Mailbag: Roy Williams Has Never Reached The Elite Eight After Losing By More Than 20

We received an interesting email this morning:

I saw your Tweet where no team has lost by 30 then won the NCAA tournament. But it seems like Roy Williams has made some pretty deep runs after bad losses. His last Kansas team made the final after losing to Arizona by 17 at home. Maybe this is the year he wins it after losing by 30!
– Erik

That prompted us to do a bit of quick research. Just how much does a regular season loss by a Roy Williams squad tell us about its postseason prospects?

As the title of this post implies, there does seem to be something of a barrier around the 20-point line. Here is a plot of his team’s worst loss each year, versus the number of NCAA tournament wins he notched. The dark line is the best fit through the points (with an R^2 of only 0.17):

It’s fairly rare for one of his teams to lose by more than 20, but when they do, it’s generally a pretty bad sign for their postseason outlook. The teams suffering such a large beatdown have gone on to win 0, 0, 1, 1, and 2 games in the NCAA tournament.

Judging by the best fit line, this most recent loss — if it were the only thing we knew about the Tar Heels – seems to point towards a mere one win in the NCAA tournament.

However, once the tournament rolls around, we’ll also know UNC’s seed. The five previous teams that lost by more than 20 were seeded only #2, #4, #8, and missed the tournament twice (the two zero-win years). That means they underperformed slightly in the tournament, since we would have expected them to win 5 or 6 games, but they only won 4.

Here’s the full season by season list:

SeasonTeamBiggest LossNCAA SeedNCAA Result
1988-89Kansas28 @ Oklahoma----
1989-90Kansas22 @ Oklahoma2L in 2nd Round
1990-91Kansas17 @ Kentucky3L in Championship
1991-92Kansas7 vs Louisville1L in 2nd Round
1992-93Kansas15 vs Long Beach St2L in Final Four
1993-94Kansas14 vs Temple4L in Sweet Sixteen
1994-95Kansas19 @ Indiana1L in Sweet Sixteen
1995-96Kansas8 vs Temple2L in Elite Eight
1996-97Kansas2 @ Missouri1L in Sweet Sixteen
1997-98Kansas11 @ Hawaii1L in 2nd Round
1998-99Kansas18 vs Kentucky6L in 2nd Round
1999-00Kansas33 @ Oklahoma St8L in 2nd Round
2000-01Kansas31 @ Wake Forest4L in Sweet Sixteen
2001-02Kansas10 @ UCLA1L in Final Four
2002-03Kansas17 vs Arizona2L in Championship
2003-04UNC11 @ Georgia Tech6L in 2nd Round
2004-05UNC13 @ Wake Forest1National Champion
2005-06UNC15 @ USC3L in 2nd Round
2006-07UNC8 vs Gonzaga1L in Elite Eight
2007-08UNC11 vs Duke1L in Final Four
2008-09UNC7 vs Boston College1National Champion
2009-10UNC32 @ Duke----
2010-11UNC20 @ Georgia Tech2L in Elite Eight
2001-12UNC33 @ Florida State??????

Tar Heel fans can’t be happy about the performance by North Carolina this weekend. And according to the historical record, they’ve got good reasons for their pessimism.

----------------------------------------------------------------- SHARE! -----------------------------------------------------------------
    • DABBERWOCKY

      We lost by 33 to FSU, not Va Tech.

    • http://www.teamrankings.com/ David Hess

      Thanks, mental lapse there. Fixed.

    • Dorothy Mantooth

      If the model has an R^2 of 0.17, isn’t more appropriate to conclude that there is no linear relationship between the variables?  In other words, you cannot say with any confidence that the size of the worst loss is correlated with the number of tournament wins.

    • http://www.teamrankings.com/ David Hess

      Dorothy, you’re a saint, but that’s actually a relatively strong correlation given a variable as erratic and non-linear as NCAA tournament wins.

      For comparison, the R^2 of NCAA seed vs NCAA wins for Roy Williams is only 0.20 (throwing out the non-tourney years). The scatter plot looks very similar — the few years with higher seeds are bad, and everything else looks relatively random. 

      Anyway, the point is not that this loss tells us everything we need to know about the rest of UNC’s season. That would be silly. The point is that the mere fact that the team can lose by such a large margin is, in and of itself, a bad sign for their future. I don’t think that’s much of a reach.

      Thanks for the comment, by the way. Always good to have some making an argument on the merits of the logic of the post, rather than just being a blind homer.

    • Sllancaster

      That is Roy.

      How about UNC History???

      The good news — maybe? — for North Carolina: there is a precedent for the program coming back from a blowout loss to win the NCAA title. The 1993 Tar Heels were pasted 88-62 at Wake Forest on Jan. 30 that season (and then lost their next game 81-67 at Duke) before going on to win the national title.

    • Sllancaster

      You can make numbers fit any argument you want to make.

      That is why we play the games!!!!

    • http://www.teamrankings.com/ David Hess

      Totally agree, I just thought this was an interesting little tidbit.

    Like this blog? Subscribe!

    Get new posts delivered automatically to your email address!

    Or Follow Via: RSS Twitter

    About Us

    Categories