The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

[granddaddyscout]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

Dave, I want to commend you for your clear and concise analysis. I am not sure how you have the time to delve so deeply into these topics, but I for one appreciate it. Based on what you and others have presented it seems that these disparity issues are somewhat unique to female sports, I am curious if you agree and what you thoughts are on the following as well.

1. Do you think that the depth disparity justifies the use of a different approach and could that argument be supported sufficiently to the NCAA. The WCHA is somewhat unique in that 75% of its teams were ranked in the top 15 most of the year. The other conferences had strong teams at the top and the drop off was significant after the top 1 or 2 teams. This seems unique as compared to the men's programs which seem to have less top to bottom disparity typically.

2. From what you have presented thus far it seems that the current system essentially weights all games the same. Minnesota as an example had a tough start for a variety of reasons, but they are playing at a very high level now. Cornell and BU (I am not bringing up the East vs West debate by mentioning this, just the best example) had strong showings for the first 2/3 of their seasons, but then showed some softening down the road. So while the polls seem to reflect the current strength, the RPI seems to attempt as least to evaluate the entire body of work. For men's sports they seem to select the 8 best teams at the end of the season more consistently. I am not trying to argue that we didn't in this case, simply that the system doesn't seem to be designed to do that.

3. Just out of curiosity, what would have happened if Minny best UW in OT and BU had beaten NU and then defeated BC in their finals.

4. Finally what does the NCAA see as its charter? Are they looking to field the 8 teams that had the best seasons or the top 8 most competitive teams at the end of the season.

Thoughts and/or comments greatly appreciated!

 

[papulaisle]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

Fly Boston to Minneapolis. Fly twenty five-thirty folks.

Is round trip commercial rate, currently. Not booked in advance would be $735 each for March 16. That is Frontier or Air Tran, Continental and United are $1,042. One stop. Total duration five hrs., plus/minus.

Non stop over $1,200.

Travel=$18,375 (25 flying) to $22,050 (30 flying). Likely NCAA can book ahead, perhaps and also get the best, much lower rates, too.

 

[papulaisle]

The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

From granddaddyscout (as above):

" 4. Finally what does the NCAA see as its charter? Are they looking to field the 8 teams that had the best seasons or the top 8 most competitive teams at the end of the season. "

I, too, would like to know what the NCAA's 'charge' is as to the athletes, member schools-all.

I also applaud and appreciate, Dave, your very careful unbiased analysis.

Dartmouth may have juju; from you, sir, I get the story.

 

[dave1381]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

Dave, I want to commend you for your clear and concise analysis. I am not sure how you have the time to delve so deeply into these topics, but I for one appreciate it.

(1) I neglect my dissertation (2) I've been looking at this for a decade and care about it, so it's pretty much second nature for me.

1. Do you think that the depth disparity justifies the use of a different approach and could that argument be supported sufficiently to the NCAA. The WCHA is somewhat unique in that 75% of its teams were ranked in the top 15 most of the year. The other conferences had strong teams at the top and the drop off was significant after the top 1 or 2 teams. This seems unique as compared to the men's programs which seem to have less top to bottom disparity typically.

The RPI is flawed in all sports and should be tossed in all sports that use it, though that's not politically feasible.

The women's hockey WCHA, because it's a successful, geographically isolated conference, suffers more from the RPI than any other sport I can think of. If any sport could justify why they need a more sophisticated system than RPI to the NCAA, it's D-I women's hockey. I would hope if women's hockey successfully adopted a better system, it would eventually spill over to other sports as well.

Just one caveat, the bottom of Hockey East was stronger than the bottom two of the WCHA this year, but teams 2-6 were underrated, but it's the latter that matters for NCAA selections.

Compared to men's, well there were times earlier in the decade when the men's WCHA was very successful and getting seriously hurt by the RPI but I don't believe that's as true any more. But in general, there's more funding for better-quality nonconference games in the men's game, the men's WCHA has another quality CCHA conference that's closer, and usually plays 8 or so nonconference games as opposed to 6 or fewer.

2. From what you have presented thus far it seems that the current system essentially weights all games the same. Minnesota as an example had a tough start for a variety of reasons, but they are playing at a very high level now. Cornell and BU (I am not bringing up the East vs West debate by mentioning this, just the best example) had strong showings for the first 2/3 of their seasons, but then showed some softening down the road. So while the polls seem to reflect the current strength, the RPI seems to attempt as least to evaluate the entire body of work. For men's sports they seem to select the 8 best teams at the end of the season more consistently. I am not trying to argue that we didn't in this case, simply that the system doesn't seem to be designed to do that.

I mean that depends on the criteria. It's not a men's/women's thing. Women's basketball has a more subjective process like men's basketball. Hockey uses a much more objective set of criteria.

Men's and women's hockey used to have a Last 16 criteria but both dropped it and men dropped it first. Problem was mainly that the Last 16 games were often all within conference so it gave an unfair advantage to weaker conferences, and it would've been difficult to fairly adjust that for strength of schedule.

The hockey criteria also doesn't consider victory margins at all, while the basketballs can (they though primarily use an RPI which doesn't consider it)

If I had to pick things to fix about the system, I don't think punishing poor late-season results is necessarily at top priority. I'd rather start by getting the overall ranking on the whole body of work right in the first place.

3. Just out of curiosity, what would have happened if Minny best UW in OT and BU had beaten NU and then defeated BC in their finals.

Off the top of my head, I believe Minnesota would've clearly had home ice as No. 4 and BC would've ended up 7th.

The rankings would've been (1) Wisconsin, (2) Cornell, (3) BU, (4) Minnesota, (5) Mercyhurst, (6) UMD, (7) BC, (8) Dartmouth.

The NCAA could have then chosen from a three-flight bracket
Dartmouth @ Wisconsin
BC @ Cornell
UMD @ BU
Mercyhurst @ Minnesota

or a two-flight bracket
Dartmouth @ Wisconsin
BC @ Cornell
Mercyhurst @ BU
UMD @ Minnesota

If a point you want to make is that Wisconsin was punished for losing the WCHA final and not letting Minnesota to get home ice, I agree. I started another thread making that point.

I can't be 100% sure the NCAA would choose the three-flight bracket. The NCAA did choose a three-flight over a two-flight bracket in 2009 which allowed 3 WCHA teams to reach the Frozen Four.

This year, it's clear to me the NCAA choose to pair Dartmouth with Cornell because it was the only way to avoid a four-flight bracket. It's not clear to me whether they paired UMD with Wisconsin purely to avoid a three-flight bracket, or whether because they believed UMD was the worst team available to pair with Wisconsin -- my guess is the latter. It's very clear to me the NCAA tightened it's purse strings from when they did a four-flight bracket in 2005, but there's no conclusive evidence yet that the NCAA tightened its purse strings from 2009 when it picked a three flight bracket over a feasible two-flight bracket (they could have paired UMD @ Minnesota that year, instead of shipping UMD to UNH).

4. Finally what does the NCAA see as its charter? Are they looking to field the 8 teams that had the best seasons or the top 8 most competitive teams at the end of the season.

Well the current criteria reflect a mixture of the two. The at-large selections and seedings definitely reflect the former. But the existence of the autobids gives some weight to the latter.

We can debate whether at-large selections and seedings should reflect more late-season improvement & success, but I believe what's of first-order importance is judging the best overall seasons correctly.

People intuitively have a problem with BU getting the No. 3 seed after losing to Northeastern and Maine, but I don't think you need to have a late-season criterion to move BU down -- BU's ranking just needs to be appropriately adjusted for losing to Northeastern and Maine at any point of the season.

 

[dave1381]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

Fly Boston to Minneapolis. Fly twenty five-thirty folks.

Is round trip commercial rate, currently. Not booked in advance would be $735 each for March 16. That is Frontier or Air Tran, Continental and United are $1,042. One stop. Total duration five hrs., plus/minus.

Non stop over $1,200.

Travel=$18,375 (25 flying) to $22,050 (30 flying). Likely NCAA can book ahead, perhaps and also get the best, much lower rates, too.

You can actually look up on the NCAA site what there estimated budget impact was for moving from a 2-team seeded bracket to a 4-team seeded bracket. I think they basically thought it meant 2 more flights, but I could be wrong. I don't have time to dig through the NCAA now cause it changed. (edit: I mean their web site changed)

 

[ARM]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

But the existence of the autobids gives some weight to the latter.

Doesn't the NCAA award an autobid to a conference, and it is the decision of the conference how their champion and autobid winner is decided? For example, it wasn't that long ago that none of the Ivy League, Big Ten, or Pac 10 had postseason tourney's in men's basketball, so their autobid went to the regular season champ.

 

[dave1381]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

Doesn't the NCAA award an autobid to a conference, and it is the decision of the conference how their champion and autobid winner is decided? For example, it wasn't that long ago that none of the Ivy League, Big Ten, or Pac 10 had postseason tourney's in men's basketball, so their autobid went to the regular season champ.

True, it is a conference decision. So to be precise, by giving conferences the discretion to give autobids to conference tourney champions, the NCAA gives some weight to rewarding late season success.

 

[Sean Pickett]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

Doesn't the NCAA award an autobid to a conference, and it is the decision of the conference how their champion and autobid winner is decided? For example, it wasn't that long ago that none of the Ivy League, Big Ten, or Pac 10 had postseason tourney's in men's basketball, so their autobid went to the regular season champ.

The Ivy league still doesn't have a post-season BB tournament.

Sean

 

[Hux]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

I agree the WCHA competition seems to serve the teams well, which is one reason I would like to see more non-conference games with Hockey East teams. On a personal level I would like to see BU schedule 2 game series against two different WCHA teams every season (home and away) or pairing with another Hockey East team for scheduling two WCHA teams. Which would be easier if the WCHA teams had more non-conference games.

Sean

I agree. Would love to see a weekend where say Minny comes in and plays BU and UNH, and Wisco plays Providence and BC.

 

[granddaddyscout]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

I agree. Would love to see a weekend where say Minny comes in and plays BU and UNH, and Wisco plays Providence and BC.

I like that idea as well, there are so many teams close together it would be great to get some variety. Certainly a better option that UW going to Robert Morris, UND going to Vermont. Unfortunatly the WCHA teams are more spreadout, but Wisco was willing to travel to St Cloud for a couple of games with Hurst and NU this year. Hope it happens soon!!

 

[FlyingPig]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

Abandon the 2 flight bracket restriction or preference. Having 8 teams means up to 4 flights.
When you have a 64 team tournament, dropping a team a seed or 2 to stay closer to home has very little impact.
In fact, it can be used as a tool to fine tune home court/home ice advantage.
Having only 8 teams in the bracket, dropping a seed can be a punishment to up to half of the bracket.
It makes 0 sense other than saving travel expenses. Seriously, if the schools are not concerned with the expenses, why should NCAA?
If the school can't afford the expensses, the solutions are simple. 1. drop out of the D1 women's hockey program 2. don't get into the tournament

 

[FlyingPig]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

1. drop out of the D1 women's hockey program 2. don't get into the tournament

I missed #3 cross country bus ride.....road trip!

 

[dave1381]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

I've been pushing the idea that KRACH should be used instead of RPI. I get the sense that some here think my "whining" has been without justification, because the teams I claimed were underrated by RPI lost yesterday (Minnesota & UMD), and the teams I claimed were overrated won (BU & BC). And if you look at the current KRACH, it still has Minnesota & UMD higher than BU & BC after today's results, while the RPI has BU&BC higher. Clearly then RPI is right and KRACH is dead wrong, right?

Not quite.

More or less, the reason why Minnesota & UMD do well in KRACH is because they have a win & tie vs. Wisconsin, and they don't have much in the way of bad losses (UMD has more, which is why Minnesota is higher). BU and BC still lack any quality results at the level of Wisconsin, while they have more bad losses.

So Minnesota losing at BC doesn't really change much about the argument that Minnesota deserved to be seeded higher. Minnesota's been a 50-50 or so proposition vs. teams at BC's level all season. Again, the argument for seeding Minnesota higher was based on their relative lack of bad losses compared to the Hockey East school (which matters not so much for predicting tournament outcomes), and their ability to compete with a school as good as Wisconsin (which presumably has some predictive power about whether they'll perform well against Wisconsin/Cornell in the tournament).

If BU or BC manage to score wins over Wisconsin/Cornell, then they'll finally pass Minnesota & UMD in KRACH, but only then. Still, the Frozen Four outcomes won't change my argument about seeding. BU & BC have not yet proven they can compete at a team at the Wisconsin/Cornell level yet, and they did not dominate their other opponents as if they were on a similar level to Wisconsin/Cornell. This is what the KRACH & Rutter rankings reflect, this is presumably why the polls had Minnesota at No. 3 entering the tournament, and this why I am firm in believing they were both BU and BC were seeded too high.

 

[Watson Rink]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

LATE SEASON WEIGHTING AND OTHER RANDOM FACTORS

It's startling that the four sports I happen to follow most closely at my alma mater have wildly different approaches to late-season weighting in determining national tournament placement, objective/subjective selection criteria, neutral venues, and transportation costs.

Women's hockey we all know about, and as mentioned above, Ivy league men's basketball rewards nothing but full-season results for the NCAA autobid, with playoffs held (as this year) only if needed to break full-season ties. On the other hand, women's squash entirely disregards early season results whenever teams play each other more than once, thereby removing the homecourt neutrality effect of home-and-away series in any given year although the late-season homecourt advantage usually alternates from year to year.

The women's squash seeding committee aims at producing close matches at the nationals based on end-of-season strength and is authorized to use highly subjective factors in doing so. For example, one year Cornell had lost all its matches to teams ranked #8 or higher, defeated all opponents ranked #24 or lower, and played nobody in between #9 and #23, so the committee slotted Cornell in at #12 simply in order to avoid rematching #11 Bates and #13 Colby for a fourth time that season (and the resulting 5-4 Colby win over Cornell proved the seat-of-the-pants seeding to have been pretty accurate). Because the national tournament is structured as a one-weekend, five-bracket, three-match elimination tournament, the key controversies about seeding involve the #8/#9, #16/#17, #24/#25 and #32/#33 slots, as a team seeded #9 can only hope to win the "B" bracket title whereas the #8 seed gets a chance to finish anywhere between #1 and #8 in the top bracket, and so forth. The geographic concentration of women's squash means that all matches are played at a neutral site (usually Yale/Choate) and only two teams, Stanford and Berkeley, ever (albeit always) have to fly. The one fly in the ointment is that not only does the Yale team usually play at home, but three of Yale's 15 courts are unique in college squash in having four glass walls instead of one, which requires some getting used to.

Trying to describe the complexity of the seeding for men's and women's crew is beyond me, but the interesting aspect is that although the brackets for the heats are intended to be equally seeded, each heat (as in many Olympic sports) advances more than one winner to the next round and, when you get to the six-team final, the lane you are seeded into may or may not prove to be as advantageous or disadvantageous as expected due to vagaries of wind and choppiness. As in squash, the tournament venue is neutral for everybody, but the logistics of transporting shells cross-country (perennial powerhouse Washington) or halfway (perennial powerhouse Wisconsin) are daunting. I don't know whether the NCAA provides subsidies.

I wonder if this crazy quilt pattern would continue you looked at all other collegiate national tournaments.

 

[dave1381]

Re: The Thread for Constructive Ideas for Improving the NCAA D-I Selection Process

I've looked the soccer, field hockey, lacrosse criteria at some point in time, I recall they're all somewhat similar to ice hockey though not exactly the same. Tennis has a complex computer rating system. I think it might be the best of the lot, but I've never looked closely at it. I assume it has to be better because the others are RPI based.