Who Knew Best In NCAA Tournament Round 1?
ESPN's crowd showed its wisdom in the men's Round of 64.
Some first-round upsets are nearly impossible to see coming. For instance, if you picked No. 16 seed Fairleigh Dickinson to unseat No. 1 Purdue Friday night, chances are you attended the New Jersey-based school — otherwise, there was little reason to pick a team ranked 312th in Ken Pomeroy’s pre-tournament rankings over the team that ranked 7th.
But others were easier to predict, depending on the source you consulted. And thanks to The New York Times’ handy NCAA tournament comparison tool, we can judge which prognosticator did best in Round 1 of the men’s tourney this year.
To grade the picks, we’ll use what’s called the Brier Score, which is simply the average squared error between a probabilistic forecast and the actual outcome. The lower average error, the closer a picker came to matching the results from the real world. Today’s contestants will be those tracked by the NY Times’ interactive:
[…] the collective picks of a pair of large public bracket contests for the men’s tournament — from ESPN and Yahoo — along with expert forecasts from FiveThirtyEight, Jeff Sagarin, Ken Pomeroy and ESPN.
By this method, ESPN’s Tournament Challenge pickers did the best job of picking the Round of 64. Among the algorithms, the best belonged to Sagarin at USA Today; my usual go-to, Ken Pomeroy, sadly finished last. (Of course, lest you think this is a victory for the wisdom of crowds over the numbers, Yahoo’s bracket contest pickers ranked second-to-last, only narrowly better than KenPom.) And of course, my sentimental favorites of FiveThirtyEight and ESPN were somewhere in the middle.
Does this mean 2023 is another year of the advanced metrics’ favored teams flopping, similar to 2022? It’s too early to say. Despite some stretches of first-round shakiness from analytic darlings like UConn and Tennessee, most of Pomeroy’s top-ranked teams are still alive in the tourney. (If anything, Purdue’s loss could make life easier for Tennessee going forward.) But when it came to identifying those first-round upsets, it was a better year to be an ESPN user picking from the gut than to purely trust the metrics.