How does your team’s bench affect their NCAA Tournament aspirations?
With the recent injury to Villanova’s Darryl Reynolds, the Wildcats were dropped to six regular rotation players and spot big man minutes from Dylan Painter. It served as my inspiration to take a quick look at a simple question — does a short bench hurt you in the NCAA tournament?
Villanova currently ranks 323rd per KenPom’s ‘bench minutes’ metric, and they’ll have seven rotation players in the tournament in the best case scenario. There are a somewhat surprising number of anticipated high seeds in a similar boat.
UCLA and Baylor are both in the bottom ten of NCAA Division I in bench minutes; Kansas, Duke, Notre Dame (and obviously Villanova) are all in the bottom fifty. How they’ve gotten there are all different, and it could be argued — especially in Duke’s case — the rotation will be deeper now, in tournament time. But the central question remains: do teams with short benches get tired in season, or more nicked up, and simply perform worse in the NCAA tournament?
It’s not a rare situation. When Wisconsin and Duke squared off in the 2015 championship game, it was a contest between two “bench poor” teams. On a squad led by Frank Kaminsky and Sam Dekker, Wisconsin’s bench minute percentage (21.2 percent) was among the 10 lowest in the country. Just seven players averaged more than 12 minutes per game, and only six of those cleared 15 minutes per game.
Duke relied a bit less heavily on the starters, playing them for only 67.9 percent of all available minutes. Their rotation was similarly short, though, especially after Rasheed Suilamon’s dismissal from the program. A six-player rotation, with spot minutes from Marshall Plumlee and Grayson Allen, carried them to the sixth championship in program history.
These are just two examples, though. As (almost) always, I attacked the question with data. The process was similar to the one I used when looking at how ‘extreme teams’ fared in the tournament (damn your premise-breaking offense, Louisville). I just linked up that analysis with KenPom‘s “bench minutes” metric, and used a combination of sports-reference and RealGM to define how many rotation players each historical tournament team had. The bench minutes metric only stretches back to 2007, so that served as the cutoff.
For a detailed explanation of expected wins, you can go to the bottom of the above ‘extremes’ article. Basically, one’s based on seed, and the other is based on team strength in simulations.
Rotation players were defined as all players who played more than 20 games and averaged 12 MPG. Heavy rotation players met the same criteria, but averaged more than 20 minutes per game.
The Short and long of it
A large-scale analysis of all tournament teams since the 2006-07 season found similar patterns no matter how I structured the analysis. Basically, shorter is better. Check out the table below, which aggregates past tournament teams based on the percentage of bench minutes. I used 25 percent as the cutoff for teams with short benches. It’s a round cutoff point that usually includes the bottom 50 or so of all Division I teams.
The sample on teams with short rotations is smaller — just 142 of the 640 teams analyzed ran with benches that small. But they’ve been quite a bit better relative to expectations, when compared to teams with longer benches.
Teams with bench minutes 25 percent and below, with an average seed of 8.3, have won between 22 and 24 more games than they’d be expected to. This is consistent whether we’re controlling for expected wins by seed or log5 simulations.
The same holds true if we do the same analysis by number of rotation players (players with 20 or more games played, and an average of 12 minutes per game). A table summarizing results is below.
The results are more muted, but we see a similar pattern — lower sample, better results than expected.
And it’s not just something seen in lower seeds. If I restrict the analysis to just protected seeds — the top 4 — the effect is similarly pronounced. Check out the table below, which summarizes the general numbers, before we dive a bit deeper.
While the sample remains smaller — about a quarter of all teams in the top 4 seeds, within the last 10 tournaments, ran with small benches — we see many of the same things. Teams with short benches continue to outperform their expectations based on seed or pre-tournament efficiency margins.
Thirty four of the 41 teams with short benches made at least the Sweet 16 in their respective tournaments (82.9 percent), while just 71 of 119 (59.7 percent) of those teams with heavier reliance on bench minutes accomplished the same feat.
As you’d expect, this same effect carried over to records as favorites or underdogs. “Favorite” or “underdog” was based solely on a team’s pre-tournament efficiency margin vs. their opponent’s. Short bench teams went 98-22 (0.817 winning percentage) as the favorite, and 13-17 as underdogs (0.43 winning percentage). Long bench? Fine as the favorite, going 237-74 (0.762 winning percentage), but terrible as the underdog, with an 18-38 record (0.32 winning percentage).
Why is this a good thing?
While I didn’t do much in-depth analysis to explain exactly what was going on here, I’ll take a reasonable, short stab. Having a deep bench benefits a team more during the regular season, being able to combat fatigue and injury with extra bodies is a real advantage. Over the course of a season, it’ll let a team pick off a game or three they may not have gotten otherwise. Those extra victories help seeding and positioning for the postseason.
But when it comes down to single-elimination games, who wins is most often about who’s putting the best players on the floor. Many coaches cut their postseason rotations down to 8 or less even if they have other options available. Much like the NBA postseason, depth and benches matter less; it’s your best vs. their best. Teams that earned high seeds with short rotations will likely have better top end players than teams with deeper rotations, almost by necessity. They managed to earn the same thing, without the marginal benefits mentioned above.
As long as a short rotation for a season doesn’t physically impact your rotation by the time you get there — through nagging or season-ending injuries — your chances certainly aren’t hurt in the Tournament. Based on the above, there’s even an argument to be made that your expectations should be higher. But mainly, I’d conclude that fans of teams with short benches — like Villanova, Kansas, Baylor, Notre Dame, and UCLA — shouldn’t worry that their teams can’t win with just 6 or 7 players. History shows they can do well in the tournament – and often do.