N.C.A.A. Bracket Picks: Where Fans and Experts Diverge

Picking a perfect N.C.A.A. tournament bracket is all but impossible.

Whoever wins a bracket contest — whether a small office pool or a big national one — is going to need at least a little luck. But there are ways to improve your chances, especially in a large pool. One avenue is examining the differences between the forecasts of mathematical models and the collective choices of fans.

The table above shows 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. Experts use models to assign probabilities of advancement for every team in each round.

The outlooks from the models and the fans are broadly similar. Both groups have No. 1 seeds like Alabama and Houston as best bets to win the tournament, and both groups don’t expect lower-seeded teams to make much of a deep run.

But the public brackets, representing the wisdom of the crowd, and models do differ widely in some ways. Fans tend to rate No. 1 seeds or well-known teams as more likely to win the tournament than experts do. For example, fans are much more likely to have Kansas and Alabama in the Final Four than the experts.

The table below shows an aggregate view of the difference between the public’s picks and those of forecasting models. Teal-colored cells represent picks that forecasters thought were much more likely to pay off. More orange cells represent picks that the public preferred more:

Higher-seeded teams tend to perform much better than lower-seeded ones, and there’s no reason to be a contrarian on that basic fact. But if you pick the same teams as everyone else, your bracket will look the same, too, something you should consider. A “contrarian” approach described in a 2005 paper said a bracket strategy should lean on favorites that “are likely to be underbet by our opponents relative to other teams in the pool.”

If you think the forecasting models know something the public does not, teal-colored cells represent potentially less risky ways to separate your bracket from the pack. Viewed this way, teams like Tennessee, Creighton and Connecticut may be a good opportunity.

There is no guarantee that this approach will work, of course, and part of the draw of the N.C.A.A. tournament itself is its unpredictability. But, no matter what, you’re going to need some luck to win your bracket. Our hope is that these tables mean you might need a little less luck than your competition.