Sunday, March 13, 2011

From 345 to 68, Epilogue

My reaction to the release of the 2011 NCAA tournament field was surprise.

My reaction to the reaction to the release was disgust.

With the lowered bar for entry -- 37 at-large teams, the most ever -- and the absolutely brutal set of performance characteristics for the 20 or so teams in line for the last spots in the field, there is simply no defending a team that didn't get in. I missed on two teams, and whereas in past years I have felt that I was right and the committee was wrong, there's no such reaction this time. The committee swapped my last team in (Colorado) and first team out (Southern California), and they chose the regular-season champion of a good conference (Alabama-Birmingham) over the fourth-place team in a somewhat better one (Virginia Tech). I had UAB as my sixth team out, largely because I didn't think the committee would value the conference championship highly enough relative to the lack of top-50 wins and the tournament quarterfinal bad loss. UAB was on my board all week and

My only question about the UAB pick is the contrast to the USC one. The two are dissonant, and serve only to further cloud the issue of what exactly the committee is looking for. UAB had a good record and a high RPI, a dearth of great wins but a lot of good ones, and performed well in its conference. USC had a mediocre record and a low RPI, a lot of great wins and a lot of terrible losses, and was T-4, with a 10-8 (11-9) record in its conference. I can understand putting either in, but to put both in is baby-splitting -- UAB paired with Harvard or Missouri State would have been consistent, as would USC and Colorado.

Committee Chairman Gene Smith wasn't helpful in explaining this oddity, and in fact, came off as delusional in post-game interviews, saying to ESPN: "This year there are a lot of good teams out there, moreso than in previous years for me."

The treatment of Virginia Commonwealth was shameful, with the collected experts in Bristol showing a complete ignorance of their resume. Despite Jay Blias' assertions, VCU was in the discussion. I had them in. Jerry Palm had them in. Andy Glockner had them as one of his last few out. Joe Lunardi had them as one of his top eight out, and I believe they were higher earlier in the week. From my last piece yesterday:


"The thing is, I'm not sure VPI is the best Virginia school in the discussion. Virginia Commonwealth has a better tournament resume, based on all the criteria listed by the committee than the four ACC teams in toto. They have a higher RPI (49) than all; they reached their conference final; they were just fourth in their conference, and the #3 team is not under consideration -- that hurts. VCU was 8-6 in true road games (the non-FSU ACC teams only played 10 road games each and none was above .500) and 12-8 in R/N. VCU had as many top-25 wins (2-3) as the four ACC bubble teams combined, and nearly as many top-50 wins (3-6) as the group (four). Echoes of William and Mary. VCU also had an 8-8 mark against the RPI top 100, which matches VPI and trails only Clemson (9-8).

"Unless the argument comes down to 'was Duke forced to play a game in your building,' I don't see how VCU ranks below any of these teams save perhaps Florida State, and then only if you give them full credit for a healthy Chris Singleton. This isn't a William and Mary argument, which relied on some key pieces of data -- ALL the data has VCU ahead of this group of teams.

"I don't know what the committee will do. I just know what the data says. I might be able to get one or two of the ACC bubble teams in ahead of VCU, but I can't see three or four."


It wasn't just that the panel thought VCU didn't belong, which is a case you can make with the last 12 teams in and the first 12 out. It's that they seemed to have no idea that they were part of the discussion. That's not a difference of opinion. That's being ignorant of what is nominally your area of expertise. How hard could it possibly be to look at a Nitty Gritty report and notice that VCU has all kinds of markers in its favor, especially relative to the ACC teams in the discussion? VCU did things they did not do. Disagreeing with their inclusion is fine; from the tone of the conversation, I just figured they were all getting VCU and VMI mixed up.

VCU was in the discussion. UAB was in the discussion. I don't think it's at all a coincidence that the two teams ESPN's staff was most offended by were from conferences that their networks rarely, if ever, televise. Nor do I think it's a coincidence that the two teams they defended most aggressively -- Virginia Tech and Colorado -- are from conferences they have deep ties to. No memos are out there, no policies are in place, but the people employed by the network are part of the established power structure, the same one that thinks it's OK to never play road games against good teams out of conference.

The call for more basketball people on the committee…I have no idea what this means other than more code, like the use of the "eye test," for "we need something that elevates that with which we are familiar with over that which we are not." I don't need more people to tell me that the BCS leagues have better athletes, maybe even better basketball players, than the rest of Division I. I need more people familiar with what the committee is trying to do and the standards they're attempting to apply in doing so. I need people who can look at Virginia Commonwealth and Virginia Tech and craft an actual argument, rather than act offended that their friends are going to the NIT.

None of the above was the real problem with the reaction. No, the most counterfactual, offensive, damaging assertion was that UAB and VCU were able to get into the tournament by gaming the system. The idea is that their high RPIs -- and this is a UAB thing more than a VCU thing -- are somehow not legitimate, achieved by playing easy-to-win games that nevertheless pump up the numbers.

First, Digger Phelps launched into a point that implicitly accused the committee of counting mid-majors that concluded, "I think it really hurts the power conferences."

Then Hubert Davis got involved: "I haven't looked at their their numbers. I haven't looked at their RPI numbers, strength of schedule numbers. I hope that's not the reason that they got into the NCAA tournament."

How can you be involved in this discussion and not know what a team's numbers are? You want to weight them differently, fine, but I'm astounded that you can go on the air, for money, and advertise that you don't know critical information.

More Davis: "I'm not saying UAB and VCU did this, but there are a number of programs, a number of teams that know this system, and they will schedule, make their schedule out to make their numbers look good."

The idea that the current system favors the mid-majors, that it's being exploited for their gain, is not only laughable, it's demonstrably untrue. Maybe the biggest problem in college basketball is that teams in the mid-tier conferences can't get games against the ones in the top six, and they absolutely can't get home games. Mid-majors have been screaming at the top of their lungs for years about wanting to play up, and the better those teams have gotten, the less access to games they've been able to get. Teams in the BCS leagues refuse, out-and-out refuse to play road games at teams in the #7-#18 conferences.

In fact, the RPI gimmickry cited by Phelps and Davis is actually the purview of the power leagues, who have taken to playing road games against bottom-100 teams in an effort to gain "road win" points in the new version of the RPI. (They understand that there's a concept in play, but don't quite grok the details.) The ACC played as many road games at Elon (2) and UNC-Greensboro (4) as they did against mid-major schools in the top 200 (6), and one of the latter games was in an exempt event hosted by one. Miami played at Florida Gulf Coast. Florida State played at FIU. Wake played at UNC Wilmington. You think Conference USA is trying to game the system? Really?

Florida State, Georgia, Penn State and Illinois are in the tournament because they got to play home games against highly-rated teams. I submit that without that opportunity -- the league forcing good teams to play at their place -- none would be in the tournament. All their cases rested on one big win at home in conference. Alabama, Virginia Tech and Colorado were similarly situated, with a single home conference win buoying the entire profile. You think UAB or Harvard or Missouri State wouldn't like to be able to play home games against Duke and Kentucky? You think they have scheduling advantages relative to the teams that choose to play a lot of bad games and never go on the road against the #7 to #16 conferences in the country? The entire ACC played six road games against mid-majors!

RPI gimmickry isn't playing better-quality competition, it's going on the road against terrible competition. If you can't tell the difference between teams 150-200 and teams 250-300 on the RPI, just looked at the damned performance of both. There's a significant gap. One set of teams is much, much better than the other. Playing the former instead of the latter is the choice teams make because teams 1-100 on the RPI -- 73 of them in particular -- won't play them, and they damned sure won't come to their place to play them.

And then their friends at ESPN go on TV and make it sound like they're getting screwed over it.

You want to be ignorant about why the selection committee would pick a team that won a top-10 conference outright, went 10-7 against the RPI top 100 and posted the highest RPI of any bubble team, fine. You want to ignore road record and top-100 wins and schedules within a conference, fine. But if you go on the air in front of millions of people and lie -- just flat-out lie -- about who has the advantages and disadvantages in scheduling in college basketball, you should be ashamed of yourself. What Hubert Davis and Digger Phelps did last night was no less embarrassing to themselves and their network than what Billy Packer did five years ago when he showed such incredulity over the selection of George Mason, a perfectly-qualified Colonial team, to the field.

You'd think everyone would have learned.

I could probably write 15,000 words on the "Bracketology" show. There were more erroneous, easily-disproven assertions crammed into two hours than I can every remember seeing on live television. There were more bad ideas about what a tournament resume than I could hope to summarize -- Digger Phelps was particularly productive on this score. I've pretty much stopped watching baseball studio shows because I couldn't take the "analysis"; this show made "Baseball Tonight" look like Yalta.

From 345 to 68, Sunday 5 p.m.

I'm assuming Ohio State beats Penn State, and even if they don't, nothing in here will change.

--


With Richmond beating Dayton in the Atlantic 14 final, there are six spots remaining in the field for bubble teams. Dayton's run wasn't good enough to get them on the board, although it did get me to at least peek at their season. Their RPI, top-50 and top-100 performance are comparable to the bubble teams, but you can't erase a T-8 performance in the conference with two wins, just one a good one.

Let's attempt to narrow down the 12-team field -- or at least put it into a coherent order -- by using smaller groupings. The ACC, for example, has four bubble teams with very similar resumes. Florida State (55). Clemson (56), Boston College (58) and Virginia Tech (61) sit in a narrow band of the RPI. They have virtually identical records (20-21 wins, 11-13 losses apiece), records against the top 25 (zero or one win, four or five losses) and top 50 (0-2 wins, 5-6 losses). Florida State finished an outright third in the ACC, ahead of the next three who tied for fourth. (However, Florida State finished the season without its best player, Chris Singleton, and went 3-3 without a quality win. Singleton is expected back for the postseason, but as I said on Twitter, unless a player is provably dead, teams will insist that he's going to play so as to not negatively affect their selection or seeding.) Two of the three fourth-place teams, Clemson and Tech, made the ACC semifinals.

Florida State stands out from the group in a positive way. They had the regular-season finish and their 8-5 road record is by far the best of the bunch. Let's slot them first. I think you can pull Boston College out and place them towards the bottom: they were the one T-4 to not make the semis. They have the worst record against the RPI top 100 (7-11) of the three and just a 1-5 mark against the top 150.

Clemson versus Virginia Tech is something of a toss-up. Both T-4, both lost in the ACC semis, they have essentially identical RPIs. Tech has wins over Duke and Penn State for a 2-5 mark against the RPI top 50. Clemson has no top-50 wins. Clemson beat Virginia Tech, but it was at home and their only matchup, so it's not a fair separator. Tech was better in true road games and in R/N games. Clemson played a slightly tougher schedule. Because I have to choose, I'll put VPI slightly ahead based on road performance and better wins.

The thing is, I'm not sure VPI is the best Virginia school in the discussion. Virginia Commonwealth has a better tournament resume, based on all the criteria listed by the committee than the four ACC teams in toto. They have a higher RPI (49) than all; they reached their conference final; they were just fourth in their conference, and the #3 team is not under consideration -- that hurts. VCU was 8-6 in true road games (the non-FSU ACC teams only played 10 road games each and none was above .500) and 12-8 in R/N. VCU had as many top-25 wins (2-3) as the four ACC bubble teams combined, and nearly as many top-50 wins (3-6) as the group (four). Echoes of William and Mary. VCU also had an 8-8 mark against the RPI top 100, which matches VPI and trails only Clemson (9-8).

Unless the argument comes down to "was Duke forced to play a game in your building," I don't see how VCU ranks below any of these teams save perhaps Florida State, and then only if you give them full credit for a healthy Chris Singleton. This isn't a William and Mary argument, which relied on some key pieces of data -- ALL the data has VCU ahead of this group of teams.

I don't know what the committee will do. I just know what the data says. I might be able to get one or two of the ACC bubble teams in ahead of VCU, but I can't see three or four.

VCU, by the way, rates ahead of Harvard due to the quality wins (3-6 vs. 1-5 top-50; 8-8 vs. 3-5 top-100). Harvard has the best at-large case of any Ivy League runner-up there's ever been, with quality wins over bubble teams (Boston College, Colorado), great stats (35 RPI, 9-6 R/N 140 SOS -- in the Ivy!) and a pretty good subjective case. The lack of top-50 wins doesn't hurt them relative to the ACC teams, but the lack of top-100 wins does, and while I'm willing to back a two-bid Ivy, I just don't know if the committee will take a team with three top-100 wins that didn't win its third-tier conference.

The lack of quality wins leads us into the other regular-season champs or co-champs on the bubble. VCU and Harvard are ahead of them all. Alabama-Birmingham, St. Mary's and Missouri State combined have two top-50 wins, and that's just not enough. UAB has the best case, with 10 top-100 wins, but the quarterfinal flameout in the Conference USA tournament was exactly the wrong move at the wrong time. St. Mary's and Missouri State have three top-100 wins apiece. It is possible the committee will look at their conference championships and tournament-final losses and put them in, but the lack of wins will probably carry greater weight.

The SEC teams are similarly hard to separate. Georgia has a much higher RPI than does Alabama, went 10-8 in the good half of the SEC, was 9-7 outside of Athens and 7-4 in true road games. They also went 3-9 against the top 50 and 5-11 against the top 100, which are really unimpressive numbers that seem to indicate Georgia cannot play with tournament-caliber teams. Alabama also had just five top-100 wins, but in the weaker SEC West -- which they did win with a 12-4 mark -- played just 12 games against the top 100, going 5-7. Their 4-4 mark against the top 50 is their strongest case, but those four wins (home against Georgia and Kentucky, at Tennessee in OT, Friday over Georgia in OT) are their whole case. They have what would be a record-setting RPI for an at-large bid (80) and they were lousy away from home (5-11).

The head-to-head results aside, I'll take Georgia ahead of Alabama. Neither is all that impressive, and both are, at best, destined to play Tuesday or Wednesday.

That leaves Colorado, which is 4-3 against the top 25, 5-7 against the top 50 and 8-10 against the top 100. That is mostly owning Kansas State, but the data counts. They played a wretched nonconference slate filled with free wins, and were terrible on the road (3-9) and in R/N games (6-11). They look like an NIT team to me, but again, seven of these teams have to get in the field.

I had taken Southern California off the board, but I went back and put them on after looking at how weak this field was. The fact is, their low RPI, high number of losses and middling finish in the Pac-10 should have left them out of this discussion. The slightly lowered bar for entry as well as the weak performance by the bubble pool gets them back in, and they actually look very good. Their RPI is higher than even Colorado's (69), which reflects their many bad losses. On the other hand, no team on the bubble has their wins -- 2-3 against the top 25, 5-5 against the top 100 -- and they were a much better team with him, going 13-10 with wins at Tennessee, over UCLA and over Arizona. The Trojans in their current state, with Fontan, have shown more than Michigan State or Florida State have in their current states, and certainly more than Georgetown or BYU have in theirs. All things considered, they're slightly behind Colorado.

In this strange year, what I know for sure -- what I think may be the most important thing to come out of the 2011 bracket -- is that the committee is going to be forced to make itself explicit about its criteria. There will be no hiding why a team is chosen or excluded, seeded up or down. The choices it makes this afternoon will be a de facto referendum on factors such as the impact of injuries in evaluation; on the value of non-majors in NCAA basketball; on winning a conference title; on unbalanced schedules within a conference; on performance away from home. Despite removing data on it from the "Nitty Gritty" sheets, is the committee still putting weight on some subset of recent games? Are they applying any or all of these criteria consistently?

To put in Virginia Tech over VCU is to say that the opportunity, by conference affiliation, to play home games against top teams is the most important path to a bid. To put in Alabama over Georgia is to abandon RPI and road record in favor of head-to-head and conference record absent context. To select UAB or Missouri State but not Clemson is to emphasize winning your league above other factors.

The committee's choices between VCU/Harvard and the ACC/SEC teams are going to reverberate, are going to make clear whether teams in non-BCS leagues are going to be evaluated on their merits. We have a situation where despite not having the opportunity to build up quality wins or RPI points in conference, these teams have been able to produce RPIs, top-25 records and top-50 records nearly identical to the big boys. Will that be taken seriously?

Finally, will the committee continue to value the ability to lose to good teams disproportionately? This trend is disappointing and is the biggest edge teams in megaconferences have. Michigan State, Marquette, Illinois, Clemson, Georgia…these teams have done as much to show that they cannot compete with the top 50 teams in America as anything else, but in some cases even I've had to put them in. If this trend continues, teams in the the BCS leagues will be even more free to ignore the ones in the next nine, counting on playing each other and their conference schedule secure in the knowledge that a dozen good losses and a dozen bad wins will add up to a bid.

Based on the factors in play, the NCAA tournament selection committee could pick any of the teams I've put in at the end of this process, or none of them. The resumes are simply so disparate this year that finding teams who have similar traits to compare is nearly impossible. (This is one reason for the smaller groupings above, which at least enabled manageable comparisons.) I do not envy the committee this year, but I grant them this: no team left out of this field is going to have a strong argument for inclusion. After about the first 30 at-large slots, everyone is wildly flawed, and if you couldn't be better than that group, you don't get to bitch to Greg and Seth, or whine to Digger and Jay. You had your chance.

Automatic bids (31): Duke (Atlantic Coast), Boston University (America East), Richmond (Atlantic 10) Belmont (Atlantic Sun), Connecticut (Big East), Northern Colorado (Big Sky), North Carolina-Asheville (Big South), Ohio State (Big Ten), Kansas (Big 12), UC-Santa Barbara (Big West), Old Dominion (Colonial), Memphis (Conference USA), Butler (Horizon League), Princeton (Ivy), St. Peter's (MAAC), Akron (Mid-American), Hampton (MEAC), Indiana State (Missouri Valley), San Diego State (Mountain West), Long Island (Northeast), Morehead State (Ohio Valley), Washington (Pac-10), Bucknell (Patriot), Kentucky (Southeastern), Wofford (Southern), Texas-San Antonio (Southland), Alabama State (SWAC), Arkansas-Little Rock (Sun Belt), Oakland (Summit), Utah State (WAC), Gonzaga (West Coast).

In (37): George Mason, Villanova, Georgetown, West Virginia, Kansas State, Missouri, Pittsburgh, St. John's, Cincinnati, Marquette, UCLA, Xavier, Texas A&M, Syracuse, Notre Dame, Purdue, Wisconsin, Nevada-Las Vegas, Tennessee, Temple, Texas, Louisville, Brigham Young, Arizona, Vanderbilt, Michigan State, Michigan, Illinois, North Carolina, Florida, Penn State, Florida State, Virginia Commonwealth, Clemson, Virginia Tech, Georgia, Colorado.

Last four placed in second round: Villanova, Florida State, Illiinois, Virginia Commonwealth
Last four in: Colorado, Georgia, Virginia Tech, Clemson
First four out: Southern California, Alabama, Harvard, Saint Mary's
Next four out: Boston College, Alabama-Birmingham, Missouri State, Texas-El Paso

I don't know how well the following is going to show up in this space. I took a run at seeding the field, taking an S-curve, trying to balance and avoid rematches and adhere to conference separation rules, but not taking into account geography.

After doing this for 40 minutes, I now have a raging meth habit. Joe, Andy, Jerry…love you, but you're nutbars.



A B C D
1 Ohio State Kansas Pittsburgh Duke
2 Texas Connecticut San Diego State Notre Dame
3 Florida Kentucky North Carolina Purdue
4 Louisville Wisconsin Brigham Young Syracuse
5 Temple Arizona Vanderbilt TAMU
6 West Virginia Kansas State Tennessee Xavier
7 UCLA George Mason St. John's Old Dominion
8 Cincinnati Richmond Utah State Georgetown
9 UNLV Villanova Missouri Michigan State
10 Penn State Marquette Michigan Washington
11 Memphis Florida State Gonzaga Illinois
12 Colo/VT Uga/Clemson VCU Butler
13 Princeton Belmont Oakland LIU
14 SPC Indiana State Bucknell Morehead State
15 Northern Colo Akron Wofford Boston U
16 UALR UCSB Hampton UNCA

UTSA Alabama State