Of all the Bracketology articles I wrote last year, none stirred more controversy or drew as many fiery e-mails as the one rating the performances of coaches in the tourney.
The statistical approach of comparing a coach's actual winning record against his expected performance based on seed position elevated the stature of some lesser-known coaches and deflated the ratings of a few high-profile legends. After last year's results, the same statistical analysis is certain to fuel debate, while a couple of new performance measures expose the surprising strengths and weaknesses of a different set of coaches.
One ranking that stayed fairly consistent after last year's tournament is the list of top tourney coaches by overall record in the modern 64-team era. As the table below illustrates, the top four coaches with more than five tourney appearances held their positions, though Mike Krzyzewski's lead over Rick Pitino and Tom Izzo tightened up due to the combination of his quick exit and their Final Four showings.
Winning the 2005 tourney certainly helped Roy Williams' statistical fortunes. He leapfrogged Jim Calhoun and Tubby Smith in the rankings. Gary Williams also moved up two slots -- just by sitting on the sidelines. Jim Boeheim's early exit and Bill Self's shocking loss to Bucknell dropped them both below Williams -- and dropped Self right out of the top 10. His replacement is Lute Olson, whose Wildcats' solid 2005 run helped pull him out of statistical purgatory. (Before all you Arizona fans fire off any e-mails, read on; good news is coming for Lute.)
Coaching overachievers
The rankings for coaching performance against seed expectations show more dramatic changes. Here's a quick recap on how this stat is figured: Over the last 21 years, each seed position has posted an average number of wins per tournament. No. 1 seeds win 3.35 games each tourney, No. 2 seeds win 2.40, No. 3 seeds win 1.76 and so on.
When the coach of a top-seeded team wins only two games, like Krzyzewski did last year, he underperforms by 1.35 games. If you tally the number of games a coach overperformed or underperformed against his seed position in each tourney appearance, you can come up with the average games per tourney that he deviated from the expected victory total. Take Steve Fisher, the top tourney overachiever. In his eight appearances, he should've won 11.6 games based on seeding. He actually won 20 -- about 8.4 games better than expectations, or 1.05 games per tourney.
As it turns out, Fisher is the only coach with more than five trips to the Dance whose performance against seed expectation (PASE) ranking remained the same after the 2005 tourney (see below). Roy Williams was the big advancer. By running the table, he increased his overperformance from .367 games per tourney to .512 and climbed from 10th place to fifth. Tubby Smith was the only other top 10 coach to advance two positions. His Wildcats' Elite Eight run pushed him into the top 10.
In statistical terms, Tom Izzo is the biggest advancer, having increased his PASE by .251 games to nearly one full game. Considering that Fisher's performance is built on the 1989 championship when he relieved Bill Frieder and the exploits of the tarnished Fab Five -- plus the fact that his San Diego State Aztecs aren't exactly tourney fixtures -- a strong case could be made that Izzo is the legitimate leader in tourney overachievement.
Bill Self was the biggest decliner, dropping from fifth place last year (+.615 games per tourney) to 13th place (+.269). Coach K and Jim Calhoun were also significant decliners. Their underperformance last year against seed expectations dropped them each two positions.
Coaching underachievers
For every coach who coaxes more wins out of his team than seeding would dictate, there's a coach who falls short of expectations. Here are the bottom 10 active coaches with more than five appearances who've performed the worst in the tournament.
Danny Nee and Dana Altman remain stuck in the top two PASE underachiever slots (that is, the two worst slots). Altman's first-round exit with 10th-seeded Creighton nudged him closer to Nee, but he'd need a couple more crater-like performances to match Nee's winless record of underachievement. The biggest decliner in 2005 was Mark Gottfried. After his eighth-seeded Crimson Tide squad overperformed in 2004 with three wins, Gottfried backslid at the No. 5 seed with an opening-round exit. Skip Prosser, Kelvin Sampson and Rick Barnes also dropped positions due to subpar tourney runs.
Bobby Knight nearly climbed out of the bottom 10 underachiever ranks -- and would have done so if Gene Keady and Bob Huggins hadn't gone into retirement. His sixth-seeded Red Raiders' Sweet 16 run raised his underachievement from .301 games per tourney to .245.
Upstart coaches on the rise
The top overachieving veteran coaches better not rest on their laurels. There's a group of upstart coaches waiting to knock the legends off their perch. Here are the top 10 coaching overachievers with three to five trips to the Dance.
If these upstarts were combined with the list of veteran coaching overachievers, the top five names here would replace Penders, Gary Williams, Calhoun, Smith and Boeheim. But that's a big if. The upstart coaches can't slip up in their next few tourney appearances. One mediocre run and they could fall off the list altogether. Consider Paul Hewitt: After his Yellow Jackets made a seed-defying run to the finals in 2004, they won only one game as a No. 5 seed in 2005 -- a respectable showing, but it still dropped Hewitt's PASE from a +.754 to +.535.
On the other hand, the upside for inexperienced tourney coaches is bigger than it is for veterans. Look at Bo Ryan and Bruce Weber. Last year, Ryan barely made it into the top 10 list of overachieving upstart coaches. With Wisconsin's Elite Eight run from the No. 6 seed, he climbed seven positions to third place. Weber vaulted from the fourth position to the top of the list by overperforming as the top seed. It's also worth noting that John Beilein was nowhere to be found on last year's overachievement list. But the seventh-seeded Mountaineers' Elite Eight run made him the top upstart coach in exceeding seed expectations.
What have you done for me lately?
The dynamics of the NCAA Tournament are different now than they were in the early years of the 64-team era. Scoring is down, guard play has taken on greater importance and upper-class-dominant teams now underperform. These are just some of the changes that have taken place over the last decade. It's worth examining whether they've brought about corresponding changes in the fortunes of coaches. One thing is for sure: The list of top-seed overachievers is dramatically different. Take a look at the top 10 overachieving coaches for the last decade (1996-2005) with at least four tourney appearances:
Notice anyone missing? Surprisingly, Coach K doesn't show up in the list of top overachievers for the last decade. As mentioned in the team article last week, since 1996, Coach K's Blue Devils actually have underperformed since 1996, winning 27 games when seed position dictated that he should've won 28. Granted, Krzyzewski has been saddled with the burden of seven top seeds -- and 23 wins -- among his 10 appearances. It's hard to build a record of overachievement when you need to reach the Final Four to exceed expectations. But his performance as a top seed hasn't been the problem. He's actually won 24 games as a No. 1 seed, a game over expectations. It's his performances as a No. 2 and No. 8 seed that have hurt him. He should've won three games in those two tourneys and only won one.
Coach K's record of overachievement was built during the first 11 years of the tourney, from 1985 to 1995. Considering his seed position, he should've won about 23 games; he actually won 39. That works out to an incredible 1.6 games above expectations per tourney. Of the 11 coaches who've gone to the Dance at least four times in both the first 11 years and over the last decade, Krzyzewski is clearly the top overperformer from '85 to '95. Here's how those 11 coaches rank:
Just as Coach K was the undisputed overachiever in the first 11 years of the tourney, Lute Olson was far and away the biggest underachiever, having fallen more than eight games below seed expectations in 11 years. But here's the good news, Wildcats fans: While Krzyzewski has underperformed over the last decade, Olson has overperformed. By winning 25 games when he should've won about 21 -- good enough for the sixth best seed-defying record among the coaches above -- Olson reversed his fortunes from the first 11 years by an average of 1.12 games per Dance. That's the best turnaround among the 11 coaches above (closely followed by Jim Boeheim's 1.10 games per appearance).
Overachievement as a percentage closer to the championship
Thanks to the sage advice of reader Sean Kates, we're introducing a new analysis of performance that quantifies how much closer a coach got to the goal of a championship than his seed position would dictate. Sean rightfully pointed out that the tournament is a closed universe. Since a top-seeded team wins an average of 3.35 games, it can only exceed seed expectations by 2.65 games. If a No. 1 seed wins five games -- 1.65 games above average, you could say it improved on expectations by 49 percent (1.65 additional games divided by the 3.35 average). But this type of calculation gives an inflated advantage to lower-seeded teams. Consider that a top seed which wins the tourney only nets a 73 percent improvement over typical No. 1 seeds. Meanwhile, a No. 14 seed that wins one measly game nets a 426 percent increase over expectations.
The more reasonable approach is to divide the number of games a team actually wins exceeding its seed expectations by the number of games it would've had to win to take the championship (based on its seed expectations). In this way, a top seed that won the tourney would rate a 100 percent improvement toward its goal of winning the championship (2.65 extra games won out of the possible 2.65 games it could've won exceeding its seed expectations). Using this metric, the No. 14 seed that wins one game would rate only a 14 percent improvement toward the championship (0.81 extra games won out of the possible 5.81 games it could've won). Yes, this calculation is particularly hard on lower seeds -- after all, who would consider the first No. 16 seed to pull off a round-one upset to be less impressive (16 percent closer to the championship) than a No. 1 seed that reaches the Final Four (25 percent closer to the ultimate prize)? But, as a couple of readers pointed out, the goal of every team in the tourney is to win the whole enchilada. The teams that go farther ought to be considered more successful, no matter what their seed position.
In that light, here are the top 10 coaches with more than five tourney appearances who've moved closer to a championship than their seeding would dictate:
The ranking isn't much different from the list of top 10 overachieving coaches by PASE, except that Izzo overtakes Fisher and Penders, who drops below Calhoun and Gary Williams. If we apply the "what have you done for me lately" rule and look just at the last decade, Pitino has overachieved by the highest percentage toward the big prize (27.8%), followed by Izzo (24.2%), Roy Williams (22.5%), Calhoun (20.1%) and Weber (20.1%). The only substantial shakeup in this order from wins per tourney above seed expectations is the movement of Calhoun from sixth to fourth and the fall of Weber from third to fifth.
Freelance writer Pete Tiernan has been studying the NCAA Tournament for 16 years. E-mail him here.