The Top NCAA Tournament Teams, According To 5 Stat Geeks

March 15, 2015 - by Tom Federico

Here at TeamRankings, we’ve developed a sophisticated strategy for winning NCAA bracket pools that relies on several different sources of input data to predict every team’s chances. The end result is our official 2015 NCAA bracket picks.

Since we spend a bunch of time on Selection Sunday gathering and reviewing team ratings and projections from different sources, we figured we’d publish some of the comparative highlights.

What Do Various Predictive Systems Think Of This Year’s Tournament Teams?

The chart below represents data from five sites: TeamRankings.com, Ken Pomeroy, ESPN’s BPI, PredictionMachine, and inPredictable.

We’ve plotted each site’s ranking (not rating) of each of the teams below as a gray dot. If multiple sites have a team at the same rank, the dot is darker. (That’s why Kentucky has just one really dark dot; the Wildcats are unsurprisingly ranked #1 by all five systems.)

All teams in the top 20 on any of the sites are included. Reflecting the great deal of consensus across these models, that represents only 23 teams.

Looking at the chart, we can see:

  • The consensus number 2 team is Arizona, which was given a 2 seed by the committee
  • The analytical models are in general agreement that Utah is far more dangerous than their 5 seed indicates
  • UNC is a 4 seed, but ranked ahead of all the 3 seeds
  • Kansas is the lowest ranked of all the 2 seeds
  • Michigan State, Ohio State, Wichita State, and Texas are all 7-11 seeds, yet the models all rank them as top 20 teams
  • There is a considerable difference of opinion across systems regarding teams like #6 SMU, #7 Iowa, and #5 Northern Iowa; SMU, for example, was ranked #19 by Ken Pomeroy and Prediction Machine, but a distant #46 by inPredictable

Here’s the full set of data that the chart was based on. We evaluated all of these data points, and thousands more, as part of our process to give you the best chance to win your NCAA pool.

Predictive Ranks of Top Teams Across Multiple Sources