Swami's seeding prediction

Swami (I'm sure reluctantly) has Cornell #1, but says that strength of schedule will probably knock us down to the #2 seed. Thoughts?

On laxpower a bunch of people were discussing Duke potentially sneaking ahead of us. Apparently head to head is way down their list of criteria, but I still think it would be ridiculous to seed a team that lost to us ahead of us based on SOS.

As I posted a week or so ago, it wouldn't surprise me at all if we aren't seeded #1. The same criteria sent us to Towson two or three years ago and knocked our seeding down from where it should have been last year. Why should this year be different?

There's an interesting thread on Laxpower where someone is proposing changes to the NCAA criteria in an attempt to deemphasize somewhat the effect of strength of schedule. To me it's nuts if one team can go 5-0 against the top 20 (however you choose to measure that, and I don't think it should be RPI) and another team that's 6-3 scores higher because it played four more games against the top 20, even though it lost three of them.

[network.laxpower.com]

I don't know how close Jim Wobus comes to calculating KRACH as we've seen it done for hockey (note the "fictitious tie" he assigns every team) but here's how his numbers come out:

[www.vaporia.com]

This list comes close to most of the polls, and, with the possible exception of Albany being a spot or two too high, is pretty much in line with my view of things.

Probably. Duke has a higher RPI than us, and a higher SoS, and more quality wins (4-3-1-0 vs. 1-1-2-2).

As to Al's point, the criteria does say "record" against top teams, not just wins, so then we're 1-0, 1-0, 2-0, 2-0, and Duke is 4-1, 3-0, 0-1, 1-0. By pure percentage we win, but more subjectively, I'm not sure 7-1 against top 10 isn't more impressive than 2-0.

Anyway, it would be perfectly reasonable, by the criteria, to put as #2, and may give the boys some motivation anyway.

You could object to the system not including H2H as a primary criteria, but, since it doesn't, it would bother me more if they did consider something they're not supposed to.

Duke hasn't even played a stronger schedule than Cornell; LaxPower has them 6th.* And, I'm sorry, any set of selection criteria that might seed a two-loss team ahead of an undefeated (hypothetically, at this point, of course) team that *beat* the two-loss team is stupid. It's not even like they're dealing with a situation like they had in football last fall where Michigan's only loss was to Ohio State, Duke also lost to a decent-but-not-great Loyola team.

* Edit: Yes, I know that's not how the selection criteria evaluate SOS, but still.

Josh '99
Duke hasn't even played a stronger schedule than Cornell; LaxPower has them 6th.* And, I'm sorry, any set of selection criteria that might seed a two-loss team ahead of an undefeated (hypothetically, at this point, of course) team that *beat* the two-loss team is stupid. It's not even like they're dealing with a situation like they had in football last fall where Michigan's only loss was to Ohio State, Duke also lost to a decent-but-not-great Loyola team.

* Edit: Yes, I know that's not how the selection criteria evaluate SOS, but still.

It seems to me there are two separate issues here. First is if the selection criteria should be changed. On that you'll have no problem convincing me they should be revised, not to mention made more objective.

The second is where Cornell should be given the current set of criteria. On this, your use of the LaxPower SoS is irrelevant. That's based on their Power Rankings, which are also irrelevant. They are as meaningful to the NCAA as KRACH or pure win % or alphabetical.

Unfortunately, we don't have any top-10 SoS I consider to be fully accurate. But the closest (since Hymie hasn't fixed his bug) is probably LaxPower: [www.laxpower.com]

The story is a little different there, huh?


There's a final issue, which is that lax doesn't even have a seeding criteria. Only a 'selection'. There is nothing to say what's suppose to be used to seed teams. I guess if they wanted to do it alphabetically, they could.

DeltaOne81
As to Al's point, the criteria does say "record" against top teams, not just wins, so then we're 1-0, 1-0, 2-0, 2-0, and Duke is 4-1, 3-0, 0-1, 1-0. By pure percentage we win, but more subjectively, I'm not sure 7-1 against top 10 isn't more impressive than 2-0.

I suspect most would say 7-1 vs. top 10 is more impressive than 2-0. But perhaps 6-0 against top 20 is better than 8-2. And remember who the "1" is in that 7-1 record. Also, how are we determining the top 10? By the--in my opinion--very flawed RPI formula?

It isn't clear to me why head-to-head should be ignored when the relative ranking of two teams that are otherwise close is being decided. For example, Drexel's win over Virginia wouldn't win any argument with me to seed Drexel ahead of Virginia--given both teams' complete body of work--but Virginia's win over Hopkins might sway me if it was a toss-up otherwise as to which of those two should be seeded above the other for, say, seeds #3 and #4.

DeltaOne81
There's a final issue, which is that lax doesn't even have a seeding criteria. Only a 'selection'. There is nothing to say what's suppose to be used to seed teams. I guess if they wanted to do it alphabetically, they could.

If lax doesn't have a seeding criteria, then we're back to Josh's point. We're the only undefeated team at this point (pending the Hobart game, of course). In fact every other team has at least 2 losses. It would be ridiculous for the committee to not seed Cornell #1 (again, pending tomorrow).

Of course it might be better to go into the tournament the #2 seed with a chip on their shoulders than the #1 with a target on their backs.

schoaff
Of course it might be better to go into the tournament the #2 seed with a chip on their shoulders than the #1 with a target on their backs.

Exactly. And with the bottom 8 teams not being seeded, we'll probably get Sienna regardless of whether we're #1 or #2.

KeithK

DeltaOne81
There's a final issue, which is that lax doesn't even have a seeding criteria. Only a 'selection'. There is nothing to say what's suppose to be used to seed teams. I guess if they wanted to do it alphabetically, they could.

If lax doesn't have a seeding criteria, then we're back to Josh's point. We're the only undefeated team at this point (pending the Hobart game, of course). In fact every other team has at least 2 losses. It would be ridiculous for the committee to not seed Cornell #1 (again, pending tomorrow).

Not only that, but I keep coming back to our H2H win on their field, which has to mean something in the grand scheme of things; to hold otherwise is simply illogical.

Interesting tidbit: Duke has played ten of their fifteen games at home and only four on the road (plus one at a neutral site in San Diego).

scoop85
Not only that, but I keep coming back to our H2H win on their field, which has to mean something in the grand scheme of things; to hold otherwise is simply illogical.

It's the NCAA. You want logic?

In contrast, after tomorrow we will have played six at home, six on the road, and one at a neutral site. Unfortunately, in its infinite wisdom the NCAA does not give extra weight to road wins. Margin of victory is another no-no, of course. We wouldn't want to encourage running up the score, as it could cause permanent psychological harm.

According to the RPI ranking on Laxpower, we're slightly higher in RPI than Duke. I don't know if that's the actual RPI used, but it does imply that we could still wind up #1 in RPI.

But as noted, this is the NC$$. Logic comes second.

What's interesting is that many pundits have the t-4 as:
CU
Duke
UVa
JHU

While the NCAA criteria seems to indicate:
Duke
CU
JHU
UVa

Sets up the same semifinal matchups and Cornell would still likely see Sienna in round one. I'd just rather not have the path to a title be through 2(maybe 3!) ACC teams. This potential path seems reasonably favorable as a 2 seed: Sienna--->7.Albany/Princeton--->3.JHU---->1.Duke/4.UVa.

I like the Red jerseys better anyway

Of course when Sunday comes and we're staring at Cornell as #6 hosting Navy or some such, I'll be typically pissed at the committee. I remember ONE year where they did right by us. I think it was '02 when the committee out and said on the (web only!) selection show "we relied heavily on straight RPI" and that helped CU considerably. Though maybe that was the women's team especially since it was a weird webcast... I forget.

Jeff Hopkins '82
According to the RPI ranking on Laxpower, we're slightly higher in RPI than Duke. I don't know if that's the actual RPI used, but it does imply that we could still wind up #1 in RPI.

Its not:
[web1.ncaa.org]

The difference is very small and hence the details of the calculations make the difference. Here is a site that explains their formula (its 25/50/25 just like LaxPower) and gets the same results as the NCAA:
[www.vaporia.com]

We're gnashing our teeth and worrying about all the ways the NCAA seeding committee can screw over Cornell. Perhaps doubly so since the NCAA didn't have a chance to stick it to the hockey team this year, so lax gets a double. Take a look at the Duke lacrosse record. First, weep for how good Duke's season schedule page is. It lists not only the score but also the ranking of the team each week and the ranking of its opponent going into the game. As for the record:

Duke has two losses.
One of the losses to Cornell.
One win by one goal (in OT).
Three wins by two goals.
One blowout win by 10 or more.

I'm not saying these should be the seeding criteria, but they are ways to look at a team's progress during the season.

This is from the Hobart write-up about yesterday's game:

At 13-0, Cornell is all but guaranteed the top seed in the NCAA Tournament

[laxpower.com]

ebilmes
This is from the Hobart write-up about yesterday's game:

At 13-0, Cornell is all but guaranteed the top seed in the NCAA Tournament

[laxpower.com]

I don't think the Hobart SID has a vote.