The Conference Tournament Effect

If you're a fan, conference tournaments are fun.  They start the wall-to-wall tournament basketball of March with a good mix of mid and low-majors fighting for their conferences lone bid, power conference teams fighting for an at-large bid, and heavyweights duking it out one last time before March Madness officially begins.

But what those conference tournaments mean in the long run, aside from seeding, is up for debate.   Some folks believe in momentum.  To them, winning the conference tournament is a launching pad to great things.  The winning team is "hot" and has proven its mettle in the unique, high-pressure tournament atmosphere of March.  You can't succeed there if you've never been there.

Others will scoff at that assertion.  Sure, it's better to win the conference tournament all else equal, they'll say, but it takes a toll.  Teams play three, and sometimes four or even five, games in as many days against top competition.  College athletes aren't used to playing that type of schedule, especially turning around four or five days later and playing their biggest games of the season.  Even the crucible of March Madness, after all, requires them to win only two games in a weekend to advance onto the next phase of the tournament.

So what's the real deal?  Let's look at the data.  I went back to 1985 and looked at the six power conferences: ACC, Big East, SEC, Big 8/XII, Big Ten, and Pac-10/12.  The Big Ten only just began holding a postseason tournament in the late '90s, and the then-Pac-10 revived its postseason tournament in 2002 after its initial run from 1987-1990.  That's a sample of n=148.

Next I took the expected wins by seeding from this handy article (credit: Cardiac Hill) and compared actual and projected wins.  The conference champions in the sample exceeded expectations by 0.117 wins, i.e. a little better than one win every two years for the six power conference champions, with weak statistical significance (z-test score: 0.20).

There was significant variance between conferences, but the only one where sample data showed winning the conference tournament had a negative effect on NCAA tournament advancement was in the Big 8/Big XII where Kansas significantly under-performed a  #1 or #2 seed several times under Roy Williams, as did Oklahoma on multiple occasions.

Looking at just the past ten years (n=60), teams gained approximately 0.285 wins, but again with weak statistical significance.

Winning the conference tournament might have an effect on how a team plays later in March, but if it does, I'm not going to be the one who finds it.