3-on-3 overtime and the RPI & PWR

[cbuckser]

Last week Brandon Thomas explained the new ECAC and NCAA rules for regular-season overtime. In short, for intraconference games, the ECAC standings award three points for a regulation win, two for an overtime or shootout win, one for an overtime or shootout loss, and zero for a regulation loss. That's what the IIHF and a lot of European professional leagues do. I think it's a far more rational and superior system to the loser-point system the NHL uses. The Ratings Percentage Index component of the Pairwise Rankings weights the results differently: 100% for a regulation win, 55% for an overtime win, 50% for a shootout win or loss, 45% for an overtime loss, and 0% for a regulation loss. So, after the first two games, according to the RPI, Cornell's winning percentage is .550, not .667 or 1.000.

It will take me a long time to get acclimated to the new system. I don't really know how to process a 55% victory. I'm curious to hear how the rest of you perceive the results.

[redice]

Last week Brandon Thomas explained the new ECAC and NCAA rules for regular-season overtime. In short, for intraconference games, the ECAC standings award three points for a regulation win, two for an overtime or shootout win, one for an overtime or shootout loss, and zero for a regulation loss. That's what the IIHF and a lot of European professional leagues do. I think it's a far more rational and superior system to the loser-point system the NHL uses. The Ratings Percentage Index component of the Pairwise Rankings weights the results differently: 100% for a regulation win, 55% for an overtime win, 50% for a shootout win or loss, 45% for an overtime loss, and 0% for a regulation loss. So, after the first two games, according to the RPI, Cornell's winning percentage is .550, not .667 or 1.000.

It will take me a long time to get acclimated to the new system. I don't really know how to process a 55% victory. I'm curious to hear how the rest of you perceive the results.

My assessment:  pretty damned stupid!!

[jkahn]

For pairwise, we are actually under .500.  Home wins are multiplied by a .8 factor, and home losses by 1.2.
So our 2-0 record gets factored down to (1.1 x . wins and (.9 x 1.2) losses, totalling .88 wins and 1.08 losses.
The committee in charge obviously thinks 3 vs. 3 overtime is fairly meaningless weighting it only .55 vs. .45, why not go back to 5 vs. 5.

[Swampy]

For pairwise, we are actually under .500.  Home wins are multiplied by a .8 factor, and home losses by 1.2.
So our 2-0 record gets factored down to (1.1 x . wins and (.9 x 1.2) losses, totalling .88 wins and 1.08 losses.
The committee in charge obviously thinks 3 vs. 3 overtime is fairly meaningless weighting it only .55 vs. .45, why not go back to 5 vs. 5.

I understand the logic: we should win at home and not need overtimes. But the weights seem arbitrary and perverse. I'd rather see 1 as a weight for home wins and a bonus, say 1.2, for away wins; similarly, 1 as a weight for away losses, and say 1.2 for home losses. Maybe use 1.1 or 1.15 instead. The idea is to give full credit for the games a team should win (home) and extra credit for the games a team should not (away), and assume the team should always win at home.

I also don't understand the second part of your explanation.

    We get .55 x 2 for the 2 OT wins, and a .2 penalty for doing it at home: 1.1 x .8. Fine.

    But what's the bit about 1.08 losses? If we won both games, where do the losses come from?

[jtwcornell91]

In the name of eliminating ties, the NCAA has actually caused more games to be counted more or less as ties.

[marty]

In the name of eliminating ties, the NCAA has actually caused more games to be counted more or less as ties.

This!  Thank you.

[Al DeFlorio]

In the name of eliminating ties, the NCAA has actually caused more games to be counted more or less as ties.

This!  Thank you.

This happens when you lose sight of the forest.

The only games where you can justify extreme measures to ensure a winner are tournament games.  In regular season games, a tie, after playing real hockey for a reasonable period of overtime, is perfectly acceptable.

[nshapiro]

For pairwise, we are actually under .500.  Home wins are multiplied by a .8 factor, and home losses by 1.2.
So our 2-0 record gets factored down to (1.1 x . wins and (.9 x 1.2) losses, totalling .88 wins and 1.08 losses.
The committee in charge obviously thinks 3 vs. 3 overtime is fairly meaningless weighting it only .55 vs. .45, why not go back to 5 vs. 5.

1. I agree completely with the 'why not go back to 5-on-5.  Just add the shootout if you have to.

2. https://www.collegehockeynews.com/ratings/pairwise/ shows our weighted win percentage of .449.  Why isn't it 0.49 ... (.88 + 1.08)/4?

[marty]

In the name of eliminating ties, the NCAA has actually caused more games to be counted more or less as ties.

This!  Thank you.

This happens when you lose sight of the forest.

The only games where you can justify extreme measures to ensure a winner are tournament games.  In regular season games, a tie, after playing real hockey for a reasonable period of overtime, is perfectly acceptable.

But what I think John is saying is that we won both nights this weekend and yet for RPI we got .55 of a win times .8 for being home so our two wins are essentially two ties for post season ranking.  I agree that 3x3 for 5 minutes isn't real hockey but with this system in place I ask why have overtime at all?  

In the ECAC the points will be more reasonable.  But for non-conference games I guess the fans get to see which team is the lion and which team is sacrificed. For post season purposes in NC games there is (almost) no reason to have OT.

[upprdeck]

lost in the end of the article is the comment on how little this team has practiced the last 2 yrs..  we have done almost no 5x3 stuff and who knows how many other things this group has to get up to speed on.  it will be a work in progress for awhile.

[Al DeFlorio]

In the name of eliminating ties, the NCAA has actually caused more games to be counted more or less as ties.

This!  Thank you.

This happens when you lose sight of the forest.

The only games where you can justify extreme measures to ensure a winner are tournament games.  In regular season games, a tie, after playing real hockey for a reasonable period of overtime, is perfectly acceptable.

For post season purposes in NC games there is (almost) no reason to have OT.

That's what I said...although I have no objection to a 6-on-6 sudden death overtime of defined duration for a regular season game, within conference or not. Going to two decimal places to quantify the results of a game is simply absurd.

[upprdeck]

on the flip side since hockey is also a fan sport.. Talked with kids at multiple schools the last few days and they love the 3x3 and shootout.  Most fans do and there are far more casual fans than hardcore.  if you remove the PWR issues it is fun to watch for many people.

[jtwcornell91]

In the name of eliminating ties, the NCAA has actually caused more games to be counted more or less as ties.

This!  Thank you.

This happens when you lose sight of the forest.

The only games where you can justify extreme measures to ensure a winner are tournament games.  In regular season games, a tie, after playing real hockey for a reasonable period of overtime, is perfectly acceptable.

For post season purposes in NC games there is (almost) no reason to have OT.

That's what I said...although I have no objection to a 6-on-6 sudden death overtime of defined duration for a regular season game, within conference or not. Going to two decimal places to quantify the results of a game is simply absurd.

I think Al and I are, as a colleague of mine likes to say, in violent agreement.  The fact that playoff games fall back to 5x5 OT with no shootout shows that all of these gimmicks are recognized as less than "real hockey".  That said, if they're insistent upon using 3x3 and shootouts for regular season games, they should be consistent and use one point system across the board.

[billhoward]

on the flip side since hockey is also a fan sport.. Talked with kids at multiple schools the last few days and they love the 3x3 and shootout.  Most fans do and there are far more casual fans than hardcore.  if you remove the PWR issues it is fun to watch for many people.

Exactly! Many fans love 3x3s and shootouts. I recall an SI piece about the then-new NHL shootout. The story talked about how players, in the dressing room after a game, or out at a bar, and it came time for a shootout, their attention was riveted. 3x3 is similar because of the higher odds there will be a score, soon. I'm fine with 3x3 college hockey OT; it is a little disconcerting that a shootout win at home counts so little.

[Dafatone]

on the flip side since hockey is also a fan sport.. Talked with kids at multiple schools the last few days and they love the 3x3 and shootout.  Most fans do and there are far more casual fans than hardcore.  if you remove the PWR issues it is fun to watch for many people.

Exactly! Many fans love 3x3s and shootouts. I recall an SI piece about the then-new NHL shootout. The story talked about how players, in the dressing room after a game, or out at a bar, and it came time for a shootout, their attention was riveted. 3x3 is similar because of the higher odds there will be a score, soon. I'm fine with 3x3 college hockey OT; it is a little disconcerting that a shootout win at home counts so little.

They're fun. I'm just a little iffy on how much, if at all, they should count.

Part of me says go back to 5x5 OT, then do an exhibition shootout purely for bragging rights, kinda like they do in the mid-season tournaments.