Meaningful Basketball Analytics Must Answer Two Questions: "Is it True?" and "Why Does it Matter?"
I analyzed NBA.com's All-2010s Team and made my All-2010s Team selections, so I am not going to specifically address the various other All-2010s Teams that have been published and publicized. I wrote about the broader topic of how to best evaluate players in Objectively Assessing Talent Versus Issuing "Hot Takes, but this subject is worth revisiting, specifically regarding the way that some of the skills that I honed in law school are applicable to basketball player evaluation.John Terzano, my law school professor and co-founder of the Nobel Peace Prize-winning Vietnam Veterans of America Foundation, explained that every word written by a lawyer must fulfill two criteria: "(1) Is it true? and (2) Why does it matter?" He suggested that we keep those two thoughts foremost in our minds as we wrote out law school papers and, especially, when we wrote our Bar Exam essays. That advice helped me to earn a spot on the Moot Court (appellate advocacy) team and it helped me to pass the Ohio Bar Exam.
Professor Terzano told me very early in my law school career that to be successful as a legal writer after having a background in journalism I would need to forget/relearn everything that I know about writing except for spelling and grammar. Journalistic writing and legal writing involve different skill sets. Effective legal writing, from a structural standpoint, consists of stating the relevant facts, citing the applicable law and then analyzing how that law applies to those facts, thereby reaching a conclusion that (hopefully) persuades the judge and/or the jury.
Those lessons from Professor Terzano not only helped me to become the lawyer that I am today but also, in retrospect, have honed my non-legal writing skills as well. Basketball writing will almost always have more flavor and spice than legal writing, but it is not a bad idea to review any piece of writing by asking, at key points, "(1) Is it true? and (2) Why does it matter?"
If a team wins several games without the services of a star player, why does that matter? The absence of one player is not the only factor that determined the outcome of those games, and a small sample size of data is inherently a less reliable basis for drawing meaningful conclusions than a larger sample size of data.
Consider a storyline that received at least a fair amount of media coverage during the 2018-19 NBA season: Toronto's won-loss record without Kawhi Leonard.
During the 2018-19 regular season, the Toronto Raptors had a better winning percentage without Kawhi Leonard (17-5, .773) than they did with him (41-19, .683); that is a true statement, but "Why does it matter?" must be asked. The 17 wins without Leonard included two wins each against the New York Knicks (17-65), the Chicago Bulls (22-60), the Atlanta Hawks (29-53) and the Washington Wizards (32-50). Other wins without Leonard came against the 19-63 Phoenix Suns, the 33-49 Memphis Grizzlies, the 37-45 L.A. Lakers, the 39-43 Sacramento Kings and the 39-43 Miami Heat. The Raptors went 13-1 in games that they played without Leonard versus teams with a .500 or worse record but they went just 4-4 in games that they played without Leonard against teams that finished with a record better than .500; the Raptors did very well without Leonard against bad teams but they did not do so well without Leonard against playoff caliber opponents.
The subset of data from the games that Leonard did not play does not suggest that the Raptors are better without Leonard than they are with him, but rather that the Raptors employed their "load management" strategy primarily against weak teams that they could defeat without his help. The Raptors went 18-16 in games that they played with Leonard against teams that finished with a record better than .500.
One could take this analysis further by looking at which other players sat out those games, which teams were playing back to back games or four games in five nights, and so forth.
Without even taking those further analytical steps, though, we have already seen enough to state with confidence that Toronto's excellent regular season record last season without Leonard is true, but it does not matter in the larger context of correctly evaluating Leonard's value and the team's overall strength.
A player's value cannot be meaningfully and definitively assessed based on a handful of games. One must consider a variety of factors about those games overall, and also about the skill set of the player. Kawhi Leonard's prowess as a two-way player may not be needed to beat sub-.500 teams, but it was vitally important when Leonard played stifling defense while unleashing a torrent of 30 point games as the Raptors defeated Orlando, Philadelphia, Milwaukee and Golden State in the playoffs (Orlando, Philadelphia and Milwaukee are three of the five teams that beat the Raptors during the regular season when Leonard did not play). Correctly evaluating basketball players involves much more than looking at a couple of numbers and pretending that you are a mathematical/statistical guru.
One good place to at least begin the search for basketball player evaluation wisdom is by looking at any piece of data—statistical or anecdotal—and applying the two-part Professor Terzano test: "Is it true?" and "Why does it matter?"
Labels: John Terzano, Kawhi Leonard, law school, NBA, Toronto Raptors, writing
posted by David Friedman @ 10:45 PM
