What seed does Oklahoma State deserve?

OKC Dave —  February 26, 2013 — 23 Comments

Web

This is the first offering in a three-part series that will examine how OSU’s current team compares to NCAA tournament teams from 2009 to 2012, a universe of 266 teams. I will attempt to answer three main questions (with a few detours along the way):

Part 1: What NCAA seed does OSU deserve?

Part 2: What kind of teams advance in the tournament?

Part 3:  How far will OSU advance in the tournament?

In all three parts, teams will be evaluated using Ken Pomeroy’s system (yes, I refer to him all the time — because he’s the best). I will evaluate these 266 teams based on three Pomeroy factors:

  • Adjusted Offensive Efficiency (points scored per 100 possessions, adjusted for SOS)
  • Adjusted Defensive Efficiency (points allowed per 100 possessions, adjusted for SOS)
  • Pomeroy Rating (Pomeroy’s rating, which is based on a combination of the two factors listed above)

You want a high offensive number and a low defensive number.

Let’s get started with Part 1: What NCAA seed does OSU deserve?

Each of the grey dots represents an team that made the NCAA tournament from 2009 to 2012 (these are the 266 teams I mentioned). I’ll wait here while you count them.

I situated the axes so that you want to be at the top right of the chart – high offensive number (“north” on the chart), low defensive number (“east”). Let’s look at a couple examples:

See the dot closest to the top right corner? That belongs to the 1-seed 2010 Duke team that beat Butler for the title.

Okay, how about the dot closest to the lower left corner (the bad corner)? That belongs to Alabama State’s 2011 team that lost in a play-in game.

I’ll get to that orange dot in a moment. First, let’s add something else to the chart:

Those numbers (1 through 16) are the average chart position for each of the 16 NCAA seeds over the last four years. For example, the average of the sixteen 1-seeds over this four-year period is 119.5 on offense and 88.1 on defense.

I found this pretty fascinating. The seeds are lined up pretty much the way they should be, progressing from the lower left up to the upper right. What is interesting to me is how some of them are bunched together in packs. 1 seeds are clearly better than the rest. 2 through 4 are kind of bunched together, and 5s are kind of on their own. Then 6 through 12 are a train wreck — no wonder there are so many “upsets” in this range — all of the teams are basically the same. 13 through 16 show you to the exit in an orderly fashion.

Back to that orange dot — yes, that is OSU’s current season. What does it mean? Well, notice how we are “east” of the 1 seed number? We are playing elite, 1-seed level defense — in fact, only 14 NCAA tournament teams out of 266 have a better defensive figure than OSU’s current 86.7.

That’s the good news. Here’s the bad: draw a straight line over from the orange dot to the pack of numbers in the middle. OSU is playing offense on a 10-seed level. 63% of the 266 tournament teams over the last four years have a better offense than what OSU will be bringing to the dance.

So, to answer the question I brought up at the beginning of this: what seed does OSU deserve? I did a little regression analysis using Pomeroy’s overall rating versus the NCAA seeds assigned to each team over the last four years and came up with a formula that is fairly accurate at assigning ratings (it was correct within two spots 69% of the time over the last four years).

Once again, OSU is the orange dot below. The line running through the dots is the closest fit to each dot, and it provides the formula for determining the predicted seed. We have a Pomeroy rating of 0.9124. When you plug that figure into the formula, it says OSU will be assigned an NCAA tournament seed of 5.1.

Part II | Part III

  • KS1010

    This is some awesome date crunching work. Keep up the work Dave.

    • KS1010

      Data*

    • Bro, I’m a Bro, bro. Slam some brews.

      Best Freudian slip of all time, bro! CRUNCH THOSE DATES, BROS!!!

  • Bro, I’m a Bro, bro. Slam some brews.

    I like the top graph the most, bro. It would be cool to see the names of the dots (!) that are right around where OSU is listed here (in the first graph) and see how they did in their respective years, bro.

    • http://gravatar.com/okcdave okcdave

      Stay tuned for part 3 bro.

  • Jess

    Cannot get enough of the stats. Great job, Dave.

  • ScKing

    I LOVED your post Dave! Great work! I’m taking a graduate stat class this semester and this totally puts my studies into perspective. Thanks for the great example of applying statistics!

  • Jon

    When the seeds are announced for the tournament this year, there needs to be a post with all of this information and graphs for this years teams. That way all of us loyal readers can hustle our office tournament pools…..

    • madsenpoke

      +1

  • Chris

    Great work! Pretty much all I understood was 5.1 seed though. You really need to switch from OKC_Dave to OSU_Rainman.

  • Doug

    Great stuff, Dave.

    I have a general question about Pomeroy. Maybe you know the answer.

    All season, kenpom has pegged OSU as one of the unluckiest teams in the nation. Right now, we’re ranked #254 (of 347) in terms of luck. But most of the articles we’ve seen recently have rightly pointed out that OSU is a few breaks away from being 7-7 in conference instead of 10-4.

    Akron, Baylor, Iowa State, OUII, Gonazaga, and KUII were essentially ties. We’re 4-2 in those games. You could throw KSU in there, but then you’d also probably have to add KUI. Any way you slice it, we have not been losing more than our share of close games, which is usually how these kinds of systems tease out the luck.

    So how is it that Pomeroy’s system pegs OSU as unlucky? My theory: we have huge margins of victory against the bad teams (TCU, Tech, etc) and have some decisive victories over a few good teams too (NCSU, Tennessee? Texas?) Contrary to conventional wisdom, the ability to consistently roll bad teams *is* a reliable marker of strength. So, to summarize what Pomeroy’s computer might be saying: OSU has been a little lucky to have won some close games, but their overall body of work indicates that some of those games shouldn’t have even been close in the first place.

    This is a little off-topic, because I know Pomeroy’s system isn’t this smart, but I felt like OSU outplayed OU soundly at Gallagher-Iba. It seemed like half of OU’s shots rolled around the rim four times before falling, and Grooms hit a dozen shots that would have required 100 takes if he was making a trick shot video.

    Anyway, I’m curious if you have any thoughts on OSU’s luck ranking.

    • http://gravatar.com/okcdave @okc_dave

      Great question. I think the main thing Pomeroy’s “luck” rating is trying to answer is this: is this team’s record misleading? In our case, despite the fact that we’ve won a few close games, the answer is no. We “deserve” our record based on how we have played throughout the season…in other words, we have played consistently well enough to justify our record.

      Very interesting thoughts, though.

  • OSUMJ

    This is very good data for people that like to see numbers. How many 5 seeds have made it to the final 4.

  • http://sadastronaut.wordpress.com Chance

    Ah man, that means we have to play a 12 seed. Great work.

  • http://gravatar.com/carsonc5 carsonc5

    Fascinating. That really backs up a lot of what we see on the floor. Good stuff.

  • http://www.facebook.com/freddyrains2003 Fred Rains

    Just awesome.

  • Dylan

    I am interested to know who that ELITE defense on the FAR right was/is.

    • http://gravatar.com/okcdave @okc_dave

      That is Memphis in 2009 (the Tyreke Evans year). They were 31-3 entering the tournament as a 2 seed. They advanced to the sweet 16 and ran into Mizzou, who lit them up for 102 points — by far the most they gave up all year.

      • Scott

        yeah but they play in Conf USA. Not exactly the ACC is it?

  • Steven Myers

    Great job Dave! can’t wait for parts 2 & 3.
    Dave (and Kyle) you HAVE to read ESPN The Magazine. They LOVE crunching numbers like this. However, this months issue is just BUSTING at the seams with number crunching. It’s the March 4th issue titled “The Analytics Issue”. My favorite ‘Tale of the stat’ in this issue is how Andrew Luck is proved to be a more valuable runner than Bob Griffin. Good stuff.

  • Dave Latimer

    Nice Work Dave…. I find it interesting from your scatter plot that on Average a #10 team is better defensively than a #7 team according to the KenPom ranking criteria. Likewise a 9 is better defensively than an 8, on average. The difference is the 7 and 8 teams are better offensively. What would be interesting is to look at those first round matchup’s individually using KenPom’s criteria and compare with who won. May lend some proof or disproof to the old adage that “defense wins championships” ( or tourney games in this case.). Very Interesting work….

    • FM_MarcusFan

      Wherever OSU ends up being seeded I think they are a tough out. Their six losses fall into 3 categories. 1) 2 gut wrenching single point losses to the #1 and #6 ranked teams in the country; 2) early road losses where a hot player or 3 pt shooting did them in (VT – Eric Green 28 pts 4-7 from three, KState – McGruder 28 pts 5-9 from three, Oklahoma 8-13 team 3pt shooting) and 3) Baylor loss where the team just didnt shoot well. It seems like they have fixed some of the defensive issues that led to the hot shooting after the first Bedlam game. All in all looking for a good tourney run.

  • Josh

    Awesome! So where is Oregon?