I have a running debate with an old friend who now lives in Kentucky, where college ball is king, about expanding the NCAA tournament. He says it’s fine the way it is, don’t fix what ain’t broke. I say the current selection and seeding process is too often arbitrary, if not biased, and that blowing the tournament open would be in the best interest of basketball, fairness, and good old fashioned logic. Pedro, this blog is for you.
There’s a lot of talk these days about expanding the NCAA tournament field to 96, and many coaches are all for it. A leading proponent of the idea happens to be Mike Krzyzewski, whose Duke team is usually well served by the current selection process. They’re a #1 seed again, and this time it’s moot whether it was deserved, unlike the #2 seed the selection committee handed the Dukies in 2009.
The expansion Coach K favors would eliminate the NIT (which the NCAA owns anyway) and basically have a play in game for 32 of the top 64 seeds. The selection and seeding process would be left in tact, and the conference tournament cash cows, automatic bids and all, would remain.
This wouldn’t affect Coach K or Duke much in most seasons. The Blue Devils would still get a top 20 seed, receive a bye as 64 teams play for the 33-64 spots. Duke’s first opponent might possibly be a little tougher, maybe not, but other than that, the powerhouses of college ball would still be the powerhouses. Yet it would create 32 more seeds for the selection committee to assign, 32 more chances for error, 32 more potential controversies and complaints. 96 is not enough.
I say open it up further. Open it wide. State high school tournaments include every team — why shouldn’t the NCAA’s? That’s right, all 347 D-1 teams* (see notes below), seeded with a new and improved, committee-proof ranking system.
Any expansion should find a more systematic way of seeding teams that takes the job from the committee, yet is not so complicated that only math majors can understand it. Such a process should do a number of things, some of which the current selection process fails to do. Miserably fails. It’s not clear year-to-year, team-to-team or conference-to-conference what committee members look at with the at-large bids, whether the RPI strength of schedule rankings, regular conference play, the conference tourneys or the nonconference games. It’s time for a new seeding process that would:
1) Preserve regular season conference play, something the conference tournament system undercuts. Win the conference tournament, get an automatic bid. It causes pain for deserving teams every year, and raises other issues. How did the Big Ten champ, Michigan State, receive a lower seed (5th) than the team that finished 4th and blew its opening conference game? That was my Wisconsin Badgers, proud owners of a #4 seed.
How did Illinois, a team that was 11-9 in the Big Ten, tourney included, miss the big dance while a team that lost 10 conference games, Minnesota, received a bid — a top 44 seed no less. Based on Big Ten play, nobody believes Minnesota is that much better than Illinois, and they’re not. But the tournament selection committee says they are. (I’m writing from Big Ten country so many of my examples will be from the Big Ten and Horizon League. But nearly every conference could take issue with NCAA selection and bracketing every year. Not the ACC of course. But the howl of injustice from the mid-majors like the Atlantic 10 gets louder all the time. )
Back to preserving the regular season: Why should one conference tourney final decide who goes to the big dance? It shouldn’t.
2) Maintain meaningful conference tournaments. I’d eliminate them altogether but the conference commissioners won’t give them up. This is the trickiest part of properly seeding tournament expansion, and the disparity between the majors and the small school D-1 conferences is wide. But if they are going to play them anyway, the tourneys have to count for something.
A standard point system (which I’ll get into later) for each round of a tournament would help equalize the NCAA seedings to some degree and give some importance to the tourney games. As far as fan interest in the tourneys, the conferences will simply rake in what they can market.
3) Preserve the importance of nonconference play. This should be easy enough to do, as this seems to be what, in theory, works, with RPI strength of schedule rankings and the polls relying on these games. But do they? The Big Ten won the ACC-Big Ten challenge 6-5 this season and Wisconsin beat Duke, which might have mattered more to the selection committee in seeding the Badgers than their loss to Illinois in the Big Ten tourney quarterfinals. But the ACC boasted six NCAA berths, and the Big Ten only 5.
How much does the committee appreciate nonconference success or failure? Many a mid-major program would question this. The Big Ten might question this, too. In a down year for the ACC, Duke lost both of their tough nonconference road tests (Wisconsin, Georgetown) yet was ranked 3rd in the RPI largely because its ACC competition was considered stronger than the Big Ten — despite losing the Big Ten challenge and disappointing tournament showings last season. All six ACC seeds were in the Top 40.
As far as the NCAA’s and NIT were concerned, 3rd place Horizon League team UW-Green Bay (21-12) didn’t help itself by beating Wisconsin in December. UW-GB wasn’t picked for either tournament. Why even play the game? The Phoenix did get an invite to the 16-team College Basketball Invitational. Wright State (20-12), 2nd in the Horizon League, was also snubbed by the 48-team NIT. They’re so disappointed they’re skipping the smaller tourneys.
Butler (28-4) ran the table in the Horizon League, earned a #5 seed in the NCAA but has no signature win the likes of UW-GB’s upset over Wisconsin. If Butler truly is a top 20 team by virtue of their record, is it really possible that their 20-win Horizon rivals don’t fit somewhere in the top 113 teams? Not likely.
St. Louis University (20-11) finished three games ahead of Dayton in the Atlantic 10 yet was also snubbed by the NIT. Billikins coach Rick Majerus is pissed.
A new system that assigns a value to each of the 32 conferences, and then sets scaled value to teams in the upper echelons of the conference would properly account for and reward both conference and nonconference success.**(see notes below) The current NCAA selection process for its tournaments is far too flawed.
4) Rank the conferences. Twelve wins in the loaded Big East is better than 12 wins in the ACC. Twelve wins in the Big Ten were better than 12 wins in the ACC, but not as good as 12 wins in the Big East. But wins against the top teams in the Big East or Big Ten were high caliber victories. A win against the upper echelon of the Horizon league wasn’t half bad either, perhaps more worthy than beating lower echelon Big Ten teams such as Indiana.
Ranking each of the 32 conferences after the nonconference schedule (and also based on the previous NCAA tournament) would add value to games like the ACC-Big Ten challenge, and provide a system for assigning value to wins. The average number of teams in each conference is 11. We’ll simply assign the teams a number, 1-11, based on their place in the conference.
But wait you say, there are 16 teams in the Big East! Tough break, ‘Cuse fans. You get the same number of points for crushing Depaul as you would beating UConn, this year anyway. The point is to ensure that the Top 128 seedings are as equitable as possible, not to engineer the bottom 128. The games would still have to be played.
If the Big East is deemed the best conference, they are assigned 32. Beat conference winner Syracuse at any point in the season, get 32 x 11 points or 352 points. Beat Depaul, get 32 x 1, 32 points.**(see notes on home vs. road games).
Big Ten was the 2nd best conference in the land this season. A win against a Big Ten team is an automatic 31 points. UW-Green Bay’s win over Wisconsin, 4th in Big Ten play, would net 31 x (11-3), 248 points. Wisconsin’s win over Big Ten champ Michigan State would net 31 x 11, 341 points.**
Sure, this feeds the better conferences and allows the best teams in the best conferences to get fat quick. But they’re fat already, and if mid-major team can come into Madison and the beat the Badgers, your team leaves with 248 points. Duke couldn’t do it. Illinois did. This system locks value on that win into the seeding process.
It doe, however, allow middle-of-the-pack teams from the Big East and ACC to amass points in conference mismatches that the best teams from lower ranked conference can’t. But this is where the injustices are occuring now — the Illinois’, the Virginia Tech’s are being left out. Top tier Horizon League and Atlantic 10 teams can’t crack the NIT. Expanding the tourney to 96 wouldn’t end this, only move the controversies down the ladder. They’re already there.
In this system, the best chance that an 8-win Big East team would score higher than a top team from the Horizon would be to knock off a heavyweight at the top. But the better Horizon League teams are amassing points against the bunnies in its conference and winning more games. Figuring about a 20 rank for the Horizon, there a lot of points to be had with 12 wins, 220 if able to upset league champ Butler this season on the road. In this scenario, Cleveland State would have earned 180 points for beating 3rd place UW-Green Bay. The most Georgetown can get by beating up on 10th place Seton Hall is 64. Six Big Easts teams, Cincinatti through Depaul are assigned a 1.
Also, conference tournaments with standardized point values for each round (call it 32 in the round of 8, 64 in the semis, 128 for the final) would help a team like Butler build points (224 total) that a UConn or Illinois or Michigan wouldn’t.
The nonconference games would offer also provide a means of amassing points while improving (lowering) the ranking of any conference. The previous year’s tournament would also decide the conference rankings, so if Cinderella couldn’t crack the round of 64 one, the teams in Cinderella’s conference could be playing a lot of play-in games, 347 down to 256 the next. The best of the small, low ranked conferences, however, would still likely not face a Kentucky or Kansas until the round of 128. The best of the mid-majors, the Gonzagas and Butlers, would have opportunites to amass enough points to steer clear of such a matchup until the round of 32 — if they get past the Northwesterns and Providences in their path through the early rounds.
Would it kill Duke to play a team with a losing record in the round of 256? I don’t think so, and the round of 128 would be mostly comprised of winning teams. It would be great for the small conference schools, if not CBS’ ratings. But NCAA basketball isn’t about Duke or CBS or the selection committee. It’s about basketball, the more the better for the fans, the players and the schools.
And the expansion would only take a week to play out. A week that would have students and alumni at hundreds of schools all over the country dancing.
* While initially I thought including all 347 teams would be possible, I’m backing off on this. I don’t think there would be that much interest in matchups between teams with losing records, nor is it at all interesting to match a #1 overall seed play the #256 seed. However, as the 2010 tournament unfolds, we’ve already seen 13-seed Murray State beat a #4, Vanderbilt. 10th seeded St. Mary’s has defeated #7 Richmond. A #11 (Old Dominion) beat a 6th-seeded Notre Dame team out of the Big East. Not only does this suggest that expansion would be a good idea, but that the selection committee’s ability to seed even the top 25 teams is wide open to question. I’m not surprised, as #11 Wisconsin’s relatively easy win over 6th-seeded Florida State in the 2009 tournament was the inspiration for a conference ranking system. If the committee can’t properly rank 32 teams, how can it be expected to fairly rank 96 or 112? It can’t.
No system is perfect, but what should the cut-off be? Should there be 32 byes and 32 play-in games to 64? That’s the proposal coach K favors but I don’t think it goes far enough. Should there be 32 play-ins to get to the 64 play-in games? That’s getting there, but it would really be better to break groups into four and play-off for the slot — also economically more beneficial. This would create 160 teams with 128 (4 x 32) playing for the 33-64 slots. Whatever the best number, the conference ranking system, the process of setting value to each win, will be critical to fairly seed the tournament.
**As this is a work in progress, a note on road games. In this points system, no credit is given for a loss, wherever it is played. But wins on the road should count more heavily than wins at home. Teams such as Wisconsin and Duke are very strong home teams, vulnerable on the road (Duke was 5-6 on the road, including losses to Georgetown and Wisconsin). Halving the point haul for wins at home, leaving full credit for road wins is a simple way to fix the disparity. Wisconsin’s home win against Michigan State would then result in 341/2 , or 170.5 points, still a strong tally. A Badger win at MSU would net the full 341.
This home/road correction creates opportunities for mid-majors such as Temple (loser to Georgetown by one on the road) to significantly increase their stock with a road upset, while equalizing home strength. Wisconsin was 17-1 at home and 6-6 on the road or at neutral sites. Duke was 20-0 at home and 6-5 on the road or at neutral sites, not including conference tourney games.
Neutral sites such as Holiday tournaments, will be considered road games for both all teams. Conference tourneys have their own round-by-round scoring system.