Quality curve suggests trouble for two seeds in second round, one seeds in Sweet 16

A few weeks back, I took the new KenPom data using his “mismatch-free” formula and compared the average efficiency values of the top 13 seeds for the chalkiest dance of the 64-team era (2007, 4.1% Madometer reading), the craziest dance (2013, 20.8% mad), and the average for the last 11 years.

I promised that I would load in the 2014 Pythag values once they became available. Well, I’ve done that—and the results are pretty illuminating. First the chart, then my comments:

2014_Seed_Curve

It looks like this year’s top seeds are slightly better than average and just a tick below the quality of last year’s one seeds. Remember: only Louisville got to the Final Four last year.

The big surprise comes with two and three seeds. They are both weaker than their historical counterparts—and significantly worse than last year. This could spell more upsets in the second round, particularly for two seeds. They’re much worse than 2013, not to mention the 2003-13 average. Meanwhile, this group of seven seeds is stronger than average, though weaker than last year.

The seed position that really sticks out is the four seed. I thought last year’s group—with Syracuse and Michigan—was imposing. This year’s bunch, led by Louisville and Michigan State, but also including dangerous UCLA and San Diego State squads, is significantly better. Their average Pythag is just .0004 lower than that of the two seed group and solidly ahead of the two seeds.

As for the 5v12 and 6v11 upset match-ups, the numbers indicate that five seeds are more vulnerable than average or last year, but 12 seeds are weaker. Meanwhile six seeds, while slightly better than the 2013 class, are still below average. And 11 seeds are well above average. Is this the year when 11 seeds take over the mantle of Cinderella from 12 seeds?

This entry was posted in Tourney Trends. Bookmark the permalink.

24 Responses to Quality curve suggests trouble for two seeds in second round, one seeds in Sweet 16

  1. Mike L. says:

    In Pete’s article (link: wp.bracketscience.com/?p=882) “So you want to win that billion-dollar bracket…”, he ended it with the line “you should at least make sure that your billion-dollar bracket looks like something the could actually happen.”

    Well, I’ve went through the task of seeing what actual brackets look like, using the last 10 years. Here are the results.

    Elite 8
    2013 : 1, 2 || 9, 2 || 4, 3 || 4, 3 (28)
    2012 : 1, 3 || 4, 7 || 1, 2 || 1, 2 (21)
    2011 : 4, 2 || 5, 3 || 1, 11 || 8, 2 (36)
    2010: 5, 6 || 5, 2 || 1, 2 || 1, 3 (25)
    2009: 1, 2 || 1, 3 || 1, 3 || 1, 2 (14)
    2008: 1, 3 || 1, 10 || 1, 2 || 1, 3 (22)
    2007: 1, 3 || 1, 2 || 1, 2 || 1, 2 (13)
    2006: 4, 2 || 1, 2 || 1, 11 || 1, 3 (25)
    2005: 1, 3 || 4, 7 || 1, 6 || 5, 2 (29)
    2004: 4, 3 || 1, 2 || 1, 7 || 8, 2 (28)

    Sweet 16:
    2013 : 1, 12, 3, 2 || 9, 13, 6, 2 || 1, 4, 3, 15 || 1, 4, 3, 2
    2012 : 1, 4, 3, 10 || 1, 4, 3, 7 || 1, 4, 6, 2 || 1, 13, 11, 2
    2011 : 1, 4, 11, 2 || 1, 5, 3, 2 || 1, 12, 11, 10 || 8, 4, 3, 2
    2010: 9, 5, 6, 2 || 1, 5, 6, 2 || 1, 12, 11, 2 || 1, 4, 3, 7
    2009: 1, 12, 3, 2 || 1, 5, 3, 2 || 1, 4, 3, 2 || 1, 4, 3, 2
    2008: 1, 4, 3, 2 || 1, 12, 3, 10 || 1, 5, 3, 2 || 1, 12, 3, 7
    2007: 1, 5, 3, 7 || 1, 4, 3, 2 || 1, 5, 6, 2 || 1, 5, 3, 2
    2006: 1, 4, 6, 2 || 1, 13, 3, 2 || 1, 5, 11, 7 || 1, 4, 3, 7
    2005: 1, 12, 3, 2 || 1, 4, 6, 7 || 1, 5, 6, 10 || 1, 5, 6, 2
    2004: 9, 4, 3, 10 || 1, 4, 3, 2 || 1, 5, 3, 7 || 8, 5, 6, 2

    Upsets: (TOTAL) : By Round
    2013: (11) 7-3-0-1-0-0
    2012: (9) 7-1-1-0-0-0
    2011: (13) 5-4-2-2-0-0
    2010: (11) 5-4-2-0-0-0
    2009: (5) 5-0-0-0-0-0
    2008: (8) 5-2-1-0-0-0
    2007: (3) 2-1-0-0-0-0
    2006: (12) 6-4-1-1-0-0
    2005: (8) 4-3-1-0-0-0
    2004: (7) 2-4-1-0-0-0

    My Thoughts:
    1. Always seems to be at least 2 two-seeds in the E8. 2005 & 2008 are the exceptions. As Pete pointed out before, two seeds look extremely weak this year. After 5 straight years of 2 two-seeds plus the highly questionable job by the committee, maybe this year bucks that trend.
    2. As Ryan Tressler pointed out in another forum, at least one 12-seed or 13-seed makes the Sweet 16. 2007 had none (but it was a chalk year), yet 2008 followed up with 2 12-seeds in S16 to balance it out. I don’t think this year is a 2007 year.
    3. No upsets in F4 or Championship games. There have been 5 opportunities in the last 10 years. 2 5-seeds, an 8, a 9, and an 11 all with a chance to spring a F4 or Champ game upset, and all 5 cinderellas coming up bare-footed.
    4. An interesting strategy that incorporates these bracket outcomes are the numbers in parenthesis in the E8 section. These numbers in parenthesis are the sums of the E8 seeds for that year. Just by looking at the 2007-2009 years when the tournament was relatively calm, seed totals were 13, 22, and 14. I think everyone holds to consensus that 2014 is not going to be a calm year, so you will probably want to shoot for a total above 22. I did not do the sums for the S16 since I felt that these sums would not be as reliable as the E8 sums. You are welcome to do them.
    5. Feel free to add your own observations below, or do any of the years prior to 2004. It seems as though these last 10 years have given us a lot of information to work with. The stability of 07-09 combined with the wildness of 10-13, and the uniqueness of 04 itself stands on its own. In 2004, 32 games in the first round with only 2 upsets (Predictable as a Kevin Costner movie), yet 16 games in the second round with 4 upsets (Like that scene in the Godfather where all the bosses get taken out at once). Just remember, make your bracket look like an actual bracket, but don’t make it look like my bracket, because I’m not sharing my prize money!!!

    • Ryan Tressler says:

      what I have found on average is this:

      Round 1:
      you can expect around 1.7 of teams seeded 13-15 to advance (aka, 1 or 2 out of the 12)
      you can also expect 4.3 teams seeded 10-12 to advance past round 1

      Round 2:
      typically 2 teams seeded 6 or 7 advance to the sweet 16
      every other year, an 8 or 9 seed will also advance to the sweet 16
      you can expect 1.2 of the 10 or 11 seeds to advance to the sweet 16 each year
      also, on average, at least one 12-15 seed advances past round 2

      SWEET 16:
      you can typically expect one 3 seed to advance to the elite 8
      1.35 of the seeds 4 through 6 advance to the elite 8 as well
      finally, you can expect one team seeded 8 through 11 to advance to the elite 8 (again, on average)

      ELITE 8:
      on average around 1 of the 2 seeds make the final four (actual number .86, so just below 1)
      1 of teams seeded 3 or 4 will advance (actual number .93)
      also, every other year, a team seeded 5 through 11 will make the trip to the final four (.55 per year)

      Again, to reiterate, these are all ON AVERAGE, haha

      • Gary Diny says:

        Thanks for the leg work Mike L!!!!
        I did the sweet 16 seed totals:
        2013=81
        2012=72
        2011=80
        2010=77
        2009=49
        2008=70
        2007=51
        2006=71
        2005=72
        2004=73
        10 year average 69.6
        Omitting 2007-2009, other 7 year average is 75.1

        Similar to Ryan posting above, in the early years of the 64 team tourney, the FF was a 1, 1, 2 and some lower seed. That has not held up in the last 10+ years, but those general trends are good to look at when searching for the “perfect bracket”

        Gary

        • Gary Diny says:

          Took the time to review all the 29 prior tourney brackets and tabulate the seed totals for the sweet 16, elite 8 and final four. I am not going to post it all here as it is saved in excel format, and I am not interested in re-typing the info. If Pete wants it, I certainly can send it to him (or others if desired).
          Sweet 16 seed total averages: Range (49 in 2009 to 88 in 1999)
          All 29 years = 70.79
          Removing top 3 and bottom 3 (outliers) = 71.22
          Average # of double digit seeds in sweet 16 = 2.207

          Elite 8 seed total averages: Range (13 in 2007 to 40 in 1990 and 2000)
          All 29 years = 24.9
          Removing top 3 and bottom 3 (outliers) = 24.35
          Average # of double digit seeds in Elite 8 =0.414

          Final four seed total averages: Range (4 in 2008 to 26 in 2011)
          All 29 years = 10.76
          Removing top 3 and bottom 3 (outliers) = 9.913

          If you are trying to build that perfect bracket, might want to play the percentages of how the seeds seem to shake out and be somewhat close to that historical. I took out the outliers to see if it made much of a difference, it did not. I am sure there is probably a better statistically way to do it, but that is not my specialty.

          I think Pete has the average numbers of how many upsets occur total per tourney and might have them per round as well.

          Gary

          • Gary Diny says:

            Other interesting info to ponder;

            Only 4 of 29 tourneys have all 4 #2 seeds making sweet 16, and only 1 since 1997 (that was 2009).
            Only 2 of 29 years have all 4 #3 seeds making the sweet 16 (2008 and 2009)

            Gary

    • John says:

      What is the Final Four breakdown for these years? (seeds)

      OR

      What site provides this information?

      • Ryan Tressler says:

        check out the actual F4/Champ post under the tips link (not the bracket in the models link) . . . it has a list of all the final four teams in that post

      • Ryan Tressler says:

        ignore that last post, it has who the model picked, not the actual final four . . . you can find the final four listed all over the place though, haha . . . check out college basketball reference, or even wikipedia, haha

  2. ptiernan says:

    You guys are absolutely blowing me away. Thanks so much. This is a great thread. So…is anyone going to “outcome match” their own model that incorporates this? Can I issue a challenge? If you guys come up with an “Ultimate Curve Fit” model, I’ll post it and give y’all credit. Any takers?

  3. ptiernan says:

    Gary, Ryan, Mike – Any desire to do a three-headed beast of a curve fitting bracket?

    • Ryan Tressler says:

      might be able to work something out here, haha

    • Mike L. says:

      I’m always up for a challenge, but I’ll probably be the weakest head of the 3. I have done a group pick in the past for fun, it turned into a chalk bracket because 3 or 4 would overrule the other 2 or 1 on upsets.

    • Ryan Tressler says:

      Pete, sent you two models to your email, one using pythag, and one using the seed by seed guidelines, let me know what you think

  4. Gary Diny says:

    Got some questions on a pool I am in…seeking thoughts. Basics of the pool is like buying stocks at each seed level 1-16 (treat the play-in games as 1 team for that seed level…Iowa vs Tenn is one team)
    You pick 1 team from each seed level you think will go far… aka long run in the tourney.
    You pick 1 team from each seed level you think will not go far…aka a short run in the tourney.
    You accumulate points for you long team wins, you deduct points for your short team wins. Yes you could end with (-) points at the end of the tourney. Points are 1, 2, 3, 4, 5, 6 for the successive rounds of the tourney.

    Interested in others thoughts on some options for each long and short at the seed levels.

    Thanks!!

    Gary

    • Ryan Tressler says:

      I like that format . . . where to begin, haha

      • Gary Diny says:

        yeah its a good pool. makes you think of both winners and “losers”. I have been just outside of the $$ (300-500) for top places and growing) the last 2 years. Last year Michigan was my 4 short, kept deducting points. IF they had lost to Kansas, I would have had 1st locked up after the elite 8 games. 2 years ago, I needed either Baylor over Kentucky, UNC over Kansas, Ohio st over Kansas or Louisville over Kentucky to finish 1st. None of those happened and I finish 4th.

        nothing sucks more than being close and not closing the deal. The short teams have doomed me Michigan in 2013 and Kansas in 2012.

        Gary

  5. Gary Diny says:

    http://www.allbrackets.com/

    Above link is where you can find that info. you need to tally the numbers by hand. Time consuming, but something to get some historical perspective.

    Gary

  6. Paul says:

    Some thoughts here… you’re not going to win a billion bucks but you will get to be inundated with Quicken Loan ads. That’s why Warren has the billion and we don’t. He’s somewhat bright.

    As for the hoops I was floored by the 1 seeds margin of victory. The lowest I can ever remember seeing. Virginia is sitting on a 10 almost single digits. Wichita for being in a conference that bad only comes up with a 15. The other 2 teams are not much better. The super 1 seeds of the past always put up margins around 17 and 18 and even better.

  7. Andy says:

    When was the last time no 2 seed made the Elite Eight? I ask because, call me crazy and tell me what you will, but I actually think it may (and perhaps likely will) happen this year.’

    Kansas: More than one freshman starter and no Embiid in the lineup right now. On top of it, although they have won three of their last five, they needed overtime to get their last win which was bracketed by two losses which were never really close. Plus, as I have mentioned time and again, Bill Self’s tournament performance is a roller coaster every couple years, and he just had a high two years ago. Stanford is also the toughest 10 seed by the efficiency ratings, and a good pick to take over the Jayhawks and untrustworthy New Mexico.

    Wisconsin: I really do not trust Bo Ryan, plus this team doesn’t rebound very well. Although they have a fair amount of momentum on their side, the bad news is….so does Oregon. Although the Ducks don’t meet the qualifications for a Cinderella, they also didn’t last year against St. Louis, while the Billikens met the qualifications to be a victim. Apart from Bo Ryan’s fluky Elite Eight run in 2005, they meet all the criteria to lose a game like this. (I’m ignoring BYU because they never win when they are the lower seed.)

    Michigan: Unlike the other three, this weak 2 seed got a very good draw to the Sweet 16, with a very weak 7 and a hobbling 10. But standing in the way is Duke, who beat them when McGary was playing (albeit in Durham, but still). This team struggles on defense and rebounding, and yes Duke does a little, but they also score like crazy and have a better turnover margin.

    Villanova: I really want to like this team. They meet the attributes of a prototypical overachieving 2 seed that could easily make the Final Four. But there are a couple things working against them. First is their less-than-stellar perimeter defense. If they get a team that shoots and makes a lot of threes, they will go down quickly, and there is such a team who they could easily get in the Sweet 16 in Iowa State. Second is what I have said before, that Jay Wright’s 30% SOAR does not excite me in terms of making a deep run. I tend to think he does better when he is a strong underdog than when he is a favorite.

    So indeed, riddle me this. When was the last time no 2 seeds make the Elite 8? I am really curious about this.

    • Gary Diny says:

      No #2 seeds in the elite 8 have happened 2x, 1999 and 1990. Each of those years only 1 made it to the sweet 16.

      Gary

      • Andy says:

        Thanks Gary, just looked and indeed you’re right! As an aside, I remember one of the teams upsetting a 2, as my sister was a student at Miami (OH) when they went to the Sweet 16 as a 10 seed. Still, I knew this had to be rare, akin to all the 1 seeds failing to reach the Final Four only twice in the modern era (2006 and 2011). If Michigan had Villanova’s draw, it would be tempting to take them out early, too! But not Texas or Arizona State, and I’m not convinced UConn is strong enough to beat Nova.

    • Paul says:

      I think your 2 seed to make it to E8 is Wisconsin. You are right about them to an extent. I don’t think its fair to hold Bo Ryan to these misses. He wasn’t the one to botch the play against Syracuse a couple years ago. Don’t forget these guys were unbeaten out of conference against some pretty good teams. You are right their weakness is their inside play. But Oregon and Creighton play right into Wisconsin’s hands. If you don’t play solid defense Wisconsin will beat you. This is the first time Ryan has had this many good players. They have a Freshman coming off the bench you will want to watch.

  8. ptiernan says:

    Yo Gary…if you can drop these tidbits in single Word file, I’ll post.

Leave a Reply

Your email address will not be published. Required fields are marked *

* Copy This Password *

* Type Or Paste Password Here *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>