What do similarity scores say about the top seeds?

When picking a bracket everyone wants to know how good the top seeds are. Which ones should make the Final Four? Go even further? Right now the standard answer seems to be Kentucky. What team though is the one you shouldn’t trust? I used similarity scores to try and determine both the strongest and weakest of the top eight seeds in the tournament this season.

The method was simple. I took the 10 most similar teams to each of the top eight seeds since 2003 and calculated the average number of points their teams would’ve earned in the past if scoring on the 1-2-4-8-16-32 scale. Teams that didn’t even make the tournament in the 10 most similar were counted as scoring 0 points. Also, I took similar teams in 2012 out, but it’s instructive to look at those first. Here are the similar teams in the 2012 bracket:

Kansas’ 2012 similarities:

  • New Mexico (#1 most similar overall)
  • Michigan State (#7)
  • Wichita State (#9)

Michigan State’s 2012 similarities:

  • New Mexico (#2)
  • Georgetown (#6)
  • Kansas (#7)

Missouri’s 2012 similarities:

  • Iona (#2)
  • Creighton (#10)

Ohio State is somewhat similar to Cal (#7) and Kentucky is somewhat similar to Michigan State (#6) as well. You’ll notice that Missouri’s similarities are two lower seeded mid-majors. That follows a trend in general for teams the Tigers resemble in terms of their statistical profile, including Air Force, IUPUI, St. Mary’s and Cornell (the Sweet 16 version). It’s safe to say that for a team like Missouri the match ups are going to be of the utmost importance.

Here now are the results of the study sorted by the highest average points per similar team:

You can click on the chart to get a larger version if you’d like. What you’ll see though is that Kentucky is a great bet. Basically, this team looks like a lot of John Calipari’s most successful  teams. More than half of the Wildcats’ 10 comparisons that have played in the tournament have at least made the Elite 8. None of them were upset in the first round.

I was most surprised to see Syracuse come out so strong in the analysis. While teams with the Orange’s profile were upset a few times, 6/10 teams made at least the Elite 8. That’s a very strong track record. Interestingly, teams with Syracuse’s statistical profile have always received high seeds. None of their comparisons were seeded lower than fourth in the bracket. Top seeded teams similar to Syracuse have a Sweet 16, an Elite 8 and two National Runner-Ups.

The bottom three teams are general considered the weakest of the top eight teams. What though is Kansas doing in that group? The Jayhawks have a score closer to Duke or Michigan State than North Carolina. Well, part of Kansas’ problem is that one of their top results is the 2005 Kansas team that was famously upset by Bucknell. Also, 2004 Stanford, which was a #1 seed, is a top comparison that lost way earlier than expected. The one team that made the Final Four with a profile similar to Kansas? 2011 George Mason.

And finally there’s Missouri. The Tigers score fewer than a point, which is the equivalent of just surviving the round of 64. I don’t think Norfolk State is going to shock the world, but Frank Haith’s team could be extremely vulnerable moving forward. No team with this profile has advanced to past the Sweet 16. That includes disappointing Ohio State teams in 2010 and 2011. If the Tigers go on a run it’ll be slightly unprecedented.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s