The Best Team Won…Because It Always Does, by Gabe Sokoloff

Each championship system values different qualities: the BCS values the intangibles factored into a poll of coaches, the single-elimination NFL playoff values clutch performance, and the best-of-five/seven series MLB playoff values tenacity (to name three). Most of us prefer one above the others, and I am interested our ability to assess these systems.

Let me clarify my purpose from the outset. I am not interested in discussing which system is best. That discussion might go: "tenacity is important...but clutch performance is the true mark of greatness. So, the NFL’s championship system is better than MLB’s". Instead, I am interested in what it really means for a system to be good or bad. The point of departure is this puzzle: we judge a system as good or bad based on its ability to identify the best team, and yet we define what qualities belong to a “best team” by referring back to that system.

A quick side-note: the subject here is both on the dry side and is unusually 'meta' for sports journalism, so I want to explain why I'm writing this at the outset. My reason is that complaints about how “the best team didn't win it all” – how a championship system often fails to reward the most deserving team – ARE common fare in sports journalism. I think then that it is important to look beneath the surface of these comments and figure out what all these terms really mean, in the hope of allowing for more meaningful complaining, especially with the Patriots’ upset of the Chargers just behind us and another Super Bowl just in front.

So, back to the puzzle. Our first premise is that a system is supposed to reward the best team. This point isn’t too controversial. But we usually mean it in a particular way: we think that the system is not creating the best team, but rather identifying it, just as a timed Olympic race identifies who is the fastest man. The notion of “fastest man” is somehow outside, somehow bigger than the test itself, which is just a means of finding to the result that was somehow always there. Even without a race, one of those men is still the fastest. Similarly, we tend to think that, system or not, there is a best team, and the championship system is just a more complex stopwatch for this more complex race among teams.

Our second premise – that we define what makes a best team by referring to our system – is as intuitive as the first. Consider this: to be the best NFL team, no one would dispute that it is more important to have a good record in games than a good record in halves-of-games. But we think this because our system rewards teams that perform well for full games, not halves. Similarly, we all believe that a team’s greatness over a season is indicated by its won-loss record, not by its points differential. But again, we hold this opinion because our system specifically rewards those teams that perform well in full games and does not (directly) reward teams for their points scored/ points allowed.

So we see that our system does not merely ‘identify’ the best team, in the way a stopwatch identifies the fastest man, but actually shapes what we consider the best team to be. But if this is true, then what of claims like "the best team didn't win it all? How can the best team not win if the best team is, by definition, the team that wins? That's the rub.

Two quick suggestions for how the best team may fail to win come to mind, but neither holds up to scrutiny. The first is that practical considerations impose a finite length on a season and playoff and so our sample size of games is too small to reliably reward the deserving winner. But this view assumes, like in science, that a larger sample size is always superior to a smaller one. But sports are not science, and this view ignores the value of winning when it counts most. We all agree that a great team is not simply one that wins nine times out of ten, but the team that wins the one time that matters.

A second (and related) answer might be that sometimes the best team doesn't win because it is the victim of bad luck. Here, we must take a hard look at our term “luck”. I would divide “luck” into two categories, which I call “in-the-game luck” and “outside-the-game luck”. When a defensive back makes an interception he would make only one time out of one hundred, his team benefits from “in-the-game luck”. A team that wins a game on a blown call benefits from “outside-the-game luck”.

“Outside-the-game” luck occurs when something prevents the system from working properly, such as a blown call, or if a player breaks the rules and doesn't get caught (e.g. steroids). In this case it is true that the best team might not win, but it is not the system's fault; instead, the system’s rules were violated. In other words, if a team loses on a blown call, we do not have a case of the best team not winning, but rather a case of an asterisk-tainted “winner”.

“Inside-the-game” luck occurs when something very unlikely influences the game – say, when the opposing team’s defensive back makes a game-winning interception he would only rarely make. It is tempting to say that, in such a case, the best team would not win. But if a player makes a play he would make only one in a hundred times, it is still the result of skill in some sense, as it’s possible that most players (or at least most people) would only make it one out of a thousand times.

I think that “luck” is a perfectly sensible term that we use meaningfully most of the time. I only mean to argue that it is beyond the ability of language to separate out factors like luck and skill as clearly as we often think we do. The fact is, once you are talking about real-life teams playing real-life games, terms like “skill” and “luck” are clumsily pieced together, and break down a whole into parts in a way that can never really work that well. Luck and skill are everywhere and they are nowhere. We might want to believe that the truly best team is the one that would win in a luck-free environment, but “luck-free” is nonsense, as are the notions of a team whose “raw skill is highest”. For these terms to have meaning there must be things in reality to which they correspond, but in real life no such thing exists. In our imaginary case of the once-in-a-hundred interception the arbitrariness and ultimate impossibility of deciding whether it was a result of luck or skill is very clear.

Given this, it’s hard to say that “luck” keeps the best team from winning; it is rarely if ever that simple. So I can now state, hopefully even more convincingly than before, that what makes a best team best are results only, where “results” = a team's status as decided by a system, despite any claims about luck or skill. The statement “the best team did not win” is self-contradictory, and so then are critiques of championship systems that are based on that idea. A championship system defines, not identifies, the best team, and can therefore never be mistaken.

One final question: how do we reconcile this view with the fact that we simply DO tend to prefer one championship system to another? I myself have always preferred the NFL’s single-elimination playoff to the MLB’s best-of-however-many games system. My reason is that I like the idea that each team has only sixty minutes of play to show itself to be superior. For whatever reason, I value that highly – perhaps I myself find clutch performance more personally challenging than tenacity, and I value in others what I myself lack. Whatever the reason, my preference for one system over another is completely personal and has no objective basis, as is the case with anyone’s preferences. These matters of taste are irrelevant to whether a system rightly or wrongly rewards the best team. I hope the reader does not resent that I have taken such a roundabout way just to illustrate that our mere preferences are mere preferences.*

*In this article I have dealt only with those of our preferences between systems that pertain to the qualities of a team themselves. There are other completely different kinds of reasons for preferring one system to another. For example, one might hold that the college football coaches’ poll is flawed because coaches may be biased. This is not the sort of complaint addressed here (and a complaint that can, I think, have an objective basis).