Last Thursday, I started a "Bracket Science 101" blog series to get everyone up to speed on some basic principles for building a better bracket. On Selection Sunday, CBSSports.com will produce its "Bracket Lab" for the second year. The tool's historical information comes from a database that I've put together on the 28 years of the 64-team tourney era. These blog posts should make it easier for you to work with the Bracket Lab once the 2013 tourney field is set.
Our last Bracket Science 101 lesson centered on the concept of PASE, or performance against seed expectations. PASE is like baseball's WAR metric, comparing performance to the average wins of a "replacement" seed.
Let's use Duke coach Mike Krzyzewski as an example. Based on accomplishments, there's no denying that Coach K is the top coach of the 64-team bracket era. He has notched more victories (79), built the highest winning percentage (.775), been to more Final Fours (11) and won more championships (four) than any other coach. Extraordinary numbers -- but do they necessarily make Krzyzewski the top performing tourney coach?
Considering that Coach K has been among the top three seeds in 24 of his 27 trips to the dance, seeding has preordained a reasonably strong record. After all, the typical top seed has a .799 winning rate -- basically averaging a Final Four run. So how can you tell whether Krzyzewski's performance, when adjusted for seed bias, isn't actually worse than some coach with a solid record built from lesser seed positions?
We created the PASE statistic to answer just such a question. Based on seeding, Coach K should've won about 70 games. The fact that his Blue Devils have actually won 79 means that he has beaten expectations by almost nine games. If you divide that by their 27 appearances, Krzyzewski's PASE works out to a solid +.322.
Does that make Coach K the top overachiever of the 64-team tourney era? Five years ago, the answer would've been yes. But his Blue Devils have struggled to live up to seed-projected win totals in seven of the last eight dances. That has dropped him all the way out of the top 10 PASE list, to 14th place among 97 active coaches with at least three tourney trips. The top overachieving coach is actually Brad Stevens. Based on seeding, he should've guided Butler to just 3.3 victories, but his Bulldogs have actually won 11 games. Stevens' PASE works out to an astounding +1.938 -- nearly two games above expectations for every dance that he has entered.
Coaching performance isn't the only factor you can measure with PASE. In fact, you can analyze the degree to which any attribute contributes to over- or underachievement against expectations. Let's look at the top-five performing coaches and teams, then break down the 10 characteristics with the highest PASE in the modern tourney era.
Top Five Overachieving Coaches
We already know where Coach K and Brad Stevens stand in terms of performance against seed expectations. Here are the four coaches that trail Stevens in the PASE race:
The gulf between Stevens and Michigan State coach Tom Izzo is staggering. Much of it, however, has to do with the difference in overall appearances. Stevens has only been to four dances, but two of them have been historic runs. As a five seed in 2010 and an eight seed in 2011, Butler was projected to win only two games -- it won 10, reaching the final both years. Call me crazy, but I don't think Stevens can sustain that sort of mind-boggling overachievement. All it would take is one or two underperforming runs, and that lofty PASE would plummet in a hurry. Look no further than Coach K for an object lesson. After his first eight dances, Krzyzewski owned a +1.765 PASE. And look where he is now. Years of March Madness tend to leave scars.
That's why Izzo's performance shouldn't be overlooked. In 15 tourney trips, he has beaten expectations at more than a three-quarters-per-game clip. Just as importantly, Izzo has been a reliable overachiever, exceeding seed-projected win totals 10 times in 15 tries. That's what the SOAR statistic at the far right of the table reflects. If PASE measures the degree of tourney overperformance, SOAR, or "seed overachievement rate," measures the frequency. No coach with 10 or more trips has beaten expectations with greater regularity than Izzo's 66.7 percent.
Arizona coach Sean Miller is right behind Izzo in the PASE race and actually has a higher SOAR, defying seed expectations four times in five tries. But like Stevens, Miller has a thin tourney resume. That's not the case with the fourth- and fifth-best performing coaches. Florida coach Billy Donovan and Louisville's Rick Pitino have both built their records of overachievement on more than 10 dances. And they've owned the highest average seeds of the top five as well, so more was expected of them.
There's a dark side to the PASE metric. For as many coaches who own positive PASE values, there are nearly the same number who are saddled with negative numbers. Of the 97 active coaches with at least three tournament appearances, 48 have an overachieving PASE, 47 are underachievers-and two (UTEP's Tim Floyd and UNC-Ashville Eddie Biedenbach) have played exactly to expectations.
So who are the five biggest underachievers? DePaul coach Oliver Purnell owns the worst PASE (-.818), followed by Colorado State coach Larry Eustachy (-.665), Pitt coach Jamie Dixon (-.567), Creighton's Dana Altman (-.515) and Clemson's Brad Brownell (-.446).
Top Five Overachieving Teams
Just as there as coaches who have shown a propensity to defy tourney expectations, the same goes for teams. In some cases, a school's entire tourney history is defined by a single coach (think Duke and Coach K). But more often, a group of coaches contributes to the overall performance of a school. Iowa State, for instance, has been led to the dance by five coaches-Johnny Orr, Tim Floyd, Larry Eustachy, Wayne Morgan and Fred Hoiberg.
So which schools are the top five performers against seed expectations? Here's the list:
You might think that Butler's ranking atop the team PASE list is all due to Stevens. But Thad Matta and Todd Lickliter are the coaches who turned the Bulldogs' tourney fortunes around. They were actually -.313 PASE underachievers in their first three trips under Barry Collier. Matta and Lickliter put together a +.890 PASE in the next three appearances. Then came Stevens.
Florida's tournament record isn't all Donovan, either. Norm Sloan and Lon Kruger coached in five of the Gators' 17 appearances and put together a +.289 PASE. Connecticut, on the other hand, has Jim Calhoun to thank for its entire +.515 PASE. If he hadn't retired, Calhoun would be eighth on the PASE list of active coaches.
Kentucky's record of overachievement is an interesting case. Six coaches have contributed to the Wildcats' performance. Here's how each of them performed:
- Joe B. Hall -- one trip, +1.482 PASE
- Eddie Sutton -- three trips, -.497 PASE
- Rick Pitino -- six trips, +.704 PASE
- Tubby Smith -- 10 trips, +.309 PASE
- Billy Gillispie -- one trip, -.545 PASE
- John Calipari -- three trips, +1.586 PASE
Aside from the Sutton era and the Gillispie hiccup, Kentucky has been a consistent overachiever throughout its 24-year tourney history. And the Wildcats have been good to their coaches, too. Pitino, Smith and Calipari have had their best stretches of overachievement while coaching Kentucky. In fact, if you back out Calipari's Wildcat years, he's actually a -.016 PASE underachiever.
Roy Williams is in the same boat as Calipari. He has been a huge part of North Carolina's tourney success, logging a +.585 PASE with the Tar Heels. But before he came to Chapel Hill, Williams had a +.053 PASE that was barely over seed expectations.
We told you whom the biggest coaching underachievers were. Who are the worst performing teams with at least 10 tourney trips? The bottom five are: Pittsburgh (-.565 PASE), Georgia (-.516), Clemson (-.478), Stanford (-.430) and New Mexico (-.425).
Top 10 PASE Attributes
When it comes time to fill out your brackets, it will be helpful to know which coaches and teams have a propensity to overachieve and which tend to fall short of expectations. But PASE analysis can be applied to any factor. In fact, at Bracketscience.com, we let you do queries through our Bracketmaster database tool on any combination of 18 statistics. Want to know whether Big Ten teams with a top-20 strength of schedule and an All-American are overachievers? You can find out. (They're actually +.961 PASE overperformers.)
I analyzed more than 35 factors under hundreds of conditions and came up with the top-10 characteristics that correlate with overachievement in the dance. I restricted the analysis to the top 12 seeds, since they're the real advancers in the dance -- and the lower four seeds tend to dilute PASE values.
Here are your top 10 PASE attributes:
The factor that leads to the highest degree of overachievement among 1-12 seeds is scoring margin. The 120 teams that have come into the tournament with an average margin better than 15 points have a strong +.412 PASE and get to the Final Four 38 percent of the time. That bodes well for high-margin teams like Indiana, Florida, Syracuse and Michigan. On average, there are usually just a handful of squads that meet this condition.
Not surprisingly, coaching success -- not just experience -- is the second-highest overachievement factor. Coaches who have made more than four Elite Eight runs are +.242 PASE overperformers. Only eight coaches meet this criterion: Coach K, UNC's Roy Williams, Pitino, Calipari, Izzo, Kansas coach Bill Self and Donovan.
The third biggest sign of overperformance is turnover margin. Teams that win the turnover battle by more than two possessions on average are +.186 PASE performers. And they own the best SOAR of any of the top-10 factors. More than 56 percent of the squads with this attribute beat expectations. Which teams figure to be high turnover generators come tourney time? Right now, Virginia Commonwealth and Louisville force the highest percentage of turnovers in D-I.
Here are some comments on the rest of the top 10 factors:
- Free-throw percentage greater than .728 -- Amazing that free-throw shooting has a stronger correlation to overperformance than any of the other shooting metrics that I analyzed, better than overall field-goal percentage or 3 accuracy.
- Starters' percentage of scoring less than 72.5 percent -- This is a confirmation of the power of a deep bench.
- At least one All-American -- More than half the Final Four teams and 22 of 28 champions have had an All-American on their squad. Right after Selection Sunday, the U.S. Basketball Writers Association announces its top 10 players. That's the list that I always use.
- Front-court percentage of scoring more than 62 percent -- So much for the myth that you need a great guard to go far in the dance. It's the teams with dominant front courts that defy seed expectations in the dance. Twelve champs have relied on centers and forwards for more than 62 percent of their points. Only two champs have relied on guards for that much scoring.
- Rebounding margin better than 5.4 possessions -- I'm not surprised that rebounding makes the top-10 list, but I thought it would've been higher.
- At least one freshman and fewer than two senior starters -- If you're tempted to pick a team just because they've got an experienced starting five, resist the urge. Teams with younger starters actually outperform older squads, more so since the NBA one-and-done rule.
- Defensive efficiency of fewer than 90 points in 100 possessions -- The KenPom stats were reliable signs of overachievement across all the various studies that I did. Only defensive efficiency, however, made the top 10.
These are the top 10 overachievement factors for all the 1-12 seeds. But not all seeds are created equal. The attributes that a one seed needs to beat expectations and make the Final Four are very different from the attributes a 12 seed needs to spring a single upset. If you're interested, I'll have a complete breakdown of the top overachievement factors for five classes of seeds on Bracketscience.com in the coming weeks.
As for the next Bracket Science 101 course, let's look into the factors that separate Final Four contenders from pretenders.