Around this time of year, those in the world of college basketball like to claim that “it’s hard to beat a team three times.” Coaches use the line as a truism, and it often spreads to their players. In his memoir, Dean Smith wrote, “It was extremely hard to beat a good team three times in a row.” As most conference tournaments ramp up this week, you’ll hear that line again. After his Michigan State team beat Maryland on Sunday to set up a rematch in the Big Ten Tournament on Thursday, coach Tom Izzo said, “Now you’ve got to beat a team three times, but it is what it is.”
On its face, the argument that the third win is tougher than the first two against the same team seems a bit questionable. In the NFL, for instance, the data shows if one team wins the first two matchups, then winning a the third straight meeting is no more difficult than winning each of the first two. But men’s college basketball’s coaches have a point. Historically, the high variability of college basketball outcomes has made two-time winners a bit shakier than usual in the third matchup between two teams.
We tested the adage by looking at all games from 2001-02 through last season. We separated the games into bins according to the difference between the final adjusted efficiency ratings on KenPom.com of the teams involved, in increments of 3 points per 100 possessions. We found that teams with the slightest advantage — those that ended up with an edge in adjusted efficiency of less than 3 points per 100 possessions — won about 54.5 percent of all matchups, no matter when they happened during the season. In the first meeting between any two teams, regardless of whether they played again, the slightly better team won 54.7 percent of the time; when teams that played before met a second time, the better team’s chances went down just slightly, to 54.1 percent.
But when two evenly matched teams meet for a third time in the same season, their previous results matter. When the teams split their first two games, the better team won 57.3 percent of the time; when the better team lost the previous two meetings, that team avenged those defeats at a 56.7 percent clip. But when the better team won both previous matchups, the third was a true toss-up — the better teams won only about 49.3 percent of the time.
Those trends hold up across each of our bins. Teams with edges of at least 3 but less than 6 points in adjusted efficiency won about 63.3 percent of third meetings when they came off two wins against the same opponent, 66.3 percent when they came off a split and a whopping 74.3 percent (109 out of 190) when they lost both previous meetings.
With an edge of at least 6 but less than 9 points in Pomeroy’s ratings, the better teams won 72.5 percent of the time overall, 79.6 percent after a split and 84.8 percent after two losses. After that, the sample starts to thin — there aren’t many cases of teams being swept in the regular season by significantly inferior opponents.
By the same token, there aren’t many teams with a strong track record of being swept in the regular season and then reversing those results in the postseason. But a few teams seem particularly adept in the art of revenge. Duke, for instance, has fended off the only two teams that have tried to vanquish it three times in one season.30 Michigan State has won four of five such chances, far exceeding the national 29 percent success rate for teams avoiding sweeps.
The champion in that regard, though, is Winthrop, whose penchant for revenge spans 10 seasons and three coaches. Five times between 2006 and 2015, the Eagles ran into a team that beat them twice in the regular season, and they won all five third rounds on their way to four NCAA Tournament berths.
The peculiar pattern of beaten teams coming to life in the postseason seems to lie in the high variability of single college basketball results. Consider, for instance, the 2002-03 UCLA Bruins, who entered the Pac-10 Tournament at 9-18. They had suffered two embarrassing losses to Arizona in the regular season — 87-52 at home and 106-70 in Tucson. When they drew the Wildcats one more time in the postseason, Arizona had already locked up a No. 1 seed in the NCAA bracket, while the Bruins were playing for coach Steve Lavin’s job.
UCLA fell behind by 15 in the second half; the Bruins’ Jason Kapono said afterward, “I’m sure our fans thought, ‘Oh no, here comes that 35-point blitz.’” But Lavin’s team pulled off a remarkable comeback, rallying to win 86-79 in overtime. It wasn’t enough to save Lavin’s job — UCLA lost to Oregon the next day, and the school changed coaches after the season — but for 45 minutes, the Bruins overperformed. No team in the past 20 seasons has delivered a greater swing than minus-71 in the regular season to plus-7 in the third matchup.
Whereas Arizona coach Lute Olson said after UCLA’s upset win, “Everyone knows I’m not a proponent of [the conference tournament],” Smith wrote in his memoir, “We planned all year long for the ACC Tournament. Everything rested on those three days in March.” He even went so far as to hold back defensive sets until the conference tournament in order to surprise opponents. Smith won 13 ACC Tournaments in 36 years at North Carolina.
Some teams entering conference tournaments this week find themselves fighting for NCAA Tournament bids. Creighton and Xavier, for instance, may each need to complete the tricky 3-0 sweep against Marquette and Butler, respectively. Wake Forest and Miami, which are both on the bubble, should play each other Thursday in the ACC quarterfinals, after Miami won both regular-season meetings. And TCU and Oklahoma will try to avenge regular-season sweeps against Texas and Baylor, respectively, which have already secured trips to the Big Dance. Heading into Thursday’s game, TCU has made it no secret where the extra motivation comes from. “[Texas] has beat us twice,” TCU guard Mike Miles said, predictably, “but it’s hard to beat a team three times.”