen- .003(maybe 4) bonus for road wins and throwing RIT out of consideration. At this point I believe that takes TUC comparisons out of the equation for Cornell because it leaves the Red with only 9 TUC games. The TUC cliff will really get interesting in the next few weeks depending on who Cornell draws in the playoffs. At this point the best thing to do is look at what comparisons are flippable for cornell to move up. When I looked last night, CC, BC, and UVm all were attainable provided Cornell gets 10 TUC games in the end. There's just a lot of uncertainty at the moment.

[Ken70]

Cornell can get one foot in the door on Friday simply by beating Dartmouth.  That will flip the VT comp. and give Cornell 9 comparisons and alone in 15th place (assume .003 bonus).  If Co. College does anything but win on Friday (while Cornell beats Dartmouth) that may very well flip that comp., putting Cornell in 14th ahead of CC based on the RPI tie-breaker.

Sure - lots of games left after Friday, but it's better to be on the inside looking out than visa versa.