The Parity Ideal

John Urschel, Offensive Lineman / Baltimore Ravens - The Players' Tribune

The idea that any NFL team can win is as central to the NFL as the idea that any American can get rich is to America. It’s been the justification for the league’s efforts to redistribute talent from weaker to stronger teams ever since the 1930s, when Bert Bell, who would go on to become NFL commissioner, led the drive to let the worst teams draft first. It’s still the explanation for the league’s hard salary cap and low rookie salaries.

The commissioner’s “core responsibility,” Roger Goodell told the Wall Street Journal during the last collective bargaining agreement negotiations, is “making sure our game is competitive.” Imagine the alternative: a league in which the highest payroll dwarfs the smallest — like baseball, where the Dodgers have a payroll four times higher than the Marlins.

All four major pro sports leagues engage in revenue sharing and have some rules related to redistribution, but there’s a big difference in how extensive and severe their policies are. (The NFL and NHL, to give one example, have hard salary caps — no exemptions — while NBA and MLB have soft ones.) According to Goodell’s logic (and conventional wisdom), without a strict redistributive system, the league would have significantly greater disparity. Instead of Any Given Sunday, the NFL would have the Yankees.

There are different ways to measure parity, of course. Its vagueness is one of the reasons it’s a convenient goal. But one simple way to think about it is to see how teams perform relative to the mean — a .500 record. You’d expect a league with good competitive balance to have more teams clustered around .500, while a league of haves and have-nots would have greater variance. What’s more, leagues set up to redress competitive imbalance — taking from winners and giving to losers — should be mean reverting (that is, their records should move in the direction of .500 in subsequent seasons). The worst teams would get better, and the best teams would get worse. In a league with fewer controls, on the other hand, the winning teams should be able to use the advantages that helped them win in the first place to sustain their dominance year to year.

That’s the myth, at least. But how true is it? To get a clearer picture, I looked at the regular season records of every team in each of the four major pro leagues from the last decade to see how they performed relative to the mean

[NOTE: The NHL follows a point system. For the sake of analysis, I treat any loss, whether in overtime or regulation, as a full loss, and do not take the single point a team may gain into account. The win percentage is given by wins divided by losses.]

The plots are intriguing, although they mostly reinforce what the astute multi-sport fan already knows: the NFL has the largest spectrum of records, followed by the NBA, then the NHL, and, finally, MLB. And yet, baseball is the sport that is famously known for competitive imbalance.

Although that is not exactly what has caught my eye in these plots. Look again. They are all slightly left-skewed.

[NOTE: the NFL plot does not follow the strict left-skewed distribution, due to the large number of .500 teams, although the median is closer to being above .500 than below, in a combinatorial sense.]

What does that mean? In these terms, it means there are more above-average than below-average teams, but more very bad teams than very good ones. Don’t believe me? Look at what happens when I divide the data for each league into four groups.


This is what you want. The majority of teams slightly above average and very few teams completely dominant. A type of hopeful parity. The kind that fans and cities can get behind.

Across all four sports, the results become even more startlingly similar when you look at how teams performed the following season. I’m talking about mean reversion. No matter the league — and no matter the kind of collective bargaining agreement — roughly the same percentage of teams in each league moved toward .500 than away from it the following season.TowardAway500

Despite the major pro sports leagues having different systems, their level of parity may not be that different. It may be that if NFL players commanded the kind of income that NBA players do, or if the NHL had the kind of salary cap MLB does, these graphs would look very different. I can’t say for sure.

Still, they’re striking. And they suggest something about the myth of parity.

It’s in the NFL’s interest, of course, to make the public think that its handling of players — who, despite playing the game that generates by far the most revenue of any American sport, have by far the lowest salaries, shortest careers, worst prospects of serious injury, and the most non-guaranteed contracts — is necessary to achieve some noble ideal.

It’s more convenient to do so than to own up to the fact that the policies that lead to greater balance among owners massively increases the imbalance between owners and players. This isn’t to say that NFL owners are more greedy than other sports owners, as anyone remotely familiar with the history of lockouts and court orders in sports can tell you. The Yankees haven’t given Alex Rodriguez contracts worth nearly half a billion dollars out of the goodness of their hearts. But for a variety of factors, in other sports, players have more bargaining power.

My guess is that there’s another reason the myth is so persistent: most people don’t think of parity in terms of the mean — in terms of being average. Most people just want to think that it suggests that winning means something, that conditions were fair, that sacrifices were made. We want to believe that any given team has a shot.


Math Meets Football: One in 600 Billion

When I performed my analysis, I found something particularly fascinating. There was a clear outlier when it came to academic majors in college football.

Read More