(goduke.com)

On Selection Sunday morning, I wondered aloud why people still get so wrapped up in the question of who gets a No. 1 seed when it’s been six years now since Ken Pomeroy showed that it really doesn’t matter in basketball terms.

The first batch of answers to my query muddled the distinction between cause and effect. Yes, No. 1 seeds have a great track record of getting to the Sweet 16 and the Final Four and winning national titles. Top seeds also tend to be the best teams.

A far better response I received was that arguing about who should get a No. 1 seed is fun and, besides, receiving a top seed is a really cool honor. No disagreements there.

Perhaps we could talk about No. 1 seeds in that vein henceforth, more like an MVP award than as something dispositive to title hopes going forward. It’s a venerable honorific with some nice history behind it, and it provides its own ready-made zero-sum boxing ring for debate. That’s fine.

We can do all of the above while also understanding that getting a No. 1 seed makes, in effect, no discernible difference in tournament performance. For example, it is indeed a potentially weighty basketball matter that Duke and Michigan State are in the same region. It makes exactly zero difference, however, what number happens to be printed beside which team.

More specifically, getting a top seed has made no significant difference in a team’s chances to reach the Sweet 16 in the KenPom era versus what that same team would have experienced as a No. 2 seed. The data is persuasive, but, before we go there, do we even need the data? Spoiler alert, what about the No. 10 seeds that teams on the No. 2 line get to play on occasion?

If we truly think opponent seed is so crucial and that it really does benefit a team on the top line, then we should make a No. 1-seed-level big deal out of a No. 10 seed winning in the round of 64. It should be a big story because that event at one stroke nullifies the whole weak-opponent advantage that top seeds are supposed to receive.

The fact that we do no such thing, however, suggests that once games are actually happening we’ve jettisoned our attentive concern for the question of which teams get No. 1 seeds. That is likely a good thing.

Indeed, instead of discussing the No. 1 vs. 2 seed question in terms of tendencies and averages, a better way to think of it might be as more like a game show. For instance, a round of 32 opponent will tend to be, on average, very slightly weaker for a top seed than for a No. 2 seed.

Round of 32 opponent strength, 2002-19

                         KenPom AdjEM
No. 1 seed opponent         +16.38      
No. 2 seed opponent         +17.76                               

The key term there, however, is “on average.” It turns out the variance in round of 32 opponent strength for No. 2 seeds is six times as large as what No. 1 seeds see. For the most part, this variance represents the difference between Nos. 7 and 10 seeds.

In short, top seeds aren’t distinguished by the fact that they get weak round of 32 opponents as much as they are by the fact that their opponents in the second round are comparatively easy to forecast in terms of their level of strength. That is far less true with any other seed line, including the No. 2 seeds.

So, if you’re the contestant on the bracket game show, the question is this:

Do you want the piece of mind that comes from being a top seed and getting a pretty well defined level of opponent strength in the round of 32, or do you want to take a shot at getting a much weaker opponent, possibly even the weakest opponent ever faced by a No. 1 or 2 seed in the round of 32 in the KenPom era, No. 10 seed Alabama in 2006 (+11.05)? By saying yes to a No. 2 seed, you’re taking a one-in-three chance that your opponent in the round of 32 will actually be weaker than the average opponent for a No. 1 seed.

Moreover, the difference in opponent strength in the round of 32 for top seeds vs. No. 2 seeds shown above is, as you would expect, overwhelmed numerically by the gulf in the quality of the teams on the top two lines themselves.

Average team strength, 2002-19

                         KenPom AdjEM
No. 1 seeds                 +28.71      
No. 2 seeds                 +25.20                               

In basketball terms, what matters is how good you are and how good your opponent is. The question of whether the NCAA hands you a No. 1 or a No. 2 seed doesn’t change the first answer at all and has, as we’ve seen, a relationship to the second answer in the round of 32 that varies between weak and outright paradoxical.

What if we had eight regions instead of four? Would we ponder how the committee drew the line between the Nos. 8 and 9 teams? What if teams were numbered from 1 to 68? Same teams, same bracket, same rims, ball, court, etc. What would that seeding debate look like?

When I was being forced to learn about semiotics at length and repeatedly in grad school, I remember thinking it was rather a waste of time. The robust annual discussions of No. 1 seeds as something that really can make or break a team’s tournament chances, however, prove that I owe semiotics a big apology.

The signifier “1” next to a team’s name exerts a completely non-sports pull on the deepest and most fundamental portion of our brains that no amount of basketball information will ever dispel. So be it. MVP awards are a cool honor and fun to debate.