Owp/oowp

[norm1909]

Not exactly new, but I missed it the first time.

Based on a consensus that “OWP” should be weighted more than “opponents’ opponents’ winning percentage,” the committee approved a two-thirds/one-third balance in weighting of those criteria in the sports of baseball, men’s and women’s basketball, field hockey, men’s and women’s ice hockey, men’s and women’s lacrosse, men’s and women’s soccer, softball, and women’s volleyball.

[Josh Carey]

Re: Owp/oowp

The "but we don't know how anything is weighted!" crowd is going to have trouble with this one...

[NUProf]

Re: Owp/oowp

The "but we don't know how anything is weighted!" crowd is going to have trouble with this one...

This is truly peculiar, when you recognize that for DI, the RPI components are weighted as WP at 25%, OWP at 21%, and OOWP at 54%. In other words, DI weights OOWP more than twice as much as it does OWP. DIII wants to do the reverse of this. Why?

[norm1909]

Re: Owp/oowp

More info here.

Official Notice Number: ADM--2011--2* (Tentative until formally adopted)

[NUProf]

Re: Owp/oowp

The current selection criteria regarding opponents' average winning percentage (OWP) and opponents' opponents' average winning percentage (OOWP) does not sufficiently recognize the scheduling variances that exist. For example, each institution has greater control over determining the opponents it plays rather than determining its opponents' opponents. Accordingly, after receiving input from the NCAA statistics staff, the Championships Committee determined it was appropriate to give more weight to the portion of the strength of schedule it has more control over in baseball, basketball, field hockey, ice hockey, lacrosse, soccer, softball and volleyball. Additionally, the location of a contest, whether home or away, can have a greater impact on the outcome of that contest; therefore, this proposal includes a means to recognize this difference. Consequently, on a sport-by-sport basis in the aforementioned sports, the committee will have the discretion, based on a recommendation by the relevant sport committee, to attach different significance to home and away contests when calculating the strength of schedule. Finally, this proposal distinguishes the use of strength of schedule criteria in other sports such as tennis and golf that do not presently use OWP or OOWP in their respective strength of schedule analysis.

This is nonsense. When you schedule a game against a team you know the strength of the league that they play in, so you do have the ability to consider OOWP when crafting a schedule. In fact this process encourages teams to schedule teams like Marian and other middle tier teams in the MCHA, NE and MASCAC, rather than middle teams in the interlock, NCHA, or MIAC. There's also an allowance for somehow weighting home and away games differently in calculating OWP and OOWP. The big question is will adjustments to those criteria be public knowledge or not? I hope there will be a clear statement as to how that adjustment will be made before the season starts.

[joecct]

Re: Owp/oowp

This is nonsense. When you schedule a game against a team you know the strength of the league that they play in, so you do have the ability to consider OOWP when crafting a schedule. In fact this process encourages teams to schedule teams like Marian and other middle tier teams in the MCHA, NE and MASCAC, rather than middle teams in the interlock, NCHA, or MIAC. There's also an allowance for somehow weighting home and away games differently in calculating OWP and OOWP. The big question is will adjustments to those criteria be public knowledge or not? I hope there will be a clear statement as to how that adjustment will be made before the season starts.

I think you hit the nail on the head. But look at it this way.

Going into a hypotheitcal game between Salem St and Norwich, Salem is 9-0 and Norwich is 9-0. Salem's opponents have a winning percentage of .500 and Norwich's have a win % of .667. If Norwich wins, they get a benefit of 2/3(.900) + 1/3 (.500) or .767. Salem's benefit is 2/3(1.000) + 1/3(.667) or .889.

Moral: Play the top teams of the weaker conferences and the middle teams of the strong conferences. Of couse if you go 25-0, it should make no difference as you will make the NCAA tournament, right??

[NUProf]

Re: Owp/oowp

I think you hit the nail on the head. But look at it this way.

Going into a hypotheitcal game between Salem St and Norwich, Salem is 9-0 and Norwich is 9-0. Salem's opponents have a winning percentage of .500 and Norwich's have a win % of .667. If Norwich wins, they get a benefit of 2/3(.900) + 1/3 (.500) or .767. Salem's benefit is 2/3(1.000) + 1/3(.667) or .889.

Moral: Play the top teams of the weaker conferences and the middle teams of the strong conferences. Of couse if you go 25-0, it should make no difference as you will make the NCAA tournament, right??

Just one minor correction to note. Your effect on your own OWP doesn't count. An opposing team's record for OWP purposes is their record in games against teams that aren't you. Thus after Norwich tramples Salem the effect is that a game is added to their resume with a benefit of (2/3)(1.000) + (1/3)(0.500) which is 0.833. Your calculation for Salem is correct.Thus, for risking an almost certain loss by playing a top team the gain for Salem is (.056) which averaged out over the course of a 25 game season is virtually nothing.

[Matthew Webb]

Re: Owp/oowp

The "but we don't know how anything is weighted!" crowd is going to have trouble with this one...

Not really. Sure we will know how OOWP and OWP will be weighted to form one metric, but it still does nothing to let us know how the committee weighs everything else relative to combined OWP. Real answer is of course they don't know themselves and it changes year to year as personnel changes, but simply submitting to that unfortunate reality is certainly no fun.

No worries. Still plenty of smoke. And mirrors. And Dasani.

[Dyce]

Re: Owp/oowp

Not really. Sure we will know how OOWP and OWP will be weighted to form one metric, but it still does nothing to let us know how the committee weighs everything else relative to combined OWP. Real answer is of course they don't know themselves and it changes year to year as personnel changes, but simply submitting to that unfortunate reality is certainly no fun.

No worries. Still plenty of smoke. And mirrors. And Dasani.

The cycle continues, with the irresistable power of a mighty Mobius strip...

[NUProf]

Re: Owp/oowp

Not really. Sure we will know how OOWP and OWP will be weighted to form one metric, but it still does nothing to let us know how the committee weighs everything else relative to combined OWP. Real answer is of course they don't know themselves and it changes year to year as personnel changes, but simply submitting to that unfortunate reality is certainly no fun.

No worries. Still plenty of smoke. And mirrors. And Dasani.

Furthermore, it's up to the committee to determine how this mystery home vs. away game component is factored into OWP and OOWP. It's not that I don't trust the committee's ability to make decisions, it's just that I feel that the whole process should be as transparent as possible, so that for example - Manhattanville would have known the committee's weights and known that they were going to be out no matter what happened in the ECAC W championship game - it might have made that victory a little sweeter for them instead of being left with a bitter taste in their mouth at the end of the season. They could have run the numbers and concluded that this is their championship game - just like Geneseo could celebrate their win over Oswego as a quasi-championship game. It shouldn't be a mystery.

It's kind of like if the NFL tie breakers were a secret and the committee met and applied them after the season was over to determine who got the Wild Card bids for the playoffs. How would that play?