Lawler’s Law states that the first team to 100 points wins. It’s catchy, simple, and often correct.
I also hate it.
Full disclosure: This was not supposed to be a hit piece on Lawler’s Law. I fully intended to use the law as an entry point into a broader discussion on how to predict wins, with a hopeful eye towards predicting Bucks’ wins in the playoffs. But I was sucked down Lawler’s rabbit hole, and long story short, the entry became the entree.
I’ll provide a brief recap of Lawler’s Law for the uninitiated before diving into my critique. The critique has three parts of increasing importance, each of which I connect to a point with broader relevance than a random law coined by a random commentator. I’ll wrap up with a consideration of predictions writ large as well as a plug for a more acceptable alternative to Lawler’s Law: the no-less-egotistically-named Morgan’s Motto.
Let’s start with some background. Ralph Lawler started commentating with the (San Diego!) Clippers in 1978, a post that he continued to hold for many decades. However, he did not start with an empty quiver. From his days of commentating for the Sixers, he remembered that Al Domenico, a Sixers’ trainer, had a saying: the first to 100 always wins. Lawler airlifted the maxim to San Diego, branded it with his convenient last name, and brandished the arrow successfully; Lawler’s Law was born. It became his catchphrase and well-known around the association. I encourage you to check out this nifty website that tracks all things Lawler’s Law.
(In case you’re wondering, Wikipedia lists two additional axioms that comprise the scientific Lawlerverse: Lawler’s Law Corollary, which states that the team that shoots greater than 50% from the field will win, and Lawler’s Overtime Law, which states that the first team that scores four points in overtime will win.)
With the boring, unbiased information out of the law, let’s get on with it.
The first prong of my attack against Lawler’s Law is the simplest. Lawler’s Law predicts which team will win the game. Teams wins games by scoring more points than their opponents. To predict whether a team will score more than the opposing team, Lawler’s Law posits that the team must be outscoring their opponents when they (the team) hit 100 points. In other words, we can predict that the Bucks will win because they are winning. Groundbreaking. Moreover, it only comes into effect late in the game, often when one team already has a chunky lead.
A parallel can be drawn to the distinction between latent and manifest variables. Manifest variables are things that are readily observable (e.g., the length of a basketball court). Latent variables are the opposite (e.g., the likelihood that a team will win a game). Social scientists know that they can’t measure intangible concepts like happiness exactly. Instead, they measure indicators of happiness - tricky stuff, like asking folks, “Are you happy?” - and use statistical modeling to infer the value of the underlying latent variable.
My issue with Lawler’s Law can be reframed as a case where the manifest variable - the first team to score 100 points first - is too close to the latent variable - the likelihood that a team will win. To be sure, to measure a latent variable, you want manifest variables that are related to the latent variable. But part of the interest in measuring latent variables is that you can see which manifest variables are its best indicators. By having a manifest variable that is so closely linked to the latent variable, it crowds out the role of other potential manifest variables that would be more interesting (read: less self-evident).
My second bone to pick is that the arbitrary threshold of 100 is a product of when Lawler first used the term. In the late 1970s, in the rare instances that teams eclipsed 100, they almost always prevailed. However, due to three-pointers, officiating changes, and a motley crew of other factors, the success of Lawler’s Law has waned over time, especially in the last five seasons. Interestingly, a piece in the LA Times found that a threshold of 114 would more accurately reflect the modern game. (For interested parties, such as those participating in data analytic hackathons, the reporter shares the nuts and bolts behind this calculation on GitHub.)
I want to link this point to discussions in this venue of how the game has changed. I have argued previously that basketball has not changed much over time, which seems contradictory with the above contention. These arguments, however, have different scopes. My prior assertion had more to do with how the game was played and whether players could succeed in different eras. Here, I mean it in a more narrow and factual sense: teams score more. This can be attributed to specific ways that the game has changed (including threes and officiating as referenced above), rather than a general sense that things were different back in the good ole days.
The law’s changing efficacy over time highlights the third and what is arguably its most crucial drawback: the tradeoff between scope and accuracy. If we set a lower threshold, it would be applicable to more games. Indeed, Lawler’s Law did not apply to many games in the 80s and 90s, and still misses the occasional slugfest in the modern era. However, it would also lose predictive power. To take the extreme case, Morgan’s Motto could be that the first team to 1 wins. This would always apply, but unlikely to significantly outperform a coin-flip in terms of accuracy. In contrast, if we set a higher threshold, it would be more accurate. The first team to 150 is likely to win. But we would not be able to apply the law to many games.
This tradeoff is central to scientific theory-building. When crafting theories - explanations for how something happens - scientists must balance a tradeoff between accuracy and scope. If I want to explain a specific phenomenon with incredible precision, then it will be harder for my explanation to generalize to related phenomena. If I want to explain a wide variety of phenomena, it will be hard to provide details of what is actually going on. Arguably, this is the central tension in science. We want to create wide-ranging theories that have maximal impact across domains, but in doing so they might not actually be helpful in these areas. We want to provide detailed explanations of things, but in doing so we might not be able to extend our knowledge to other facets of life.
The point of this article is not just to rain on Lawler’s parade. It is to rain on Lawler’s parade and observe what the resulting fracas reveals about how we predict which team will win a game. When making these predictions, it is important to consider the difference between latent and manifest variables, changing historical tides, and the tradeoff between scope and accuracy. To ignore these factors results in rules of thumb that are self-evident, dated, and often inapplicable: in other words, Lawler’s Law.
And that’s okay! As much as predicting victory has spawned industries (sports media, sports betting, and so on), the fundamental reason that we watch sports is that we don’t know what will happen. As nice as it would be for the Bucks to repeat ad infinitum, it would get boring. For all its flaws, Lawler’s Law would be even worse if it were always accurate.
With this backdrop, I’d like to return to Morgan’s Motto. It is always applicable. It stands the test of history. It does involve one team being ahead, but is so far removed from the end of the game for this concern to hold weight. It even avoids the fallacy of calling itself a law; Lawler’s Law is not a law because it is not always correct (see: gravity).
Most importantly, though, it is incredibly inaccurate (and likely biased towards teams with big men who win the tip-off to boot). Based on the nuts and bolts referenced above, the first team to score wins approximately 55% of the time (which is honestly higher than I expected!). Thus, Morgan Motto’s has a tantalizing shred of accuracy, yet preserves the sense of unpredictability that entices us to tune in to basketball in the first place.
I therefore stand for Morgan’s Motto and hope you will join me. (And if someone else has already had this idea? I’d only be following in Lawler’s fabled footsteps.)