By Henry Johnson
Nobody likes a ball hog. Neglected pickup teammates mutter under their breath about the hero at the local gym. The sentiments extend to the collegiate and professional levels, where post-game analysts regularly chalk up victories to scoring balance. The basketball world seems to have reached a consensus long ago: there’s a special circle of hell reserved for those who take an unfair share of their teams’ shots.
But is scoring imbalance actually detrimental? To test this, I gathered regular season data from the 1995-1996 NBA season up to the present. By looking at winning percentage and some measurement of scoring balance, it’s possible to determine whether sharing the ball is linked with regular season victories.
Finding a suitable measurement, however, is a bit hairy. Standard deviation of points per 36 minutes is a tempting metric for scoring balance, but this isn’t an entirely fair number. All else being equal, teams that score more points will show less balance, just as a dataset of (2,4,6) has a higher standard deviation than a dataset of (1,2,3). This method might only lead to the obvious conclusion that teams who score more points win more games.
Luckily, it’s possible to adjust for this fact by calculating each player’s points per 36 minutes and dividing by team points per 36 minutes to get his share of team scoring as a percentage. The standard deviation of this new number should function as a good measurement of scoring balance.
As it turns out, there is a positive correlation between winning percentage and standard deviation of scoring. The correlation coefficient is 0.132 over 562 team-season. An OLS regression on winning percentage returned a coefficient on the metric that was significant to the 1% level, strongly suggesting that directing scoring towards your top scorers does indeed boost your chances of winning.
Still, it’s possible that this relationship is driven by top players contributing to both imbalance and success. Superstars not only make teams good, but also do a disproportionate amount of scoring, so it stands to reason that the link should be strong.
When I remove each team’s top scorer from the data, the correlation coefficient drops down to 0.047, which supports the idea that leading scorers were exacerbating the relationship between imbalance and winning. A similar regression to above only returned a p-value of 0.26 of the coefficient being significant.
So which teams should find solace in this news? When top scorers are included, teams with the least balanced scoring are Oklahoma City, Chicago, and Charlotte. On the other end of the spectrum, Denver, Detroit and Washington are most effective at spreading points around.
Once top scorers are removed, Chicago is the least balanced, followed by Oklahoma City and Charlotte. Denver, Detroit, and Washington remain the most balanced when leading scorers are taken out of the equation.
Overall, the data suggests what we all know: the NBA is a superstar’s league. There is a small subset of humans who can be an NBA megastar and if your team is lucky enough to have one, give him the ball.