It’s championship week in men’s college basketball, the last chance for schools to make an impression on the selection committee before the NCAA tournament field is announced Sunday evening. If you’re a bubble team, you’re hoping a strong showing in your conference tournament can persuade the committee to put extra weight on those games, perhaps to the point of ignoring a previously spotty track record.
Figuring out how much weight to give a late hot streak — as opposed to a team’s season-long résumé — can be tough for the committee. Fans filling out their brackets face the same decision. Conference tournaments might be the first chance you’ve had to see some schools play; a deep conference run could make the difference between marking a team down for an early exit or slotting it into the Final Four. But should it? Or is it simply recency bias to think a breakout conference tourney performance matters in the NCAA tournament?
To take a few preliminary stabs at answering that question, I computed pre- and post-game Simple Rating System (SRS) scores for every conference tournament and NCAA tournament game since 1985 (when the NCAA field expanded to its familiar 64-team format), using them to establish each team’s expected win probability1 going into a given game.
If teams that had surprising conference tournament runs (relative to their pre-tournament ratings) tended to carry that magic over into the NCAA tourney, we might expect there to be a relationship between how many “extra” wins a team had in each tournament. Take the 2010-11 Connecticut Huskies as an example: They won five Big East tournament games against an expectation of 2.5 (the ninth-most-surprising conference tournament performance of the past 30 years) and then proceeded to rattle off six NCAA tournament wins versus an expectation of 3.6 (the 12th-most-surprising NCAA run in the same span).
For UConn, the conference tournament was a stepping stone to bigger things.
But here’s the catch: Those Huskies were the exception, not the rule. Across the entire population of NCAA tournament-bound teams since 1985, there’s practically no relationship between how much a team outperforms its expectations in the conference tournament and the same metric in the NCAA tournament.2
An alternative way to look at whether conference tournament momentum leads to better NCAA outcomes is to see whether teams whose SRS ratings changed substantially during conference tournaments saw a commensurate change during the NCAA tournament. But again, there is essentially no relationship between a surprising performance in conference tournament play and in the NCAA tourney.
The admitted flaw in both approaches is the same one I ran into when evaluating which college basketball coaches outperform NCAA tournament expectations based on seeding. Such a method ostensibly captures underperformance in the final game of a tournament, but it doesn’t detect the missing future wins expected of a favored team going forward. Moreover, these results shouldn’t be taken to say that conference tournaments have no predictive value. A team’s post-conference-tourney SRS is slightly more correlated with its eventual NCAA tournament wins than its rating before the conference tourney began.
However, this analysis does serve as a warning against putting too much emphasis on the conference tournament relative to a team’s entire body of work, especially when it comes to picking unexpectedly hot conference tournament teams to go further than you’d otherwise predict for teams with their résumés.
Check out FiveThirtyEight’s March Madness predictions.