The Impact of Changing teams: The Josh Smith Edition

Yesterday it was announced that Josh Smith had signed with the Los Angeles Clippers, his third team in less than a year. However, for as poorly as he played in Detroit, the Rockets were every bit as happy to have him down the stretch last season.

Smith has been known for his poor shot selection and his time in Detroit was mostly a replay of what happened in Atlanta[1. Although interestingly, despite his own poor shot selection, his teammates’ did see an improvement in their own shot selection when he was on the court]. But what happened when he changed teams? Houston and it’s Moreyball shot selection is known to be among the best in the league and a big part of that reason is due to the presence of James Harden, who improves his teammates’ shot selection by 0.05 points per shot[2. On the scale of aXPPS, this is a big difference. Other players like Dwight and Patrick Beverley also improved their teammates’ shot selection but not to the same degree.].

Smith, like many others who changed teams, saw a difference in his shot selection, as measured by aXPPS, after making the switch in teams. Let’s look at how aXPPS changes for the players who changed teams last year (minimum 50 FGAs on both teams):

The orange line represents no change in shot selection and as we can see, most players saw an improvement in their shot selection after switching teams[3. To be exact, 29 out of the 46 players saw an improvement in their shot selection. Again, with a minimum of 50 FGAs on both the old and new teams.] Smith posted the 2nd highest improvement (+0.111), behind Chris Johnson who barely cleared the 50 FGA threshold on both teams. But surely, the quality of teams has to do with the improvement in shot selection right? Let’s look at this by plotting the difference in aXPPS of the player versus the difference in SRS of the teams:[4. To be exact: aXPPS difference = aXPPS of the player on the New Team – aXPPS of the player on the Old Team while SRS difference is: SRS of the new team – SRS of the old team.].

The R² isn’t that high but we see a pretty clear linear relationship. Furthermore, there were only four players who went to better teams and didn’t see an improvement in their shot selection (upper left quadrant). Plumlee, the dot right above the line in the upper left quadrant, barely saw an increase in team quality and yet, his Usage did go up when going to Milwaukee. And out of those four players, three of them saw an increase in Usage when changing teams, which will negatively affect your shot selection. Nate Robinson, who went to the Clippers, was the only player who didn’t see an increase in Usage after moving to a better team. And of course, Robinson barely cleared the 50 FGA minimum with the Clippers.

We can also see the difference in going from the laughable Knicks to playing with LeBron, where both J.R. Smith and Iman Shumpert were among the biggest risers in their shot selection. J.R.’s increase was the third best while Shumpert’s was the sixth best. Additionally, we see Prigioni, Rivers and Mozgov among the top 10 highest risers and it’s no coincidence that they joined teams with Harden, CP3 and LeBron respectively[5. So you mean going to a team with a superstar is a good thing? Tell me something I didn’t know.].

How about if we look strictly at the teams’ Adjusted Offensive Rating[6. Found on this Basketball-Reference page.] versus the players’ shot selection difference:

The relationship isn’t as strong as the previous graph but there’s still a clear linear relationship. We also continue to see the added benefits of getting out of that Philadelphia offense as three of the top 16 guys who saw the largest improvements in their shot selection left Philadelphia.

Although, those stats about Philadelphia are surprising for another reason: the 76ers offense, despite how bad it was, actually had excellent shot selection, ranking third among all the teams last year:

As we can see with this list, having good shot selection doesn’t necessarily lead to a good offense just like having bad shot selection doesn’t necessarily lead to a bad offense. However, we do see many of the known analytics friendly teams towards the top of the league in their shot selection while teams known to be skeptical of analytics[7. Such as the Lakers, Knicks, Nets and TWolves.] are towards the bottom.

I’ll end with this: Players are a product of their environment. Take a good player who has a high Usage on a bad team and put him on a good team, where he handles less of the load, and you’ll see his shot selection improve and as a result, his efficiency. Having Josh Smith going to what was one of the best offenses in the league last season is probably a good thing for his own performance.