The NBA’s marginal revolution in triple-doubles

by on March 5, 2016 at 3:05 pm in Data Source, Economics, Sports | Permalink

If [Draymond] Green is truly gunning for triple doubles, there should be an increase in his production in the third stat category in games once he has secured a double-double. When Green has locked up a points/rebounds double-double, his assists per minute increases to 0.27 from 0.21. Similarly, when he has a points/assists double-double, his rebounds per minute rises to 0.44 from 0.28 and when he has an assists/rebounds double-double, his points per minute goes up to 0.69 from 0.40.

Here is the WSJ piece, which covers other players as well.  I still think that sports as a medium for teaching both economics and statistics to younger individuals (and others) remains a somewhat underexplored opportunity…

1 Balop Oakes March 5, 2016 at 3:52 pm

Analysis might be flawed.
Need to compare against averages of games where he had triple doubles before he got to a double-double.
You would expect his average attempts to be higher on good games.

2 Jaunty Rockefeller March 5, 2016 at 6:23 pm

How would he ever have a triple double before having a double double?

3 too hot for MR March 5, 2016 at 7:24 pm

Well certainly he wouldn’t with that kind of negative attitude.

4 Dan March 5, 2016 at 3:54 pm

They chose to focus on Green because he had the largest number of excess triple doubles, so the effect size is biased upward. Even if there was no causal effect, we would expect the player who ended up with the most excess triple doubles to have done better than usual in the third stat after securing a double double. (By random variation, some players will do better than usual in the third stat after securing a double double, and some will do worse than usual, and one of the players who does better than usual will wind up being the one with the most excess triple doubles.)

I suspect that there is a genuine causal effect, but these are the wrong numbers to look at in order to test if it’s there and estimate its size.

5 Mike March 8, 2016 at 8:40 am

Yes! Well theorized and well said.

6 John March 5, 2016 at 6:47 pm

You’re missing the obvious. NBA stats for assists are highly subjective. It is likely that the home team stats recorders are juicing the stats to make the games more noteworthy.

7 Derek March 5, 2016 at 6:57 pm

But they also found positive effects for when assists were reached first…

These results really aren’t surprising at all to anyone who watches much basketball. I believe that draymond is actually a little infamous for chasing triple double stats. Another thing to look at would of course be how this number-chasing differs in close games vs blowouts.

But it’s good to be skeptical about the methodologies on some of these. The hot hand study’s widespread acceptance was totally wrong.

8 WarrenPeese March 6, 2016 at 1:31 am

The question I have is whether Green is getting triple-doubles because he’s on a high-caliber team, or whether his triple-doubles are contributing to this team’s elite status. Either way, it’s commendable for him because it takes a certain skill set to make it happen.

9 Foobarista March 6, 2016 at 11:45 am

You don’t need fancy stats to know that Draymond occasionally chases triple-doubles. He even said so:

10 Smash March 6, 2016 at 11:47 am

Draymond is often making the assist from 4-on-3 situations, after receiving a pass from Curry who’s being trapped. It certainly helps that he’s got elite shooters around him, but not everyone can make the right pass in those situations. He’s definitely contributing to their elite status, but perhaps more on the defensive end.

11 Mike March 6, 2016 at 1:25 pm

I asked my stats professor at Wharton why we don’t use more sports examples. He said the girls didn’t like that. This was circa ’93. Take it for its worth.

Comments on this entry are closed.

Previous post:

Next post: