Stats class: understanding height

There are few things more talked about during a basketball telecast than height. Len Elmore rambles aimlessly talking about 7-footers. Dick Vitale has his All-Airport teams. You couldn’t go five minutes of a Kentucky without hearing about how long Anthony Davis was. Height matters. It does.

But how?

Does it help on offense? On defense? Do you want tall guards, a big front line, or do you just need a nimble 7-0 guy in the middle and who cares if he has stone hands?

In his groundbreaking book Basketball on Paper, Dean Oliver attacked this idea. To determine a team’s average height he weighted each player by the number of minutes played. After all, why care about the huge Lithuanian on the end of the bench if he never plays?

Oliver looked at NBA teams and found that 20 of the best 25 offenses over a 25-year time-span were taller than the league average, as were 17 of the best 25 defenses. Conversely, looking at the worst performing teams, 8 of the worst 10 (and 14 of 25) offenses over that time-span were shorter than average, as were 17 of 25 defenses.

So clearly, height has an effect.

Ken Pomeroy took that analysis one step further at Basketball Prospectus and calculated correlation coefficients for height effect on offense and defense. He found that there is indeed an correlation, but it isn’t terribly strong (0.27 on offense, 0.38 on defense). So he set out on a quest to find a more meaningful statistic and ended up filtering height down to just the biggest players on the floor. Of 100% of any games minutes, he looked at those played by the tallest 40% of any teams players. He called this ‘effective height.’

And effective height, it turns out, has a much stronger correlation to what’s happening on the court – but only on the defensive end.

Remember that based on average height, tall players accounted for 14% of the variation in defensive play (0.38 correlation coefficient). But just looking at effective height, it accounted for 24% of the variation in blocked shots, 23% of the variation in 2-pt%, 20% of the variation in eFG%, and 18% of the variation in defensive efficiency.

So if your team is coached by someone who covets athletic 7-footers who don’t know what to do with the ball (I’m looking at you Leonard Hamilton), that’s alright – just stick them in the middle and surround them by guys who can score.