As the field of the NCAA Tournament was unveiled on Selection Sunday, there was a lot of to-do about the job the committee did (or didn't do) in selecting the field. And while the last few teams they let into the field were certainly curious -- made even curious-er by the whooping Clemson put on UAB last night -- I maintain that the worse injustice is the pitiful job they did seeding teams.
There seemed to be little rhyme or reason or their decisions. Were they rewarding performance in the conference tournament? (UConn, Washington.) Or not? (Kentucky, Florida.) Was a challenging out of conference schedule important? (USC's inclusion, Virginia Tech's and Colorado's exclusion.) Or not? (UAB's inclusion.)
I started thinking about it all over again tonight when I read Nate Silver's awesome post over at his cool NY Times blog, FiveThirtyEight. Baseball fans will remember Silver from his previous life as a writer for Baseball Prospectus, most notably as the inventor of the PECOTA prediction system. He's been writing about politics for a couple of years now, but he's taking a break from that to take a look at the impact of seeding in the NCAA Tournament.
Specifically, he suggests the tournament has a pretty enormous seeding problem. Not, as you might imagine, in the performance of the committee; instead, he says there's a systemic problem with the setup of the tournament itself:
Are you better off [being seeded 10th rather than 8th]?
The answer is almost certainly yes: the No. 10 seed is intrinsically a better position than the No. 8 seed. So, for that matter, is the No. 11 seed. The 12th seed is also better than the No 8. As are the 13th and 14th seeds. And possibly even the No. 15 seed, depending on your objective.
Welcome to the strange intersection of bracketology and bracketonomics, in which the worse a team's seed, the better off it may be.
He crunches a crapload of numbers to prove his point, and I highly suggest you read the whole post. The idea is that because No. 1 seeds tend to be much better -- relatively speaking -- than the rest of the field, the longer you want to wait to have to face one of those teams so that you can advance deeper into the tournament. He goes in search of some answers, mostly in terms of fairer ways to seed the field. But there's a big part that I think he misses.
Before you can try to come up with a better system, you've got to answer one fundamental question.
Just what is the purpose of the NCAA Tournament, anyway?
The simple answer, of course, is that it's held to crown the champion of Division I basketball. But like most everything in life, it's not that simple.
Consider how many times each of the seeds have won the NCAA Tournament since its expansion to 64 teams in 1985:
- No. 1 seeds: 17 titles
- No. 2 seeds: 3 titles
- No. 3 seeds: 3 titles
- No. 4 seeds: 1 title
- No. 5 seeds: 0 titles
- No. 6 seeds: 1 title
- No. 7 seeds: 0 titles
- No. 8 seeds: 1 title
- Nos. 9-16 seeds: 0 titles
In 26 tournaments, the 104 combined No. 1 seeds -- making up 6.2 percent of the field -- have won 17 of 26 possible championships (65.4 percent). The rest of the field -- 1,024 teams, 93.8 percent of the field -- has won just nine titles (34.6 percent).
In truth, if the whole goal was to crown the champion in the most fair and efficient way possible, the field would be cut back to 32 teams again. Nobody's asking for that, so that must mean it's about more than just crowning a champ.
Now, I know we could answer whether it's truly about crowning a champ by pointing to one thing -- the play-in games -- and consider that as enough proof that this is about so much more than just crowning a champion. But stay with me, because there's a larger point to be made.
One thing it's obviously about is the crazy upsets that are highly entertaining. For sure, that has value. But I want to go in a different direction. Think about why there's so much uproar over the last few teams to be selected. Sure, there's the theoretical chance that one of them could win the tournament, but none of us actually believes any of these teams has a realistic chance to win six consecutive games. So why the big deal?
As we all know, it's because simply making the tournament has become the big deal over the years. (Something we WSU fans know all too well, thanks to how infrequently it's happened for us.) More often than not, coaches are hired and fired based not on whether they win national championships or even make Final Fours, but on whether they're selected for the Big Dance in the first place. Even if a team has no realistic chance of winning the whole thing, there's a symbolic achievement to being included.
There are other symbolic achievements, too. The largest, of course, is making it to the Final Four, but even that's rarified air. For many programs, simply winning one game is an enormous accomplishment. For others, it's winning a couple of games and making it to the second weekend. Again, this is something we Cougs know well -- we look back at 2008's Sweet 16 appearance with pride despite the fact we ended up nowhere near a championship and were run off the floor by North Carolina.
And that, to me, is the crux of the issue. Is the point of the tournament to simply crown a champion, and everything else is just accessory? Or have the other symbolic achievements taken on such weight that pretending they don't exist is foolish?
After reading Silver's piece, my gut reaction was to dismiss his criticisms out of hand, mostly because my thinking went along with the former: The tournament is to crown a champion. The seeding isn't set up to benefit teams seeded 7, 8, 9 or 10; it's set up to most benefit the highest seeded teams. Sure, it's unfair that the 8 and 9 seeds have to face a No. 1 seed in the second round, but that's the price of creating a "fair" bracket.
But the more I think about it, the more I think Silver's point has a lot of merit.
Ignoring reality in any part of life is silly, and the same goes for the NCAA Tournament. Because only about 25 percent of the field has a realistic chance of winning it, how far you advance has indeed become meaningful. And because that's become meaningful, I think it's only right to think about alternative ways the field could be seeded to better reflect this. When a team that is an 8 or 9 has less of a chance of making it out of the second round than a 12 seed who beats a 5 seed and then faces a 13 seed who beat a 4 seed, we need to rethink the system.
Silver's most intriguing solution is to seed the field like a tennis tournament -- slot the top teams, then randomly draw everyone else. While this certainly would preserve the craziness of the tournament, I don't like it. As Silver points out, the biggest problem with the current system, on a theoretical level, is that there's incentive to lose in order to avoid the 8/9 line. He acknowledges that's unlikely for anyone to be able to carry out a plan to become a No. 11 seed very precisely, but the fact that the incentive exists is problematic in and of itself. His solution presents an almost equal issue, though: It removes some of the incentive for a team to be the very best it can be because it might get stuck playing a good team in the first round anyway. Imagine Ohio State facing West Virginia in the first round because of the luck of the draw.
I don't think we should be introducing more luck into the bracket than there already is.
So, I started thinking about something outside the box where the incentive for a team to be as good as it possibly can be is enhanced from what it is now, but also where the achievement of making it to the second weekend is inherently recognized as its own achievement. Here's what I came up with.
Continue the 1 through 16 seeding system in each region.* Create four pods, each led by the top four seeds. Why the top four seeds? Because, as we've seen, they're the ones most likely to win the tournament, and this system will benefit them in that endeavor.
*For the purposes of this exercise, I'm assuming 64 teams. I'm not going to waste brain power trying to figure out how to make my perfectly logical system fit in the with NCAA's current made-for-TV setup. Not worth my time.
However, instead of seeding the teams in the current fashion -- in which the No. 1 seed is guaranteed to play no better than the average (No. 8 or 9) seeded team in the second round, the No. 2 seed guaranteed to play no better than a slightly above average (No. 7) seed in the second round, and so on -- place the three worst teams in the same pod as the No. 1 seed, the next three worst teams in with the No. 2 seed, and so on. Here's how each pod would look:
|No. 1 Seed Pod||No. 2 Seed Pod||No. 3 Seed Pod||No. 4 Seed Pod|
In the first round, the top seed in each pod would face the worst seed in the pod, and the two remaining teams would face each other. The winners would then face each other in the second round for the right to advance to the second weekend. The winner of the top pod would face the winner of the fourth pod in the Sweet 16, while the winners of the two middle pods would face each other.
Here's what my system would look like this year, using the seeds already established by the Selection Committee. Let's test my system against our two original criteria.
Does it incentivize teams to achieve the best seed they possibly can? Yes. Absolutely yes. Let's start by taking a look at the title contenders.
Instead of a No. 1 seed facing a terrible team in the first round and an average (No. 8/9) team in the second round, it will face two pretty terrible teams in the first weekend. The incentive to be a No. 1 seed becomes even higher than it is today, when it's already pretty high, but not high enough. Consider the case of Kansas last year. Think the No. 1 seeded Jayhawks wouldn't have loved to change places with No. 6 seeded Tennessee in its own bracket? The Vols needed only to beat No. 11 seeded San Diego State and No. 14 seeded Ohio -- thanks to the Bobcats' upset of Georgetown -- to get to the second weekend. Meanwhile, after whipping No. 16 seeded Lehigh, the Jayhawks had to face a dangerous Northern Iowa team in the second round. This system rightly rewards the No. 1 seeds for their accomplishment by essentially giving them a walkover to the second weekend.
The No. 2 seeds? Their first opponent will be slightly better than what they would have faced under the traditional format (No. 13 vs. No. 15), and it's a not insignificant difference. However, they'll face a worse team in the second round than they would have under the previous format (11 or 12 rather than 7 or 10). It's the same for the 3 seeds. And while it might be tempting to think the 3 seeds are getting hosed because their first round game is now against a 10 instead of a 14, Silver shows that the difference in quality between a 10 and 14 is basically the same as the difference between the 13 and the 15.
If anyone has a complaint, it's the No. 4 seeds, but my response is going to sound rather snotty: Tough crap, fellas. If you were a serious contender for the title, you would have earned your way up to a 1, 2 or 3 seed, because, as we've seen over the past 26 years, only three teams seeded 4th or lower have ever won the tournament. So I'm not really worried about what this is going to do to their title chances. Besides, they should still be better than both the teams they'll play.
It's at this point -- when it's time to talk about the rest of the seeds -- that we need to revisit Silver's original dilemma, while simultaneously tackling our second criteria.
Does this system inherently recognize the achievement of making it to the second weekend as a worthy accomplishment? Yes. Here's how: With each move up in seed line, the path to the second weekend becomes easier in the sense that you begin to avoid teams that are likely to be substantially better than you until the second weekend.
Sure, the 14 and 15 seeds have an advantage in their first game over, say, the substantially better 7 seed who will face a 4 in the first round. But we've already established that making it to the Sweet 16 is the achievement we're shooting for, not merely winning one game, and one cannot argue that in this respect, it's much, much better to be the No. 7 seed, which will potentially play two teams that are near its ability level. If we look at the 8/9 seeds -- the original impetus behind Silver's piece -- they are likely to face a No. 3 seed for the right to go to the second weekend after again facing each other. That's no easy task, but it's not the virtual death sentence that the second-round matchup is under the current system.
To win two games and get to the second weekend, any team not seeded in the top four is going to have to win a game against a similar team and a superior team -- the only difference is the order in which the games are played.
This system also removes some of the luck currently introduced into the bracket by the committee when they move teams up or down a seed line to conform with bracketing principles. In my system, teams are essentially broken into "tiers" after the first four seeds. Yes, there's an "advantage" to being a 15 instead of a 16, but since we've really got our eyes on the second weekend, the only real difference is between 14 and 13, 11 and 10, 8 and 7.
This is beauty of it, and why I think it trumps the current system: The higher your seed, the more reasonable your path to the second weekend, and from there on out, the process of determining a champion can truly commence over the final four rounds of the tournament.
Of course, as we wrap this up, there's one undeniable fact that can't be ignored.
It will never happen. Not my idea, not Silver's idea ... none of it.
My system removes some of the inherent fun of the tournament in removing some of the chances for upsets that make this contest so unique. Although I think the "upsets" are overstated a times given the number of first round games that are basically not competitive, this system sacrifices some of the opportunity for the memorable upset in the name of a more fair system across the board. That will never fly with TV suits, which relish the idea of viewers tuning in because something might happen.
And while Silver's most interesting system certainly would be wildly entertaining, it introduces so much luck, the coaches would never go for it. I do, however, think Silver's idea would have a lot of merit if the tournament was ever substantially expanded to 96 or 128 teams.
As long as the NCAA is making piles of money with the current setup, there's just no incentive to change it for the better. Only changing it for the worse to increase inventory is really a plausible option.
Still ... we can dream. Right?