Wednesday, March 16, 2016

"Explosiveness" of NCAA Men's Basketball High Seeds

With the annual NCAA men's basketball tournament getting underway, I wanted to apply a measure I originally developed for the Golden State Warriors, namely offensive "explosiveness," to the leading teams in March Madness.

Reuters news-agency blogger Chris Taylor contacted me a week ago, as part of his investigation of statistical tools that might inform March Madness predictions. I told him about explosiveness, which he included among his "Seven tips for crunching March Madness math." As Taylor characterized my explanation of why one might want to study explosiveness, "Winning teams need to be able to come back from behind, or pull away [in] close games." It remains to be seen whether explosiveness has any predictive power in this year's NCAA tourney. However, as I told Taylor somewhat flippantly, "It seems to work for Golden State!"  In the remainder of this posting, I discuss explosiveness in greater detail.

The explosiveness statistic measures high-scoring bursts in short stretches of time. Instead of looking at 12-minute quarters in the NBA, I looked at 6-minute "eighths" of games. An explosive burst in NBA play is defined as scoring 18 or more points in 6 minutes (3 points per minute), which if maintained over an entire 48-minute game would yield a whopping 144 points. In the Warriors' first 25 games of the present season, in nearly one-fourth of all their 6-minute blocks did they register an explosive burst.

For the 40-minute length of college-basketball games, I looked at 5-minute blocks (one-eighth of regulation game-length) to see how often teams scored 15 or more points (i.e., 3 points per minute). Due to time constraints, I analyzed only the top eight projected teams in the field (i.e., all the No. 1 and 2 seeds in the bracket) and looked only at each of these teams' final 10 regular-season games. I did not include overtime periods, so each team had a total of 80 5-minute blocks. As shown in the following table (on which you can click to enlarge), the eight teams varied greatly in their explosiveness.

As can be seen, Xavier (Ohio), the No. 2 seed in the East region, was the most explosive team among those studied. Playing in the Big East Conference, the Musketeers recorded 13 explosive (15-point or more) bursts in their 80 5-minute blocks. Here are several examples:
  • In one game alone, vs. Creighton on March 5, the Musketeers pulled off three bursts (play-by-play sheet). Xavier went from having 44 points at halftime to 60 points with 15:00 min remaining; from 66 points with 10:00 left to 82 with 5:00 to go; and 82 with 5:00 left to 98 points at the end. 
  • At Seton Hall on February 28, a game in which the Musketeers trailed 45-26 at the half and ultimately lost 90-81, Xavier rallied feverishly in the second half. XU went from 43 points with 10:00 left to 58 with 5:00 left; and from 58 with 5:00 left to 81 at the final buzzer, a 23-point super-explosion (play-by-play)

Kansas (No. 1 seed in South region) and Oregon (No. 1 in the West) followed with nine explosive bursts each. Oklahoma (No. 2 in West) surprised me with only four bursts, given the excellent three-point shooting this season by the Sooners' Buddy Hield. Virginia (No. 1 in the Midwest) had no explosions in its final 10 regular-season games. 

I next inquired into what other skills and styles of play might contribute to a team's explosiveness (or lack thereof). The first things that occurred to me were that an explosive team was likely to play at a fast tempo (i.e., shooting early in the shot clock, generating many possessions per game) and be good at shooting the three. Results only partially supported these hypotheses (see the grey columns in the chart above). Note, however, that whereas explosiveness was only measured in teams' final 10 regular-season games (because combing through play-by-play sheets is time-consuming), all the other statistics are based on teams' full seasons.

Xavier indeed plays at the fastest pace  (74.5 possessions per game) among the No. 1 and 2 seeds, corresponding to the Musketeers' explosiveness. However, Oklahoma plays nearly as fast (73.7 possessions per game), but had low explosiveness. Virginia plays at a very slow pace (62.4 possessions per game, dead last among the 351 Division I men's teams), which seems to go a long way toward explaining the Cavaliers' lack of explosiveness. (Tempo/possession statistics are available here.)

Kansas, with nine explosive bursts, thrives on the three-pointer. Not only do the Jayhawks have one of the nation's highest shooting percentages from behind the arc; they also are attempting more treys than they have in the past. (All shooting, rebounding, and defensive statistics cited here are from the NCAA statistics webpage.)

Michigan State, with seven explosive bursts, is first in the nation in two categories: three-point shooting percentage and defensive rebounds per game. Limiting opponents to one shot and making shots yourself should contribute to explosiveness. However, the Spartans are quite low in turnovers forced per game, which presumably works against MSU being able to score quickly.

Oklahoma, which shoots well, cleans the defensive glass, and plays relatively fast, remains an enigma.

Keep an eye out for whether the eight teams seeded No. 1 or 2 make it to the Final Four. Either the explosiveness statistic will go out with a bang or be a dud.

No comments: