The March Madness magic of margin

If I could know only one thing about a tourney team when I filled out my bracket besides its seed, it would be a tough call between Ken Pomeroy’s Pythag calculation and the straightforward scoring margin average.

Scoring margin is one of the few metrics I track that has a reliable correlation to tourney overachievement. Simply put, the higher a team’s scoring margin heading into the dance, the better its PASE performance through the tourney. This holds true for both the favored seeds (one through six) and the Cinderella seeds (seven through 12). Take a look at the chart below. With each incremental point in scoring margin, there’s a corresponding rise in performance against seed expectations:

2013_margin_increment

Look at the blue bar first. This represents the PASE values of one through six seeds as you add a point to their scoring margin. Almost every one of these teams has a margin better than five points per game, so it’s no surprise that the PASE is nearly right at expectations (+.003). But as we narrow down on high-seeded teams that beat opponents in the regular season by successively better margins, the tourney performance increases. At more than 10 points a game, the PASE is +.092. At more than 15 points, we’re talking a +.400 PASE—nearly half a game overachievement.

The PASE for Cinderella teams seeded seven through 12 climbs even more exponentially, doubling every couple points to a margin better than 11, then taking off after that. When you look back at some of the more surprising runs of the last few years, they were made by these middling seeds with higher than usual scoring margins. Eleven seed George Mason in 2006 owned a 10.6ppg margin. Stephen Curry’s Davidson Wildcats had a huge 15.8-point edge over opponents. Remember 12 seed Cornell in 2010? They beat teams by 12.1 points a game.

In fact, if you take the average scoring margin at each seed position, then evaluate the tourney performance of teams above and below that margin, you’ll find that at all but three of the 16 seed positions, teams with a higher than average scoring margin are overachievers. This list of the average scoring margins at each seed position should come in handy when you’re pondering your brackets:

  • 1 seed – 15.6 ppg
  • 2 seed – 12.4 ppg
  • 3 seed – 11.1 ppg
  • 4 seed – 10.1 ppg
  • 5 seed – 8.6 ppg
  • 6 seed – 8.0 ppg
  • 7 seed – 7.7 ppg
  • 8 seed – 7.5 ppg
  • 9 seed – 7.3 ppg
  • 10 seed – 7.2 ppg
  • 11 seed – 7.6 ppg
  • 12 seed – 8.1 ppg
  • 13 seed – 8.3 ppg
  • 14 seed – 7.7 ppg
  • 15 seed – 6.1 ppg
  • 16 seed – 2.5 ppg

Now, take a look at how teams above and below these margins fare in the dance at each seed position:

2013_margin_seedavg

The difference in PASE performance between teams above and below their seed’s average margin is most extreme for the higher seeds. Top-seeded teams with margins above 15.6 points a game are +.339 PASE overachievers, while those below average underachieve at the same rate. The gulf for two seeds above and below the 12.4ppg margin is the second most sizable gap of all seeds. And the gulf for three seeds is the third biggest.  Four and five seeds also show a significant performance gap, but then the results are mixed for six through eight seeds. Nine through 12 seeds reassert the dominance of above-average margin teams before the results converge to expectations for the lowest four seeds.

My takeaway from this is to pay more attention to the scoring margins of the higher seeds. I’d be very skeptical of advancing a top seed with a lower than average margin too far in my bracket. Look at the last three dances: six one seeds had margins below 15.6 per game and not a single one beat seed-expected win totals. That includes MSU, Syracuse and North Carolina last year.

Need more evidence of the enduring magic of margin? I wanted to see if the supremacy of high-margin teams was a recent trend, or held true throughout the modern tourney era. So I divided the 28 years of 64-team brackets into four seven-year periods and evaluated the PASE performance of teams with margins below eight points, between eight and 12, and more than 12 points a game. Here’s what I found:

2013_margin_foureras

Check out the orange bars. These represent the teams with the highest margins. In all four seven-year periods, they’re consistent overachievers, logging PASE values between +.100 and +.196. For the entire span of the modern era, these squads are +.162 PASE overperformers.

The light blue bars correspond to the PASE numbers for teams with middling margins. They play the closest to seed expectations, averaging -.089 PASE underachievement overall. Their performance for each of the seven-year periods doesn’t deviate quite as much from seed expectations as that of the teams with the lowest margins. While these teams actually overperformed between 1992 and 1998, they’re -.103 PASE performers overall, with their worst numbers coming most recently.

After Selection Sunday, you’ll have dozens of numbers to pour over as you fill out your bracket. Make sure you give margin a little more weight than others.

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

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>