Sunday, March 14, 2010

comments on the ncaa tournament

south:

look, i don't think the south is as bad as the gut reactions given to its composition. baylor is a fairly strong three seed among the three seeds. pitt isn't much better than baylor, new mexico isn't proven, and georgetown is a good team that got hot, but had some bad stretches. i don't see baylor as a weakness here. richmond is underseeded, and could have easily been put in the top half of duke's draw. the bottom half of this draw is deceptively strong.

it's true that villanova is a fairly weak two seed right now, but they were strong throughout the majority of this season and will be an extremely tough out should they win a few games. notre dame is hot right now, and old dominion isn't a team anyone wants to play. purdue is a weak #4 without hummel, but a & m, cal, and louisville (along with, for that matter, siena and utah state) make duke's side of the bracket tricky.

while this bracket may look a little soft at the top, seeds #5-11 are flat out tough teams to eliminate. i think the south has very good depth in the middle, and that we'll see this bracket acquit itself nicely as the tournament progresses.

sweet 16: duke, a & m, nova (could easily take richmond here), and baylor.

west

people are up in arms about the south when they should be up in arms about the strength of the west. vanderbilt is overseeded (in my view). here's a team that got way too much credit for a sweep of an opponent. people complain about louisville's profile for sweeping syracuse- what about vandy's for sweeping tennessee? out-of-conference results are mixed; wins over st. mary's and mizzou are complemented by losses to illinois, cincy, and wku, all who are not in the field. they performed well in conference, but couldn't beat kentucky, and lost to both georgia and south carolina. butler should have been the four here. gonzaga and florida state are both sort of boring teams; minnesota was a last spot, and so the real competition for the 'cuse comes from pitt and kansas state. both are good teams, but they're definitely not amazing, and they're both only really fringe title contenders. kansas state has a real chance at going down to byu, as does pitt to xavier in the second round.

sweet 16: byu, pitt, 'cuse, butler

midwest

this is a nasty bracket. ohio state's side is just miserable, and the georgetown/tennessee potential showdown in the second round is tantalizing- provided tennessee gets past a red-hot sdsu team. georgia tech is going to be a tough out for any team at this point, and oklahoma state took down kansas. kansas, meanwhile, should be able to take out unlv or northern iowa, but would have a potentially difficult matchup with michigan state in the sweet 16. either way, this bracket is misery.

an interesting seeding note should be the seeding of new mexico state as a 12. this has mismatch written all over it and seems really curious to me. i like michigan state here, although maryland may be able to get through a few rounds on the disrespect card.

sweet 16: ohio state, kansas, tennessee, and michigan state.

east:

kentucky gets the misery of being in a bracket with west virginia and temple, two teams you never really want to play. i think temple beats cornell- if they don't, wisconsin certainly will, and i think i don't want to watch wisconsin play. new mexico is a bit of a wild card: they could be explosive enough to cause any team fits, or they could lose to marquette in the second round, who may not even get past washington.

kentucky doesn't want to play texas in the second round. that's a nightmare matchup- pittman is a beast inside, and there are plenty of guards to throw at the young cats. that being said, john wall should be able to do a lot of damage- opposing point guards have had their way for most of the year. west virginia could easily have problems with a team like clemson, which will likely be just as physical as the mountaineers.

sweet 16: uk, wvu, marquette, temple

in addition, i'd like to rate the 13-16 i think are POSSIBLE (not necessarily probable).

0-0.5%: lehigh, etsu, playin, ucsb, robert morris.

thanks for coming, i guess. not liking the chances.

0.5%-1.5% vermont over syracuse, morgan state over wvu.

these are at least possible, if wildly improbable. syracuse will likely put a hurting on vermont for past sins, but vermont is pretty underseeded compared to, say, houston. morgan state is a decent team.

1.5%-3% north texas over kansas state, ohio over georgetown.

ohio has a chance since the mac is reasonably competitive. north texas has a chance because k-state will be nervous.

3%-5% oakland over pitt, houston over maryland, montana over new mexico.

both pitt and maryland can look shaky at times, although houston isn't likely to be able to take advantage of maryland. montana is a good team, but new mexico will likely blow them out.

5-10% sam houston over baylor, murray over vandy.

both of these games are faux-rivalry games, so the lower seeds have added incentive to play hard. they're also likely run into each other in high school. baylor has a high seed to play up to, and murray state seems due to win one of these types of games.

10%+ siena over purdue, wofford over wisconsin.

wofford might get slowed to death, but wisconsin has a hard time putting teams away. siena is nearly a coin flip against purdue at this point.

Saturday, March 06, 2010

how to expand the tournament to 96 teams.

i wouldn't expand the tournament, but if it were to happen, here's how i think it should happen:

1. every conference champion is an auto-bid.
corollary: every conference champion is guaranteed a top-16 seed if they win their conference tournament.
2. every conference tournament champion is an auto-bid.
3. every top-40 rpi team is an auto-bid.

What would that look like this year?

Here are the "locks" from the Top-40 RPI:

Kansas, Duke, Syracuse, Kentucky, Kansas State, Villanova, West Virginia, New Mexico, Pitt, Baylor, Vandy, Texas A & M, Purdue, Georgetown, Temple, Tennessee, Butler, Wisconsin, Xavier, Maryland, BYU, California, Northern Iowa, Clemson, Texas, Michigan State, Oklahoma State, Richmond, Ohio State, Rhode Island, Utah State, Missouri, SDSU, Wake, Georgia Tech, Gonzaga, Old Dominion, Louisville, Florida State, and Siena.

Note that, in all likelihood, there is a lot of overlap between 2 and 3.

In addition, the non-RPI top-40 conference champs: Stony Brook (America East), Lipscomb (Atlantic Sun), Weber State (Big Sky), Coastal Carolina (Big South), UCSB (Big West), UTEP (C-USA), Cornell (Ivy), Kent State (MAC), Morgan State (MEAC), Quinnipiac (Northeast), MURRAY STATE (OVC), Lehigh (Patriot), WOFFORD (Southern), Sam Houston (Southland), Oakland (Summit), Jackson State (SWAC) Troy (Sun Belt).

We're up to 57 teams. For argument's sake, I'll assume some crazy stuff so the following teams win:

WCC: ST. MARY'S won already
WAC: Nevada
Sun Belt: MTSU
Summit: Oral Roberts
SWAC: not Jackson State
Southland: SFA
Patriot: Lafayette
Northeast: Robert Morris
MEAC: Delaware State
MAC: Akron
Horizon: Wright State
CUSA: Tulsa
Big West: Pacific
Big South: WINTHROP won already
Big Sky: Montana
Atlantic Sun: ETSU won already
America East: Maine

This still leaves us with 22 at-large births in a worst-case scenario format.

Some notes:

From the A-10, 4 teams were top-40 RPI and I did assume that one of those teams won.
Pac-10 is really hurt in this scenario as I assumed Cal won.

Here are my 22 at-large:

ACC: Virginia Tech (6 RPI top-40 already included)
Big East: Notre Dame, Seton Hall, South Florida, and Marquette (5 RPI top-40)
Big 12: no more (7 RPI top-40)
Big 10: Minnesota and Illinois (4 RPI top-40)
Pac 10: Arizona State and Washington
SEC: Mississippi, Mississippi State, Florida (3 included already)
C-USA: Memphis, UAB, Marshall
A-10: Dayton, St. Louis, Charlotte (4 RPI top-40 already included)
Colonial: William and Mary, Northeastern
MVC: Wichita State
WCC: Portland

The following teams are still excluded: Cincinnati, UConn, Wright State, Harvard, Fairfield, Iona, Illinois State, Arizona, IUPUI, Vermont, VCU, Charleston.

A couple of points:

One is that some conferences have too many births. Many conferences would stand to get two births, and it could be argued (albeit somewhat unconvincingly) that conference champions would lay over and die to get another team in their conference in. There are two problems with this: one, it definitely doesn't happen in the larger conferences. Look at the history of the ACC tournament- lock teams still knock out their counterparts. It only could happen in the small ones. To eliminate this problem, one could easily incentivize the conference tournament somewhat, possibly by guaranteeing a top-16 seed to a conference champion who wins, ensuring that the lower-conference teams have a relatively better chance to win their first round game.

Another is that this model assumes straight-forward behavior in all the major conference tournaments this year. The SEC is a notorious crapshoot, and the Pac-10 appears to be similar this year. I would like to claim that I get around this assumption somewhat by assuming that every small conference (as of today) has a surprise champion and takes two bids. This is wildly unrealistic, and it is likely that even more at-large bids would open up.

The main advantage I see with this model is that, while it does not have the 96 best teams (or even close), it does suitably incentivize regular-season play. Marginal teams like Cincinnati, UConn, Virginia, or Arizona are not rewarded for very poor showings because of the potential of multiple slots being taken up by less-qualified teams. If rule 1 were eliminated, probably 10 conference champions would not be selected by the tournament committee, and instead we would see a handful of teams from big conferences with poor records and no real profile. The 22 teams that were selected already contain stretches such as Charlotte, Marshall, Minnesota, and Seton Hall: teams with little on their resume to discuss.

I think the real point of this exercise is to show how yucky a 96-team tournament could be. I don't know if a tournament would really be enhanced by taking out Morgan State *should they not win their conference tournament* and putting in a Cincinnati, since both teams are not likely to make a sweet 16. Most bubble teams do not make big runs in the tournament.