clock menu more-arrow no yes

Filed under:

James Harden’s impossible box scores, explained

New, comments

How can someone go 1-of-17 from three-point range and still put up more points than shots? This is how.

NBA: Memphis Grizzlies at Houston Rockets Troy Taormina-USA TODAY Sports

Last Sunday, James Harden had one of the strangest box scores in memory. The details that jump off the page (and ended up in chyrons and app alerts): the reigning MVP scored 36 points on 32 field-goal attempts while going 1-of-17 on three-pointers.

The 36 points on 32 shots is good. Typically, you want your scorers to have more points than shots. That indicates better than one point per shot, or the equivalent of shooting 50 percent off the floor. Sort of. We’ll get to why that’s misleading in a minute.

The 1-17 from three line is an abomination. Before Sunday’s game, no NBA player had ever taken 15 or more three-pointers and made just zero or one. Due to injuries, the Rockets don’t have many offensive options other than Harden shooting. Harden is very in sync with the Moreyball standard that threes and free throws are the best shots. So he kept driving (we’ll get to that) and taking threes. The threes kept missing.

So how does a guy who shoots 1-17 on three-pointers end up with what looks like an efficient box score? It’s the free throws. Harden went 15-16 from the line.

Here’s the thing that makes that whole “36 points on 32 shots” efficient-looking line misleading. The free throws are factored into the 36 points (rightfully so). The free throws are not factored into the 32 shots (wrongly so).

Free throws aren’t free. Other than and-1s and technicals, free throws actually do use up possessions just like field goal attempts. But due to ease of quick assessment or whatever, they are rarely counted in surface-level “36 points on 32 shots” type analysis.

Yes, Harden scored 36 points with 32 field-goal attempts, but those 16 free throws also used possessions.

How many? Unfortunately, the box score won’t tell us. (This is an improvement the otherwise highly advanced modern boxscores could add.) We can estimate that the number of shooting possessions from free throws is roughly half since there are two free throws per normal trip, but this neglects and-1s, three-pointer fouls, and free throws. Metric-minded analysts use a 0.44 multiplier as an estimate (this is the figure cooked into True Shooting percentage), but that’s not very elegant in shorthand: no one wants to say that Harden scored 36 points on an estimated 39.04 shooting possessions.

A quick perusal of the play-by-play for Sunday’s game shows that Harden’s 16 free throws came on just six shooting possessions (three three-pointer fouls, three standard two-shot fouls, and one and-1 that doesn’t count as a shooting possession itself because the field goal attempt is already counted). Sixteen free throws in six shooting possessions is pretty wild itself: it completely subverts the standard 0.44 multiplier. (That means that True Shooting percentage might actually underrate Harden’s scoring efficiency, but that’s another spelunking for another day).

Those six scoring possessions that ended with free throws can then be added to the 32 field goal attempts for our total number of shooting possessions used by Harden to get those 36 points: 38.

So Harden scored 36 points in 38 shooting possessions. That doesn’t seem so efficient after all, but it’s more accurate and reflective of Harden’s total scoring efficiency against the Magic than the 36 on 32 line.

Now we just need to find an elegant way to incorporate the points created by Harden’s 12 assists with just four turnovers into that quick-reference line ...