Mar 4, 2015; New Orleans, LA, USA; New Orleans Pelicans forward Anthony Davis (23) reacts after a basket by guard Tyreke Evans during the fourth quarter of a game against the Detroit Pistons at the Smoothie King Center. The Pelicans defeated the Pistons 88-85. Mandatory Credit: Derick E. Hingle-USA TODAY Sports
So words to start this post? Or a few graphs? Let’s go with the graphs:
There’s this perception that the NBA is a point guard dominated league. If that’s the case, what gives? In the graphs above, we can see that power forward and center have a higher mean Real Plus-Minus (RPM), with power forward having a higher range of RPM values than point guard as well. Meanwhile the mean and range RPM for point guards is fairly similar to the small forward position. So maybe it’s not a point guard dominated league?
When talking with Steve Ilardi over twitter, one of the co-creators of RPM, he brought up some interesting points:
@knarsu3 I wonder what those distributions (and means) would look like on a minutes-weighted basis . . . ?
— Steve Ilardi (@dr_ilardi) March 5, 2015
@knarsu3 Yes, all would be shifted higher, but I have a feeling some positions (esp PG) might shift more than others . . .
— Steve Ilardi (@dr_ilardi) March 5, 2015
I decided to apply a pair of filters to the box plot graph to help answer this question.[1. It’s not the best way to do this but it seems to illustrate the point perfectly fine.] I filtered by MPG >= 15 and GP >= 10. This essentially limits us to the guys who have played 150 minutes over the season and we can see a difference:
The point guards made a big jump as Ilardi expected. The mean RPM for point guards is now higher than small forwards and roughly equal to the centers. The range is still roughly equal to that of small forwards, if not very slightly higher. The other group that made a massive jump was the power forwards, who now have the highest mean RPM.
So what does that jump indicate? There’s a lot of very low minute point guards who are dragging down the overall mean RPM of the position. By filtering out some of those guys, we see point guards make a big jump. The same appears to be true for the power forwardss. How about if we now try to filter by starters? I’ll apply a MPG >=30 and GP >= 20 filter here.
Everything makes a big jump upward[2. As expected. Coaches generally play the right guys and so higher minute guys are going to have higher RPMs.] but what’s interesting is that power forwards have the highest mean RPM with the smallest range too. Basically, all of the power forwards who are getting over 30 minutes a game and have played at least 20 games are really good.
A few more interesting tidbits: the range for centers seems to have grown. The range for centers was pretty small in the previous graphs. That’s not really the case in this one though and the range seems to be more similar to the other positions. Also, the range for point guards is pretty large which is somewhat surprising because there are so many good point guards in the league right now[3. Or maybe there are so many good overrated PGs in the league right now?]. Finally, the revenge of the small forwards: the mean RPM for small forwardss is basically back to even with point guards. So there seems to be a lot of 15-30 minute a game small forwards who were dragging down the mean RPM for their position relative to the point guards.
Of course, surely, if we look at ORPM and DRPM, we’re going to come to different conclusions. We’d expect to see point guards with a higher mean ORPM and we’d expect to see the bigs with a higher mean DRPM. Is this true? Let’s first look at ORPM:
As expected, point guards have the highest mean ORPM and a higher peak range as well. But that might be the least interesting aspect of this graphic. Instead, note the small-ish range for the small forwards and the number of outliers[4. Guys like Lebron, Carmelo, Durant.]. Additionally, note that the range and mean for power forwards and shooting guards are fairly similar. Usually, we expect our bigs to have smaller contributions on the offensive end compared to the guards and wings[5. Supported by the data as far back as 2008, when Eli Witus, now with the Rockets, calculated Offensive and Defensive Adjusted Plus-Minus and found the positional averages.]. Instead, we see that power forwards and shooting guards have had fairly similar offensive distributions and means[6. Shooting guards are still slightly higher both in range and mean but not by much.], albeit with shooting guards having a rather large outlier. Gee, I wonder who that could be….[7. It’s James Harden if you really don’t know. But who are we kidding, who doesn’t know that? Morey certainly hopes you do.]. It’s no secret that shooting guard is fairly weak as a position right now, but maybe it is a secret that the power forward position is pretty strong right now—so strong that it’s essentially as impactful on the offensive end as the shooting guard position. Finally, not surprisingly, Centers have the lowest mean ORPM and the lowest range as well.
Let’s look at DRPM now:
As expected, bigs are a lot more impactful on the defensive end while point guards are the least impactful position defensively[8. Every reader right now: Thanks Captain Obvious…Do I really need a graphic to tell me that?]. While there isn’t really anything surprising about this graphic, outside of maybe PG having a wider range of DRPM than shooting, it is interesting to see the defensive impact measured in a descending order by position[9. The previously mentioned Eli Witus study found the exact same descending order by position.]. And if nothing else, at least it’s a cool graph right?