The magic of margin, revisited

The next two feature articles I’m working on are the all-consuming 2014 Seed Guide and my analysis of Key Performance Indicators. In both of these pieces, the value of scoring margin figures prominently. The fact is, scoring margin is just this side of seeding when it comes to identifying tourney contenders and pretenders.

I’ve shown the chart below in previous years, but it bears updating. With each additional point of scoring margin by which you segment tourney teams, the performance against seed expectations increases. Let me put that more simply: teams that beat opponents by an average of, say, eight point points beat seed projected win totals at a higher rate than those with seven-point margins…and teams with nine-point margins are bigger overachievers than the eight-point club. And so on and so on. Want proof? Here’s the chart:

2014_margin_value

This never ceases to amaze me. Whether you’re looking at favored seeds (1-6) or underdog seeds (7-12), scoring margin is a reliable guide to identifying overachievement. Let’s take the higher seeds first. For these teams, owning a scoring margin higher then eight points a game yields a +.035 PASE. That rate increases to +.108 with more than a 10 point margin, +.176 at more than 12 points and a stellar +.400 at more than 15 points a game. In all, 112 one- to six-seeded teams have come into the dance with more than a 15-point margin. Nearly 80% of them (89) have reached the Elite Eight, 46 have made the Final Four—and 16 have cut down the nets. By seeding these teams should’ve won only 310 games…but they actually notched 355 victories.

The numbers are just as impressive for the Cinderella seeds. Look at how the PASE leaps from the 12-point to 15-point margin groups.  It’s practically linear. The 12-point club owns a +.107 PASE. The 13-point club adds +.158 to their PASE. Then the 14-point club adds another +.113 and the  15-point teams add +.146. Granted, only 12 of these higher seeded squads owned average margins above 15 points. But four of them reached the Elite Eight. They should’ve won fewer than eight games by seed projections, but actually pulled off 14 surprises, for a hefty PASE of +.524. To put this in perspective, only 25 of the remaining 684 teams seeded seven through 12 made the Elite Eight, a 3.6% rate versus 33% for our higher margin teams.

Scoring margin also proves helpful in identifying pretenders. Take top seeds, for example. The average one seed owns a 15.6 average scoring margin coming into the tourney. The 57 top seeds with a worse margin than that are -.336 underachievers with a Final Four advancement rate of 29.8%. The 59 top seeds with at least that margin are +.325 overachievers, and they’ve reached the Final Four at a 50.8% clip. Think about some of the more disappointing top seeds of the last two years: MSU, Syracuse and North Carolina in 2012 and Kansas last year. All of them owned “below-seed-average” margins.

As I unveil the 2014 Seed Guide and the KPI Study in the coming weeks, you’ll see how important margin is in separating overachievers from underachievers at every seed and in most seed match-ups throughout the dance. I just thought it would be a good idea to raise the consciousness of margin in its own right. If you remember nothing about Bracket Science other than that, it will serve you well.

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

10 Responses to The magic of margin, revisited

  1. Gary Diny says:

    Pete,

    As usual great analysis. Was wondering if scoring margin has any correlation (or any type of close correlation) with win percentage. Obviously the more games you win, the higher likelihood of a more positive scoring margin and vice versa. Additionally, how well does scoring margin factor in with later rounds (Elite 8, Final Four). Or does record in close games, record against top 20 become a good predictor. My thinking is that those later round games are more likely to be against much stiffer competition and success in close games and against upper level teams (based on poll rankings).

    Thanks again for the great information!!!

    • ptiernan says:

      All my analysis of win percentage hasn’t shown this sort of correlation. Your question prompted me to examine four conditions:

      1. Top seeds with above average win % and above average margin: +.133 PASE
      2. Top seeds with above average win % but below average margin: -.103 PASE
      3. Top seeds with below average win % but above average margin: +.547 PASE
      4. Top seeds with below average win % and below average margin: -.462 PASE

      It’s interesting that the top seeds with plus W% and margin aren’t the top performers. It’s the top seeds that win at a lesser rate, yet have higher margins. I’d have to do more research on the “close games” hypothesis to give you a better answer. But this is a start.

    • Tom says:

      Not really.

      There’s probably some correlation because if you’re winning a bunch of blowouts, then you probably have a good team. But close games can have a huge impact on winning percentage while having a minimal impact on your average victory margin. For example, let’s say that over a 30-game schedule, you play 6 games that are decided by five points or less. If you go 6-0 in those games versus 0-6, that’s a 20 percent difference in your winning percentage… but at most, you’re talking about a one-point difference in average victory margin.

      • Gary Diny says:

        I get what you are stating, that part is fairly obvious. But does that scoring margin factor in in the last 3 rounds of the torunament (elite 8/final 4/championship game).

        • ptiernan says:

          I’ll have to dig into that a little further. You can look at the Toss-up Picks feature under TIPS+ and review how teams with the higher margin perform in games where the seed difference is three or fewer.

  2. Tom says:

    Real quick, list of teams currently having average scoring margin > 15:
    Louisville (+20.6)
    Iowa (+20.2)
    Arizona (+19.2)
    Oklahoma State (+19.0)
    Louisiana Tech (+17.0)
    Pittsburgh (+16.8)
    Creighton (+16.6)
    Michigan State (+16.0)

    Louisiana Tech?

    • ptiernan says:

      Of those, I like Louisville, Iowa, Arizona, Creighton and MSU. I saw the Hawkeyes against Michigan tonight. Even though they lost, they’re a tough, deep squad that can score. They might surprise come March.

  3. Tommy Gallagher says:

    Does the average margin of victory take into account SOS? For example, KU has the fourth-youngest team in the nation but has also played the second-toughest schedule according to kenpom.com. Is there a sliding grade, an adjusted scoring margin to take that into account? Or is this strictly by the numbers and KU’s margin will skew down compared to, say, Oklahoma State (103rd), Pittsburgh (129th) or Louisville (171st).

    • ptiernan says:

      Tommy – There used to be an adjusted scoring margin stat. I would agree that you’ve got to look at the margins in context. That said, many of the 7-12 seeds with high margins didn’t have great SOS’s, so it might not have that huge an impact. Interestingly, if you’re a member and you read the “Toss-up Picks” feature, you’ll see that, in many instances, having a tougher SOS heading into the dance doesn’t confer an advantage in games where the seed differences between teams is three or less. I can do a quick analysis on margin vs. SOS…gimme a couple hours.

  4. AC says:

    Was wondering if you ever got a chance to that quick analysis on margin vs. SOS. I’m also curious about this.

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>