Tuesday, November 20, 2007

Todd Bozeman? Really?

Ken Pomeroy had a nice article today evaluating last year’s new coaches. I came up with my own coaching metric last spring, and after reading his article, I’m reminded what a difficult task it is to generate a true coach ranking.

Pomeroy’s coach ranking (actually his adaptation of a Dean Oliver idea) highlights my biggest problem with my own rankings. Within any one year, it is clear that the best coaches are the coaches that take an unsuccessful program and win more than expected. In fact, this sounds terrific, until you evaluate the flip side. Coaches that maintain a level of excellence can never rise to the top of the list.

For example, over the last 5 years, I essentially rated Ohio St.’s coach Thad Matta as the top coach in college basketball. That is because he led to a significant improvement at three straight schools. This sounds good in principle, but did it really prove that Thad Matta was a better Coach than Billy Donovan? Billy Donovan won two championships but mostly just maintained a program at the highest possible level.

To put it another way, if you lived to enjoy the end of UCLA’s glory days, was your reaction, “Eh they should win every year”, or “There goes the greatest coach of all time”? I think the answer has to be the later, and if that’s true, you really cannot overlook the most important metric of all, victories.

After all, if you put Anthony Grant at North Carolina, I’m not sure he would win the ACC or compete for a National Title. But, I’m sure that if you put Roy Williams at VCU, he would make an impact. Just look at what John Calipari did after he returned to college to coach at Memphis. From his Memphis Bio,

“Although the Tiger program wasn't as far down as UMass was when Calipari arrived, even the most loyal Memphis basketball supporters would admit their beloved Tigers had fallen on hard times in the mid-to-late 1990s. From 1993-94 through 1999-2000 (seven seasons), Memphis posted only two 20-win campaigns (1994-95 and 1995-96), and the Tigers had consecutive losing seasons in 1998-99 and 1999-2000.”

But now, “Calipari's 181 victories and 25.9 wins per year are the most by a Tiger mentor in his first seven seasons.”

Clearly a big name coach can step into a mid-level school and raise them to prominence. So for that reason, I think any reasonable measure has to reward success as a whole, not just improvement.

A Silly Criticism

To be fair, Pomeroy knew this wasn’t a tool to evaluate all coaches, and that’s why he looked at only new coaches in his article. Moreover, he also pointed out that new coaches often struggle because they inherit difficult situations. And I’m sure if you pushed him further, he would agree that there are other important factors to consider when evaluating new coaches.

For example, look at how the improvement metric would compare the new coaches Bob Huggins (now at West Virginia) and Frank Martin (now at Kansas St.). If they have equal success this year, this type of metric should rate them the same. After all, both programs were borderline NCAA teams last year. But since Frank Martin runs the same system as Bob Huggins, and because his recruits were already in place, if Bob Huggins takes West Virginia as far as Kansas St. this year, I’ll consider it a minor miracle.

In the day and age when Moneyball fanatics like to think that statistics can solve everything, including removing the need for scouting, I’m pleased to note that a regression can’t observe everything. But its still kind of fun to look. After all, why else would we have a reason to discuss Todd Bozeman?

Three Other Problems with My Rankings and Other Coach Rankings

1) Let’s assume for a moment that turning a program around will always be more impressive than maintaining a program. Then, what window do you use to measure that success? As Pomeroy said, new coaches often inherit situations where the program has bottomed out. If a team is without talent, even a top coach may not be able to get that team to win immediately. This suggests a longer time window is appropriate. It suggests we should look at coaches after five years, and see whether or not they have improved their programs. But a longer time window is also not without its downfall. After all, Roy Williams looks like a miracle maker over a short-time horizon, taking Doherty’s disappointments and turning them into National Champions. But relative to the long history of North Carolina, he’s only maintained the tradition. While Calipari clearly revitalized Memphis, the school had a tradition of success dating back to a 1973 title game against UCLA.

2) Is it easier or harder to win at a mid-major or a major school? For example, would you be more impressed with Roy Williams maintaining success at North Carolina or Mark Few maintaining success at Gonzaga? Yes, North Carolina has more tradition and a larger fanbase to fall back on, but Gonzaga faces fewer hurdles in conference play in the WCC.

3) Is it the coach or the program? The biggest problem in answering all these questions is that we don’t have enough variance in the coaching data, (we don’t have enough coaches switching between jobs) to truly identify the ability of any one coach. I think this is particularly true when it comes to recruiting. We don’t know how much the recruiting comes from the school and how much it comes from the coach. In some cases (think Duke), the two are so intertwined that it would be pointless to try to sort it out.

Of Course

So who is the best coach? I guess the reason I created my rankings is to ask the question “Best at what?” Best at Recruiting? Best at Development? Best in the Tournament? And if I’ve already bored you with this level of detail, just take the easy way out. Say John Wooden.

Tournament Action

Michigan St.’s big lead evaporated Monday night against Missouri, but they hung on for a victory. I’m sure if John Gasaway was still writing a daily column he’d discuss the fact that Michigan St.’s 17 turnovers were not terrible, particularly given last season’s MSU turnover rate and Missouri’s defense which is designed to create turnovers. But the obvious effect of all that defensive pressure was a lot of easy baskets. Michigan St.’s starting front court all shot over 50% and two of the players Gray and Morgan were perfect from the field.

The Big Ten success continued elsewhere late Monday as Illinois built a huge lead on Arizona St. Sadly for those of us Illinois fans hoping to watch the game, the large run to start the game was largely interrupted by bonus coverage of the Michigan St. vs Missouri. Not that I’m bitter or anything.

Today’s games were mostly less inspiring.

Maui: Chaminade held tough with LSU but lost in the consolation bracket and Princeton looked tired against Arizona St. in the other semifinal. Marquette on the other hand looked fresh and dominant knocking down 11 of 16 three pointers in the victory over Oklahoma St. This of course led to the incoherent rant from Doug Gottlieb about how Oklahoma St. cannot stop penetration by guards. Uh dude, I’m sure that was important and all, but that would have been a much better talking point if Marquette’s guards hadn’t just hit 11 three pointers. Just saying.

CBE Classic: Missouri outlasted Maryland, and Michigan St. vs UCLA is just kicking off.

Tommorrow Night we have the finale of the Maui Invitational, the NIT Semis, the start of the Great Alaska Shootout, and a lot of people sitting in airports.