The Quality Curve collapses—and that could mean more upsets

It’s been 20 days since the last time I checked up on the KenPom Top 20 Quality Curve. For those who haven’t read the previous posts, there’s a connection between the quality of the top teams and tourney unpredictability. The 2011 tourney saw the worst top 20 teams in terms of KenPom efficiency since the data became available in 2004. Conversely, 2007 featured the highest quality top 20. And what happened? The 2011 dance tied a record for upsets (13) while 2007 broke records for yawn-inducing predictability, with just three upsets.

When last we left the Quality Curve, it was more of a cliff then a gentle slope. The first 13 teams were at or above the average quality of their historical counterparts. Then, between Ohio State and Arizona, there was a precipitous drop. The final seven teams in KenPom’s top 20 were historically bad—worse even than the 2011 tourney’s collection of vulnerable squads.

So how has the curve moved in the last 20 days? In a word, it’s collapsed into mediocrity. Check out the slide, animated in groups of four—as if seed by seed:


The top four teams are the same ones we had three weeks ago—Florida, Indiana, Louisville and Gonzaga. All of them own efficiency values better than the average team for their ranking. Theoretically, they would be as prone to an upset as your average top seed. (Remember: just because KenPom rates teams in the first to fourth position doesn’t mean they’ll be one seeds. (Florida? Not likely.)

The next four KenPom teams (Duke, Pittsburgh, Ohio State and Kansas) get progressively worse than the average teams for their position. This is a big change from three weeks ago. For one thing, the teams were Wisconsin, Duke, Syracuse and Pitt. For another, they were right on that average curve. So what does that mean? It means that the presumptive two seeds are weaker than the historical average.

The same dynamic is in play for the ninth through 12th most efficient teams. Twenty days ago, the teams were Michigan, Miami, Kansas and Michigan State. Today, it’s still the Wolverines and Spartans, but Wisconsin and Syracuse have dropped into this grouping. Not only that, but the efficiency of all these teams has sagged. Now, instead of being average “three seeds,” they’re closer to the worst grouping in the last nine years. So—you could be looking at super weak three seeds. Does that mean a 3v14 upset is in the offing? Maybe a 3v6 or 3v11 surprise?

The teams ranked 13th through 20th—from Georgetown to Virginia—have changed names, but just like 20 days go, they remain historically bad for teams in their position. These are the teams that would be expected to wind up as four and five seeds. The 4v13 and 5v12 games are already adventurous enough. Throw in weaker than weak four and five seeds, and you could have round-one mayhem. On the other hand, soft fours and fives could pave the way for top seed chalkiness into the Elite Eight.

I’ll update this Quality Curve after Selection Sunday so you can decide for yourself how unpredictable you think this dance will be.

This entry was posted in Measuring Madness. Bookmark the permalink.

19 Responses to The Quality Curve collapses—and that could mean more upsets

  1. Ryan Tressler says:

    Looks like we might be creeping towards a crazy tournament! It is definitely more fun that way, haha. It will be interesting to see how the lower seeds rate compared to previous years once the tournament and seeds are set.

  2. John says:

    As you said UF wont get a 1 so the 1 seeds will drop on the chart a bit and the 3 seeds (if UF is a 3) will come up. But it does look good for some early upsets.

    • ptiernan says:

      The average curve here just follows the AP Top 20 irrespective of seed. After the brackets are set, I’ll compare the average KenPom value of each seed with their historical counterparts. That should be interesting.

      • John says:

        That would be interesting.

        BTW I gave a quick look to Lunardis 1-4 seeds and how they looked on the road/neutral compared to overall record. Looks like there are a lot of teams that will be in that <.150, <.125, <.100 area. Even with the tournaments going on. Zags best in this category. Indiana is ok but can improve with a good run in the Big 10.

        • John says:

          I bring that up because it may play right into what you are saying here about an upset laden tournament that perhaps helps a couple 1 seeds advance.

  3. ptiernan says:

    I’m looking forward to tracking the curve by the actual seed positions. As you know, the Selection Committee never goes by KenPom. Nor should they, really. Otherwise, Pitt would be a two seed!

    • Ryan Tressler says:

      Agreed on the seeding and on Pitt (they always find a way to disappoint, hopefully being under-seeded will help, haha) . . . I think Butler is about to be set up for a disappointing tournament . . . i get the sense they are going to be quite over-seeded this year as a probable 5-7 seed (when they have played more like a 10-12 seed). It MIGHT be enough to give Brad Stevens’ PASE a little bit of a hit this year with a first round loss (I emphasize “might” because it is Brad Stevens after all and you can never count his teams out in the tournament, haha). I also can’t wait to see where VCU’s seeding settles at, I would put my money on him keeping his PASE lead on Stevens after this tournament (when he can officially take his place at #1 on your list for coaches, haha)

    • Michael Moses says:

      So you have Kenpom data from previous years pre-tournament? It will be interesting to see how the kenpom ratings for each seed this year compare with previous years. Yes this year appears to be top heavy, but there are a fair amount of teams that have great Kenpom ratings but will be underseeded (UF and Pitt).

      • ptiernan says:

        Yes, I’ve got the data for nine years. The key will be where all these underappreciated efficient teams get seeded.

  4. Jeff Waltz says:

    Pete, keep up the great work! I love reading your site this time of year. Hope all is well!

    • ptiernan says:

      Thanks Jeff. Is this Junior? If so, I hear you’re doing good things with the girls program. Keep up the good work.

  5. Ryan Tressler says:

    Pete, loving the info as always and I’ve always appreciated how you take the time to respond to every person possible in the comments, you’re the best . . . any chance we see the old forums make a comeback and also, are you planning on having a chat this year? That was always fun in past years . . . keep up the great work, looking forward to Sunday, haha!

    • ptiernan says:

      The new site doesn’t allow for chat. Sorry–I had to skinny down because I just couldn’t devote as much time to the site. That was the theory anyway. I don’t see why the Comment section that we have now can’t work the same as the forums. You can reply to other people’s comments, right?

      • Ryan Tressler says:

        Yes an very true, haha . . . I guess the Forum feature made a central location for the comments, but it really isn’t that important, it is probably easier this way anyway since all the most recent comments are shown in the right panel anyway . . . I was really more interested in the chat, but that is understandable, haha and its not like there can’t be some form of chat amongst readers in the comments section anyway (only difference being you would have to refresh the page)

        • Ryan Tressler says:

          plus, the entire world is pretty much on twitter (I have one, though I never use it, haha), so any possible chat could easily happen there

          • ptiernan says:

            I have to admit, I still don’t get Twitter, even though I post stuff there. I got a bunch of followers, but I’m not one of those guys who farts out random thoughts while watching games. I just have no interest, either in writing or reading that sort of stuff.

        • ptiernan says:

          Ryan – Maybe what I’ll do is write a few blogs asking provocative questions and invite debate within those.

          • Ryan Tressler says:

            That sounds awesome . . . and yea, I’m with you about Twitter . . . like I said, I never use it, I don’t get the appeal, haha

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>