Will Tonight’s Champion Really Be the Best Team?

A thing sometimes asked of sports is whether the playoff format succeeds at crowning the best team as champion. This is asked of Major League Baseball from time to time. It’s asked of NASCAR. It’s been asked of the NHL, and of the NFL, and it’s never asked of soccer leagues like the English Premier League because that one, at least, doesn’t have a playoff. It’s never asked of college basketball, which is odd, because if I had to think of the most random conceivable playoff format, it would look a lot like a 64-team, single-elimination bracket without any byes.

I’d guess people don’t ask this of college basketball because the NCAA Tournament is so fun. That’s part of it, anyway, and it’s a fair stance. You could switch to a tournament format involving every league’s regular season champion, perhaps a few wildcards as well, and byes for the best teams, but in that style of tournament, Saint Peter’s and North Carolina and probably even Arkansas wouldn’t have made the field. At the same time, though, part of why this isn’t often asked is probably that it doesn’t seem necessary. The best team doesn’t always win the NCAA Tournament. But it often does.

KenPom is not, of course, the end-all-be-all for evaluating college basketball teams. It’s very good, though, at indicating how well they should be expected to perform in their next game, so good that it almost always parallels the famously accurate Las Vegas odds within a few points. Going off of KenPom, in these twenty KenPom-era NCAA Tournaments, the best team has won ten times. Half the time.

That’s more than should be expected. Gonzaga, clearly the best team on paper entering this tournament, was less than 30% likely to win it all heading into the first round. Some years, the best team in the country is a bigger favorite than that, some years, they’re a smaller favorite, but that 30% number is probably a generous estimate of the annual likelihood the best team wins. Somehow, though, the best team keeps winning.

Why might this be? Part of it could be the small sample. Twenty of something is not very many. Part of it could also be what KenPom is built to do. By referencing a system designed to evaluate teams in a forward-looking manner, recency becomes important, and a team that’s just won six straight games will naturally have strength in their recent results. Were we looking at a hypothetical system that evaluates the whole of teams’ play, over the season, the result might be different. KenPom doesn’t measure who the best team was in a given season. It measures who the best team would be if one more game were to be played.

Another part could be that our understanding of probability over the course of this particular tournament is lacking. It could be that consistency is more important than we realize. It could be that being a well-rounded team is of high importance, as many say with the “___ percent of national champions ranked in the top 25 of both offense and defense on KenPom” stat. It could be that something special happens in the NCAA Tournament which leads the best team to play even better than themselves. It could even, as many salivating fans might love to read, have something to do with officiating.

Here are the final KenPom rankings of the national champion in each of the system’s 21 years. Should Kansas win, they’ll likely be ranked 2^nd or 3^rd. Should UNC win, they’ll likely be ranked 15^th or 16^th.

A thing about this list that jumps out even more than the relative dominance of 1^st-ranked teams is how many champions finished in the top three and that none finished between 4^th and 7^th. This is likely just small-sample stuff, but a question can fairly be asked as to whether there’s something meaningful here. Is this the famous Kemba Walker trope, applied outward to Carmelo Anthony and Shabazz Napier? UNC would match the surprise factor of teams like these, and perhaps exceed it. Were it not for that Baylor game, we might look back on Caleb Love the way we look back on Anthony. In fact, we might look back at Love that way anyway. Meanwhile, Kansas would match the baker’s half-dozen of top-three champions not ranked first. There’s nothing too historic, in this sense, that could happen tonight.

The associated question with all of this is what happened to the top-ranked teams that lost. How good were these teams, who managed to hold on to the top ranking despite losing within the tournament? Here’s the list of all 21 who’ve finished at the head of the pack. Gonzaga will almost certainly finish there again this year.

Again, there’s an odd clustering of teams outside the center here. Two Championship Game losers, two Final Four losers, Two Elite Eight losers, Four Sweet 16 losers. Is something up with the Sweet 16? A possibility here is that the teams that lost in the Sweet 16 may have seen their ranking dip more by winning and continuing to play. Perhaps KenPom really did have them overvalued and while the loss hurt, underwhelming close wins would have hurt more. In 2014, Arizona was rated as the best team in the country going into their Sweet 16 game, and they won that game narrowly over 23^rd-ranked San Diego State, after which they dropped down to being ranked 2^nd, where they then stayed.

Really, this is as much an exercise in understanding KenPom as it is asking about the tournament’s ability to crown a true champion. But that does bring us back to the original question, by making us ask how to determine who the best team is in the first place. One way is KenPom. Another is a 64 (now 68)-team single elimination tournament without byes. It’s a subjective thing.