NCAA Tournament Probabilities: How Big Is UConn’s Home Crowd Advantage?

The Sweet Sixteen begins tonight, and the NIT Final Four is set. Our NCAA Tournament and NIT probabilities are updated accordingly. As a refresher on how we calculate those: We run 10,000 Monte Carlo simulations, starting with current kenpom ratings and letting them adjust somewhat to results within each simulation. For example: In simulations in which San Diego State upsets UConn, San Diego State will carry a slightly higher rating into its Elite Eight matchup. We’ve calibrated these in-simulation adjustments not to how much kenpom generally adjusts to specific results in the real world, but instead such that they align with long-term average changes in kenpom rating. We’re not running the simulations fully “hot,” then, but instead running them “lukewarm.”

One important thing to note with these:

We don’t account for crowd advantage.

Our simulations account for home court advantage, but in a binary way and a simplistic way, one in which only true home courts count and every home court is equally advantageous. They don’t account for situations like those in which UConn is playing in Boston or Indiana State is playing in Indianapolis. This was an intentional choice—home court advantage is more than just crowd, and we wanted to keep things simple—but it’s definitely a vulnerability within our model which we’d like to investigate further in future years.

To that end…

How UConn’s Final Four probability and those of the other teams in the East Region would change if we made TD Garden a full home court for the Huskies, and how it would change if we made it half a home court:

These are big changes. Here they are as true betting odds (those that would exist if there was no vig from the sportsbook—i.e., fair odds if betting between friends):

The situation isn’t limited to UConn. As we mentioned above, Indiana State’s set to be in a similar boat with the NIT Final Four. Arizona likely has some advantage in the West Region. Purdue likely has some in the Midwest. Houston might have some in the South. The degree to which these advantages exist is mysterious—it’s hard to find enough data on teams playing close-ish to home, and it’s harder still to account for differences in the size and dedication of each fanbase. But while our model has UConn as only 53.1% likely to make the Final Four, the number’s probably closer to 3-in-5.

For the Archives