Before analytics revolutionized the sport of basketball, many fans would judge the quality of a player by what is referred to as their “box score statistics”; statistics recorded for each game like how many shots a player took, the amount of times they rebounded the ball etc. In simple terms, players who did things that were quantified and showed up on the box-score (ie scoring a lot, getting a lot of blocks) were seen as great players and those with lesser box-score statistics were seen as weaker players.
Eventually, a statistic named Regularized Adjusted Plus-Minus (RAPM) was developed that changed the nature of how fans analyze player performance. Rather than look at boxscore statistics, RAPM tries to directly measure the only thing that really matters in a basketball game – how much a player contributes to outscoring the opposing team.
In short, RAPM measures how many points were scored by both teams when a player is in the game, using Bayesian inferences to adjust for the quality of the 9 other players on the court with them.
Rather than measure things that may loosely correlate with a player’s impact on winning, RAPM directly tries to measure impact. RAPM does not tell you how a player contributes or information like how good of a shooter they are; all it does is measure empirically how much a player impacted the score of the game while they were playing.
RAPM reveals a lot of fascinating things:
- Many popular players who score a lot and more generally have great box-score stats have negative impacts while many of the most impactful players have bad box-score stats. (i.e. Russell Westbrook vs Danny Green )
- The same player can make a hugely positive or negative impact, depending on who they play with (or against). (i.e. Jae Crowder)
- A great player who plays in a non-complementary way, can cause four otherwise great players to all have negative impacts. (i.e. Rudy Gay)
- Some of the most impactful players are not skilled players who are heralded and in fact rarely touch the ball, but players who do small mostly unseen things that enable other great players to flourish. (i.e. Nick Collison)
So how did this change how I see the world?
Whenever we are trying to optimize for something, we often look at what data can be quantified or easily analyzed. One of the takeaways from RAPM is to ask the question – how much of the story is actually being quantified and how much does that data actually correlate with whatever you are trying to achieve. If the data fails to capture a lot, and/or is only minimally correlated with whatever we are trying to optimize for, we often give it far more weight than it deserves.
Often in life, we act in a way that makes us feel as if we are being productive or look impressive to others rather than what most positively contributes to the goal. Sometimes we can add more value by doing less or by staying out of the way. More critically, sometimes we can be orders of magnitude more impactful by enabling other people to flourish than anything we could do ourselves directly.
RAPM helps me appreciate the importance of group dynamics and specifically, the role of complementary skillsets. Four work colleagues can be hanging out and having a terrible time, until a fifth colleague is added, who without talking much, sufficiently changes the dynamics that it makes the environment much more enjoyable. Conversely, four friends can be hanging out and having a wonderful evening until a fifth person is invited, who despite not saying much, changes the dynamics sufficiently that nobody is enjoying themselves anymore.
Sometimes in life, we feel down and that we are bad at whatever task we are focussed on. RAPM makes it clear the question is never an objective isolated good or bad, but are we working alongside the people and on the tasks that enable our skillset to be most impactful?
Would the Beatles be a better band if they replaced Ringo Starr with John Bonham? does this make John Bonham bad? Does this make Ringo Starr good? It’s an irrelevant question.
So often we focus on the question of what/who is better rather than given the dynamics at play and what is trying to be optimized for, what could have the most positive contribution.
In short, RAPM has made me appreciate the importance of asking, before all else – how can my actions most positively contribute to whatever I am trying to achieve.
RAPM developed gradually and organically over time, with contribution after contribution from basketball fans posting on their blogs and congregating on a public message board. A small number of passionate fans on blogs and a message board truly changed the nature of the sport. It took many years before the ideas raised on that message board started to permeate throughout the basketball world but over a decade later, the impact is clear. Many of the original bloggers/posters (some with basketball backgrounds, some with high status careers, some nobodys) ended up working in the NBA/media in very prominent positions.