I’ll trade you my 8-seed for your 12!

Now that the brackets are out, rabid sports fans and the clueless alike turn their attention to the 2013 NCAA Tournament. Before I get into a few bracket strategies tomorrow, I wanted to talk about a strange phenomenon that I think most people don’t realize: the tournament format is unfair. I’m not referring to the selection committee botching the seeding—obviously it does—but there is actually a fundamental flaw in the design of the NCAA Tournament.

I think it’s fair to assume two things.  One, that the tournament’s goal is to reward better teams with higher seeds because, in theory, higher seeding gives those teams a better chance to make a deep run. Two, that while every team wants to win the tournament, there are still marginal benefits to making it even one round further. For instance, making the Sweet Sixteen or Elite Eight can give a team like the 1999 Gonzaga Bulldogs the recruiting power it needs to con players into consciously choosing to spend five years in Spokane.

The current bracket system definitely gives higher seeds the advantage in the first round, where the best seed gets to play the worst, the second-best plays the second-worst, and so on. But the second round is where things get screwed up in the current format. Let’s use Oregon’s draw as an example.

Huskies fans may have been elated to find out that the Ducks only got a 12-seed in this year’s Madness. However, those same fans may not be happy to hear that it was actually to Oregon’s advantage. I did some napkin work and calculated Oregon’s chances of making the Sweet Sixteen at something between 14 and 20 percent. But had they been an 8 or 9-seed, those chances would have dropped to something below 7 percent. Then Nate Silver came along—maybe you’ve heard of him—and he pinned Oregon’s chances of reaching Sweetness this year right inside my range at 17.5 percent. Silver gave the 8 and 9-seeds in that draw, Colorado State and Missouri, a combined 14.5 percent.

It’s actually not hard to figure why this phenomenon exists in the Tournament format. The 1-seed is almost always head-and-shoulders better than the rest of the region, and the 8/9-seed has to play them in the second round for a berth into the Sweet Sixteen. Every advantage an 8-seed gains from a better matchup in the first round, they lose it and more in the second-round matchup with the 1-seed. Below I have summarized Silver’s theoretical probability of each of the teams seeded 8-through-12 making the Sweet Sixteen. The “Seed%” at the end is the probability that at least one makes it through from each seed.

Team

  Seed

  Sweet 16%

  Seed%

NC State

8

13.8%

42.08%

Colorado State

8

3.8%

North Carolina

8

14.2%

Pittsburgh

8

18.6%

Temple

9

2.4%

23.11%

Missouri

9

10.7%

Villanova

9

4.0%

Wichita State

9

8.1%

Colorado

10

17.4%

40.54%

Cincinnati

10

10.7%

Oklahoma

10

11.9%

Iowa State

10

8.5%

Bucknell

11

14.0%

42.03%

SMU/MTSU

11

13.3%

Minnesota

11

13.9%

Belmont

11

9.7%

California

12

12.8%

38.74%

Oregon

12

17.5%

Akron

12

5.9%

Mississippi

12

9.5%

8-seeds, despite generally being superior teams to seeds below, gain no advantage beyond the first round this year, according to Silver. And this is a strong year for 8-seeds, too! Pitt is considered a top-10 team nationally by both Pomeroy and Sagarin.

But if you don’t like the theoretical world of prediction, here are some fun historical data. Since 1990*, only twelve 8/9-seeds have made the Sweet Sixteen, but sixteen 12-seeds have made it that far over the same period, more than the 8’s and 9’s combined! Here’s a chart depicting how often the given seeds have reached the Sweet Sixteen and the Elite Eight during the past 23 Tournaments. You’ll notice that 10, 11 and 12-seeds have dominated the 8 and 9-seeds.

Seed

  Sweet 16

%

  Elite 8

%

1

80

87.0%

64

69.6%

2

57

62.0%

42

45.7%

3

51

55.4%

25

27.2%

4

43

46.7%

14

15.2%

5

30

32.6%

7

7.6%

6

28

30.4%

9

9.8%

7

16

17.4%

6

6.5%

8

8

8.7%

5

5.4%

9

4

4.3%

1

1.1%

10

20

21.7%

6

6.5%

11

10

10.9%

4

4.3%

12

16

17.4%

1

1.1%

13

4

4.3%

0

0.0%

14

1

1.1%

0

0.0%

15

0

0.0%

0

0.0%

16

0

0.0%

0

0.0%

Whether you look at it theoretically or historically, the 8 and 9-seeds really get the shaft in the current format. A solution would probably involve an asymmetrical format that includes byes for higher seeds, but perhaps that conversation is for another day.

*Leaving out 1985 – 1989 was not by design. I just happened to have a data set going back to 1990. Adding in Villanova’s run as an 8-seed in 1985 does not affect the data above to any noticeable degree.