Neil Paine over at Basketball-Reference released a study on Wednesday looking at which offensive stats stay most consistent when a player changes roles. It's really a fascinating read considering how much debate is generated when a player changes teams, positions or, you got it, roles.
Paine defines role changes through year-over-year differences in usage percentage (USG%) which is the percentage of possessions a player uses when he's on the floor. This makes sense, of course, because the amount of possessions you take up is indicative of your role within the team's offense. Using the 1,036 players whose USG% changed most YOY since 1974, he looks at the following stats:
- True Shooting % (Scoring efficiency that takes into account 2FG%, 3FG% and FT%)
- Assist Rate (% of teammate FG the player assisted)
- Turnover Rate (Turnovers per possessions used)
- Free throw Rate (FTA/FGA)
- Offensive rebounding % (% of available OReb grabbed while on court)
For our purposes, I'll take his results and (attempt to) apply the correlations to Raymond Felton as we're all expecting him to change his role under new scheme and coach Mike D'Antoni.
High correlations indicate that the stat is more likely to be representative of a player's skill independent of his role in the offense, while lower correlations mean the stat is more a product of the player's situation.
The results:
Stat Correlation TS% 0.627 AsR 0.905 ToR 0.724 FTr 0.802 OR% 0.932
Paine finds that true shooting percentage is the least consistent stat during a role change because an increase in USG% means an increase in shots, which increases the likelihood of missed shots. Similarly, turnover rate is the second-least consistent because the more possessions a player takes up, the higher the chances are of turning the ball over. Assist rate and offensive rebounding rate are the most consistent because, as he states, "those two stats measure tendency as much as ability." For instance, we all know LeBron James will continue to pass the ball in Miami the same way we know Al Harrington will continue to not pass the ball in Denver. Meanwhile, free throw rate is in the middle because an increase in shots tends to translate to jumpshots as opposed to shots at the basket where a player is more likely to be fouled.
Is all that clear? Good, cause this is where I actually have to do some work.
But first, applying these trends to Felton without context would be irresponsible. We need reference - basketball reference, if you will. Luckily another point guard on a two-year contract named Chris Duhon changed roles in his first year under Mike D'Antoni, so let's see if the trends ring true using per 40 minute totals.
| YEAR | USG% | TS% | AR | TOR | FTR | ORR |
| 2008 | 13.4 | 50.8 | 36.8 | 10.0 | 0.2 | 1.5 |
| 2009 | 14.8 | 56.9 | 36.3 | 14.1 | 0.3 | 1.6 |
| CHANGE | +1.4% | +6.1% | -0.5% | +4.1% | +0.1% | +0.1% |
As expected, Duhon's usage increased under D'Antoni as he controlled the ball more than he ever could in Chicago. What jumps off the screen is the +6.1% increase in TS%. Duhon goes from being ~3% under the league average to ~3% over it with increased shot attempts.
How'd he do it? Well, he basically went on a tear in '09 by making 21% more threes while attempting -17% fewer long-twos (16-23 feet shots) per 40 mins. Increasing his efficiency on a more valuable shot while decreasing the number of inefficient ones he took did the trick. It was definitely an anomaly, though, considering his career TS% is 52.4% (4.5% lower than his first year under D'Antoni) and that he regressed to 50.1% last season. Duhon's turnover rate also increased significantly with his usage bump, which you would also expect. It was easily the highest of his career.
The spikes seem to support Paine's study though. These two stats are wildly affected by the role Duhon had to assume. He needed to shoot more threes and facilitate the offense far more than he needed to in Chicago.
The -0.5% dip in his assist rate is a bit of a surprise because of his increased totals in '09, but it's fairly negligible. In '08, 60% of Chicago's shots were assisted. In '09, only 55% were assisted for the Knicks. Perhaps there were just more to go around for him in Chicago.
But onto my point. What can we expect from Raymond Felton? His efficiency increased across the board last season. It'll be even higher, just like Duhon's, once he's under D'Antoni. He should certainly be ranked higher than the 23rd best point guard in the league, right? Not so fast.
| YEAR | USG% | TS% | AR | TOR | FTR | ORR |
| 2009 | 22.0 | 48.3 | 27.9 | 11.5 | 0.25 | 2.2 |
| 2010 | 19.2 | 52.5 | 29.0 | 10.9 | 0.20 | 2.5 |
| CHANGE | -2.8% | +4.2% | +1.1% | -0.6% | -0.05% | +0.3% |
Felton's -2.8% decrease in usage last season directly affected his increased scoring efficiency. He attempted -11% less threes and -26% less long-twos per 40. Simply attempting fewer difficult shots helped increase his TS% by 4.2%. But get this, it was still below average! And you may scream small sample size but his previous TS%'s read like this: 50%, 48.1%, 48.3%. Pretty consistent, I'd say. And all with higher usage rates than he had last season.
Paine's findings are at play here again. The significant change in usage results in much different efficiency. This time, a decrease in usage was the cause. This is discouraging cause we're assuming that both his usage and efficiency will go up with his new role on the Knicks. Problem is, he'll naturally attempt more threes and long-twos in this system which is what lowered it in the first place.
So, there's reason to question Felton's improvement. There's reason why he only received a two-year contract. But there's also plenty of reason to be optimistic if you're more half-full than I. D'Antoni still tends to get the most out of his players. Felton's stats suggest he should pass more and maybe he will with someone the caliber of Amar'e Stoudemire on the receiving end. And the study itself says that scoring efficiency is the most inconsistent stat during a role change. Only time will tell if Felton's offensive production, like Duhon's, was an aberration.