Although many analysts seemed to believe that the Texas Longhorns were pretty certain to get a bid in the 2012 NCAA women's basketball tournament, Jessica Lantz has made the point (here and here) that they should not have gotten a bid, particularly not over the Oklahoma State Cowgirls.
Of course, whatever we feel about Texas and their recent tournament history of first round exits, the question that had to be answered was who exactly should have gotten in over Texas if people don't buy that OSU should get the nod.
Jessica made a few suggestions that I took a closer look at over the weekend but she might have gained a new ally in her doubts about Texas' tournament worthiness after Selection Monday: Tim Mulholland of SB Nation's Streaking the Lawn thought the absence of his Virginia Cavaliers in the field of 64 was a "shocking and a massive snub, nothing less."
The 'Hoos final RPI was right around 50. Not great, I'll grant you, but not horrible either. Kansas (53), Texas (56) and Michigan State (62) all made it into the tournament. The Jayhawks and Longhorns both had sub-.500 records in the Big 12 and neither reached the 20-win plateau. Kansas was 4-10 in their last 14 games. Texas was at least marginally better at 6-8 in their last 14 games. Virginia? 8-6.
The Selection Committee Chair said that the fact that Virginia didn't beat any of the teams above them in the ACC is what kept them out of the dance. I understand that being a knock on us, but when all of those teams are seeded 4th or above in the tournament, that hardly seems like something worth knocking us all the way out of the tournament.
That is somewhat odd reasoning given how highly the committee seemed to think of the ACC and another tidbit mentioned in Mulholland's analysis: a quality win over the second-seeded Tennessee Volunteers.
So was Virginia the biggest snub?
While selection sheets rely heavily on RPI, strength of schedule, and quality wins to answer that question, another way to look at this question is to look at the quality of a team's performance over the course of the season, which is less tied to wins and losses than how well a team played - a team can win despite performing poorly and lose despite playing extremely well. Taken a step further, a team can pull off an upset simply because of a unique mismatch or advantage that wouldn't exist against any other team in the nation. Rather than wash that all away when looking at the tournament field, it's actually useful to take performance into account when attempting to figure out the quality of a team over the course of an entire season rather than crediting them for a win on a good day.
Dean Oliver's Four Factors provide one way to look at performance statistically by looking at how well a team performed relative to opponents in four crucial areas (listed in order of significance): shooting efficiency, turnover rate, offensive rebounding rate, and free throw rate. It may not surprise you that national champions routinely outperform their opponents in all four categories throughout a season because they are able to matchup well with most any opponent; Sweet 16 teams can get by with one weakness if the matchup is right, but teams with weaknesses relative to opponents in shooting efficiency and turnover rate will struggle to get out of the first round because their inefficient use of possessions makes it difficult for them to overcome the range of matchups that a tournament gauntlet presents.
Reframing performance
So to reframe Jessica's question a bit in light of Mulholland's disappointment, one might attempt to figure out which teams were the biggest snubs by asking the following question: which bubble teams were best equipped to respond to the range of matchups they might face in the tournament?
Whereas RPI and SOS are tools of results-based analysis of performance, this question is asking for a process-based analysis of performance or how a team played to get the outcomes they got. Women's Basketball State's State Ratings are helpful for that type of analysis because they take exactly that into account by including elements of the Four Factors along with strength of schedule and location-based factors, as you can read more about at their site. Although these ratings haven't been publicly available long enough to test them against tournament performance, they provide a different way to look at who might deserve to get into the tournament by way of comparison with RPI.
So the following is a list of the projected at-large bids before Selection Monday based on Charlie Creme's Bracketology that might've been less secure than people thought, projected bubble teams, and a few of the teams Jessica suggested might have deserved consideration statistically. In either contradicting or reinforcing the standard metrics used to make decisions, the State Rating - as well as the Sagarin ratings - might shed a different light on who really should be upset about not making it.
What matters here is the differentials between a team and their opponents moreso than their own performance alone: what you want to know for effective field goal percentage (eFG%) and turnover rate (Tov%) is the extent to which a team outperformed their opponent. I included rebounding percentage here as well, although offensive rebounding percentage is more important for these purposes; for reference, Liberty led the nation with a rebounding percentage of 62.9%, meaning they out-rebounded opponents 62.9% to 37.1%.
The numbers
Ordered by State Rating (as of 3/11/12), beginning with the highest rated omission from the tournament.
School | State Rating | RPI | Sagarin | eFG% diff | Tov% diff | Reb% |
James Madison | 38 | 34 | 66 | + 5.6 | + 1.7 | 53.6% |
Memphis | 46 | 84 | 89 | + 7.1 | + 5.1 | 52.1% |
Arizona State | 47 | 63 | 50 | + 8.6 | + 2.9 | 50.3% |
Louisville* | 48 | 22 | 21 | + 10.1 | + 0.5 | 52.5% |
Tulane | 49 | 70 | 95 | + 9 | - 0.4 | 54.6% |
Georgetown* | 51 | 28 | 19 | + 5.7 | + 9.4 | 52.1% |
Oklahoma State | 53 | 65 | 43 | + 8.3 | + 1.6 | 51.5% |
Arkansas* | 55 | 31 | 36 | + 6.1 | + 5.8 | 53.3% |
Michigan* | 59 | 44 | 45 | + 8.5 | + 5.7 | 46.2% |
Vanderbilt* | 60 | 49 | 28 | + 14 | - 0.9 | 51.9% |
Oklahoma* | 61 | 27 | 20 | +10.2 | - 1.5 | 49.9% |
Virginia | 65 | 49 | 28 | + 3.4 | + 12.2 | 49.7% |
DePaul* | 67 | 19 | 27 | +10.3 | +2.3 | 48% |
LSU* | 68 | 18 | 34 | + 12.7 | - 4.1 | 53.4% |
Central Arkansas | 69 | 91 | 117 | + 11 | + 3.1 | 53.3% |
USC | 70 | 38 | 49 | + 4.2 | - 0.5 | 52.7% |
Texas* | 71 | 55 | 35 | + 8 | + 1.2 | 51.4% |
Iowa* | 76 | 42 | 47 | + 9 | - 0.5 | 50.6% |
Temple | 78 | 45 | 46 | + 6.3 | + 5.2 | 55.4% |
Florida* | 79 | 41 | 44 | + 7.1 | + 0.9 | 54% |
Wake Forest | 81 | 84 | 52 | + 6.8 | + 4.7 | 48.7% |
Michigan State* | 83 | 62 | 48 | + 9 | + 0.9 | 55.1% |
Kansas* | 91 | 52 | 39 | + 7.9 | + 0.4 | 50.9% |
Iowa State* | 98 | 47 | 30 | + 8.4 | - 2.3 | 54.3% |
Kansas State* | 111 | 16 | 37 | + 4.4 | + 3.3 | 49.6% |
North Carolina | 113 | 89 | 41 | + 9.1 | + 0.4 | 53.9% |
Observations:
- So are Lantz and Mulholland right to pick on Texas? Texas fared pretty well by State Rating and Four Factors standards - what actually hurt their State Rating was penalties for four home losses, 10 conference losses, and their RPI. Otherwise, they were top 20 by the State's statistical ratings.
- But still Mulholland is probably right about Virginia. Virginia fans might have a bigger gripe than Oklahoma State fans - that Texas got selected while Virginia didn't is somewhat odd. What likely hurt Virginia's State Rating was their statistics: they didn't shoot well at all this season (their eFG% of 41% was 248th in the nation), which obviously influenced their ability to score. But Virginia was an outstanding defense, ranking third in the nation in opponents' turnover rate (hence the 12.2% differential) and holding their opponents' shooting efficiency beneath theirs. But does defense matter in the tournament?
- Defensive-oriented teams were not held in high regard. Five teams on the list above had a top-20 defensive rating in the nation and were left out of the field of 64 (ASU, JMU, Tulane, UNC, and Virginia). Granted, all of these teams had major flaws and none of these teams had a particularly strong strength of schedule, which did them no favors.
- So why was UNC seen as a top bubble team? I never figured that one out and their State Rating does nothing to help understand that. Their defense was solid, but the number that really stands out is their narrow turnover margin, which is more significant for that particular team than most others. To put that turnover margin in perspective, three of UNC's top four usage players were post players (Laura Broomfield, Chay Shegog, and Waltiea Rolle). Post players need someone to get the ball in order to score. UNC ranked 266th in the nation in turnover rate and had not one efficient distributor.
- So should JMU have gotten a bid? Given their in-conference strength of schedule, JMU probably needed an out of conference strength of schedule ranked higher than 103 and they missed an opportunity with a loss to Virginia. Statistically, while they're a strong defense, they also don't have any particularly dominant strengths that bode well for tournament success.
- Negative turnover differentials could be the beginning of an upset watch. Iowa, Iowa State, LSU, Oklahoma, and Vanderbilt are all teams that got in the tournament with negative turnover differentials during the season and that doesn't particularly bode well for success. Oklahoma will have a home court advantage, which will help considerably especially against a Michigan team that has the numbers to justify their bid off the bubble. Even moreso, Iowa State will need everything their fans can give them against a solid Green Bay team. But Vanderbilt could struggle with a Middle Tennessee State team that forces opponents into turnovers at a high rate and LSU will have a hard time against San Diego St if versatile senior LaSondra Barrett is less than 100%.
- Kansas, Kansas State might be the bigger points of contention for Oklahoma State. Kansas State stands out above as the team with the widest variation between State Rating, RPI, and Sagarin. What really hurt KSU's State Rating is that they simply don't have any prominent statistical strengths, though they don't have any particularly big weaknesses...except for rebounding. A team that neither shoots (ranking 185th in the nation in eFG%) nor rebounds well (190th in rebounding percentage) is not likely to go far in the tournament. Jessica already touched on Kansas but the Jayhawks do at least shoot well, which makes them a team that could get hot against a higher-seeded opponent.
- What about Michigan State? As a Michigan alum, I would've had to recuse myself from any judgment of MSU had the Wolverines not taken their rightful place back in the tournament. But with UM in, I can say that MSU has solid numbers and is a particularly strong rebounding team, which should help them despite a slim turnover differential.
So should OSU and Virginia have made the field?
Both OSU and Virginia had a strong case for making the tournament when you look at their quality of play over the course of the season. And if you look more comprehensively at the two teams' performances, you could make the argument that they should've been there over at least a couple of the teams that did make it.