Duke owns best tourney record, but are they the top performer?

For more than 20 years, Duke has reigned as the winningest tourney school of the 64-team era. Despite their win percentage dropping for a third straight dance, the Blue Devils still own the best tourney record. In fact, the top six programs held on to their positions after the 2013 tournament. Here’s your top ten winningest schools:

  1. Duke – 82-24, .774 (1)
  2. North Carolina – 74-23, .763 (2)
  3. Connecticut – 48-15, .762 (3)
  4. Kentucky – 65-21, .756 (4)
  5. Kansas – 72-26, .735 (5)
  6. Florida – 38-16, .704 (6)
  7. Michigan – 33-15, .688 (9)
  8. Louisville – 43-20, .683 (12)
  9. Michigan State – 46-22, .676 (7)
  10. Syracuse – 47-23, .671 (10)
    OUT: UCLA (8 to 12)

Not surprisingly, the biggest advancer on this list is Louisville. By adding six wins to their totals, the Cardinals earned the eighth best tourney record, a jump of four positions. Michigan, their finals opponent, was the only other team to improve, moving from ninth to seven on the list of winners. Meanwhile, UCLA suffered the biggest decline. The Bruins’ opening-round loss dropped them from having the eighth to the 12th best record.

If you think tourney results are a better way to measure team performance, the rankings don’t change much. Duke is the only school with four championships in the 64-team era, so they’re the most successful team. North Carolina, Kentucky and UConn each own three tourney crowns, but the Tar Heels have made it to the Final Four nine times. They get the nod over the Wildcats (six Final Fours) and the Huskies (four). The only team with top-ten caliber tourney hardware that isn’t in the list above is Arizona. They’ve reached eight Elite Eights, made four Final Fours and cut down the nets once. Those are the ninth best team results—even if the Wildcats own just the 14th best winning percentage.

There’s one other important way to measure a team’s tourney performance. We touched on it in our November 13 post on coaching performance. Any evaluation of the top teams needs to factor in the bias that seeding confers on tourney outcomes. Put it this way: How can you tell whether Duke’s performance, when adjusted for seed bias, isn’t actually worse than some schools with solid records built from lesser seed positions?

For a decade, I’ve used a statistic called “Performance Against Seed Expectations,” or PASE, to compare a team’s actual winning record to its expected performance based on seeding. PASE is a simple concept. Every seed has recorded an average number of wins per tourney in the modern era. The average top seed wins 3.35 games per dance, two seeds win 2.41 (almost one game less), three seeds win 1.86 , four seeds 1.53 and so on. When a top-seeded team wins four games to reach the Final Four, they overperform by .65 games. But when a four seed does it, like Syracuse last year, they beat expectations by 2.47 games.

If you add up a team’s year-by-year performance, you can come up with the average games per tourney that they deviated from the expected victory total. Take Arizona, who I just got through saying had the ninth best results but only the 14th best winning record. How could that be? Well, the Wildcats have earned an average seed of 4.7. And based on that seeding, they should’ve won 47.7 games. But Arizona’s actually won just 46 games. In fact, their PASE value is a disappointing -.064 (the 1.7 game deficit divided by 27 trips). That’s a pretty weak PASE. Of the 69 schools with at least 10 tournament appearances, Arizona is just the 36th strongest performer against expectations.

So if Arizona is a shaky tourney performer, who’s an elite overachiever? Here are the top 20 active PASE performing teams in the modern era (out of the 69 qualifiers with at least ten appearances):

2014_Team_PASE

You’ll notice that many of the schools with the best records are not among the top PASE performers. The three teams whose ranking has suffered the most when factoring in the bias of seeding are Kansas, Duke and Syracuse. The Jayhawks are barely among the top 20 overachievers, despite owning the fifth best overall record. Winning percentage leader Duke owns just the ninth best PASE. And the Orange are just 16th on the PASE list. One of the reasons these schools don’t have elite PASE numbers is because they have a penchant for getting upset in the tourney. In fact, Duke is the most victimized school (upset 10 times) and Syracuse and Kansas are tied for second with Oklahoma at eight upsets.

It shouldn’t come as a surprise to see Butler solidly atop the list of overachieving schools. After all, they reached the Final Four once as a five seed and again as an eight seed. Those two runs alone put them 8.2 games above seed-projected win totals. Actually, the top ten on this list isn’t all that different from last year. The top four teams are the same in both years. North Carolina slipped from fifth to seventh on the list, making way for Louisville and Michigan. The Wolverines climb was the most dramatic of any school. They were just the 16th best PASE performer last year. But their run to the championship as a four seed boosted them 3.5 games above projections. That helped vault their PASE to +.440, good enough for sixth place on this list.

When it comes to overachievement dependability, however, the Tar Heels get the nod as the top school. See that column to the left of PASE titled “SOAR”? SOAR stands for “Seed Overachievement Rate.” It’s a simple measure of the percentage of times a team has beaten seed expectations in their appearances. The Tar Heels have defied seed projections 16 times in 26 tries, good for a 61.5% SOAR. UCLA is the second most frequent overperformer at 59.1% and Michigan State noses out in-state rival Michigan for third (56.5% to 56.3%).

This entry was posted in Team Ratings. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *

* Copy This Password *

* Type Or Paste Password Here *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>