With free agency in full swing, I’ve been thinking a lot about player value. Here I define player value as how well a player performs given their salary. I know this is slightly confusing given how we use "valuable" to describe MVPs, but just bear with me here: value relative to salary, not value as in talent relative to the team or to the league. I look at a snapshot of value for the 2013-2014 season, so just know that both the salaries and player production here are based off a single year.
Also, I’m not an economist or a statistician… just a lifelong Wolves fan.
My main points are more or less ideas that come up frequently at Canis Hoopus, so I’m not reinventing the wheel here:
1. There’s a lot of variation in the relationship between player production and salary, this is why there are teams like the Spurs and teams like us (or the Bucks for those of us Wolves fans in Madison). Identifying solid players early and not paying them very much is a critical part of running a team (even the Heat are finding this out).
The key exceptions to this are superstars. Like we’ve often discussed here at CH, superstars are worth more than max contracts (I’d say that’s a great working definition of a superstar). Identifying who those players are and having the management ability to keep them is the difference between paying Klay Thompson max money and paying Kevin Love max money.
2. Player value relative to their position makes wings more valuable than big men. Point guards have a whole different relationship between production and salary that make low salary players more valuable than high salary players relative to the other four positions.
3. In the short term, we are very unlikely to replace Kevin Love’s value from any trades … obviously.
Bonus Content: Spurs… obviously.
1. The Relationship Between Player Production and Salary
Just some background here: I did a basic regression between a player’s 2013-2014 cap salary and win shares per 48 minutes (WS per 48). Then I subtract the player’s WS per 48 from the WS per 48 of an average player at that player’s salary (on the plot below that would be the distance between the player name and the line). This gives me a rough (with plenty of flaws) estimate of how valuable that player is given his contract. If the value is positive, the player over-performed his salary in 2013-2014. If the value is negative, he underperformed his salary.
I use WS per 48, but you can imagine the difficulty in trying to quantify a player’s productivity. There are a handful of really great metrics, this is just one that I’ve been more or less comfortable with for a while. Just fair warning: a huge amount of the variation in this relationship could be explained by the nuances of WS per 48 (which can be unfriendly to good players and friendly to bad players - I'm not well versed on which kinds of players fall into each category) and not actual player production/ability. You could sub in your own favorite comprehensive statistic.
I limited the player pool to those who played at least 500 minutes and obviously those who have readily available salary and WS per 48 info (basketball-reference.com). And just to reiterate: this is only a single year of information.
I plotted names rather than points because the outliers (e.g. the only names you may actually be able to make out) tell a pretty interesting story. You see players who dramatically overperform given their contract (Kawhi Leonard and Kevin Durant) and players that dramatically underperform given their contract (Anthony Bennett, Evan Turner, and Joe Johnson). You even see the cluster of superstars who are off the charts in WS per 48 and are being paid max money.
2. Player Value Varies by Position
From the current lack of production at the wing position in the NBA, we know that some positions are more valuable than others and GMs will pay accordingly. Here, a player's position does seem to play a role in determining the relationship between production and salary, however the relationship tends to break down a bit within certain positions (some of these are pretty ugly, e.g. centers and wings).
So I figure this can give us two ways to look at player value:
1. Player’s value relative to all other players (Value Overall)
2. Player’s value relative to their position (Value Position)
Generally, this gives SGs and SFs more value than PFs and Cs. The line that represents the average performance given a salary is significantly lower for wings and higher for bigs (meaning it’s easier for wings to overperform). Our very own Corey Brewer is an excellent example of how the difference between wings and bigs affects a player’s value:
Player |
Team |
Value Overall |
Value Position |
Corey Brewer |
MIN |
0.000 |
0.011 |
Based on his performance last year, Corey’s about average given his contract compared to all other players in the league. Relative to other SFs, Corey is actually slightly (thought not by much) above average, with +0.011 WS per 48 over the average SF making his salary.
The increase in performance given an increase in salary (the slope of the lines) isn’t significantly different among these four positions. However, the slope is different for point guards (compared to the other four positions), which changes how we interpret value in PGs.
What happens with PGs is pretty interesting in that small contracts become more valuable and large contracts become less valuable relative to that same positive relationship in the other positions. This has the largest effect on guys like Chris Paul and Patty Mills. This is also a major plug for getting solid value from PGs on good contracts.
Here’s the top 25 player values at their position:
Player |
Team |
Value Overall |
Value Position |
Kevin Durant |
OKC |
0.143 |
0.170 |
Patrick Mills |
SAS |
0.090 |
0.114 |
Kawhi Leonard |
SAS |
0.105 |
0.114 |
Brandan Wright |
DAL |
0.131 |
0.105 |
MIA |
0.107 |
0.102 |
|
Stephen Curry |
GSW |
0.106 |
0.100 |
Kevin Love |
MIN |
0.106 |
0.099 |
HOU |
0.086 |
0.097 |
|
Isaiah Thomas |
SAC |
0.068 |
0.096 |
Anthony Davis |
NOP |
0.114 |
0.089 |
MIA |
0.125 |
0.089 |
|
Goran Dragic |
PHO |
0.080 |
0.088 |
Damian Lillard |
POR |
0.068 |
0.087 |
Mason Plumlee |
BRK |
0.095 |
0.082 |
TOR |
0.084 |
0.082 |
|
Manu Ginobili |
SAS |
0.072 |
0.080 |
MEM |
0.070 |
0.077 |
|
HOU |
0.087 |
0.074 |
|
Darren Collison |
LAC |
0.051 |
0.074 |
Brook Lopez |
BRK |
0.072 |
0.074 |
Chris Paul |
LAC |
0.113 |
0.073 |
Pablo Prigioni |
NYK |
0.044 |
0.070 |
Cory Joseph |
SAS |
0.064 |
0.069 |
C.J. Watson |
IND |
0.046 |
0.069 |
Paul George |
IND |
0.047 |
0.068 |
And the top 25 least valuable players at their position:
Player |
Team |
Value Overall |
Value Position |
MIL |
-0.121 |
-0.129 |
|
Anthony Bennett |
CLE |
-0.126 |
-0.115 |
UTA |
-0.140 |
-0.115 |
|
Dennis Schroeder |
ATL |
-0.137 |
-0.111 |
Tony Wroten |
PHI |
-0.106 |
-0.101 |
MIL |
-0.106 |
-0.098 |
|
Kendrick Perkins |
OKC |
-0.084 |
-0.097 |
Jeffery Taylor |
CHA |
-0.099 |
-0.092 |
Josh Smith |
DET |
-0.110 |
-0.090 |
Rajon Rondo |
BOS |
-0.074 |
-0.089 |
Diante Garrett |
UTA |
-0.116 |
-0.087 |
Chris Kaman |
LAL |
-0.055 |
-0.084 |
Brandon Davies |
PHI |
-0.071 |
-0.084 |
Kevin Garnett |
BRK |
-0.070 |
-0.078 |
CHI |
-0.068 |
-0.074 |
|
MIL |
-0.063 |
-0.073 |
|
Andrea Bargnani |
NYK |
-0.064 |
-0.073 |
Jason Terry |
BRK |
-0.083 |
-0.073 |
Enes Kanter |
UTA |
-0.049 |
-0.073 |
DEN |
-0.078 |
-0.071 |
|
Deron Williams |
BRK |
-0.030 |
-0.070 |
Andrew Nicholson |
ORL |
-0.057 |
-0.069 |
WAS |
-0.056 |
-0.068 |
|
LAL |
-0.081 |
-0.067 |
|
Eric Gordon |
NOP |
-0.078 |
-0.067 |
In the 25 most valuable players, you see a good mix of superstars over-performing on their max deals and young guys who are still on their rookie contracts. In the bottom you see all the strike-outs on guys that the league thought were worth max or near-max money as well as some overpaid role players (like J.J.). You also see a ton of guys who were playing through injuries/bad fits in systems.
The Wolves look pretty good at their core (though this will change if we give Ricky max money and he doesn’t improve), but as you might expect, our bench is loaded with some pretty nasty deals (sorry Alexey…). Also, here’s a numerical basis (not just how fun he is to watch) for never letting Ronny Turiaf go.
Wolves:
Player |
Team |
Value Overall |
Value Position |
Kevin Love |
MIN |
0.106 |
0.099 |
MIN |
0.044 |
0.037 |
|
Ronny Turiaf |
MIN |
0.040 |
0.028 |
Ricky Rubio |
MIN |
0.011 |
0.024 |
MIN |
0.014 |
0.021 |
|
MIN |
0.047 |
0.012 |
|
MIN |
0.005 |
0.011 |
|
Corey Brewer |
MIN |
0.000 |
0.011 |
MIN |
-0.008 |
0.002 |
|
MIN |
-0.037 |
-0.026 |
|
Jose Barea |
MIN |
-0.070 |
-0.055 |
MIN |
-0.072 |
-0.066 |
3. Kevin Love is Way More Valuable than a Max Contract.
The odds of us matching Kevin Love’s value are very slim. The story is pretty grim if we look at value as measured by how dramatically he outperforms his current contract (+0.105 WS per 48 over an average player at his salary) and next contract given this year’s production (+0.094 at the average player making 19M-ish). The moral of the story is that Kevin Love is one of those players who will be underpaid on a max deal and we’ve mismanaged our way out of having him on our team.
The major "centerpieces" of a Kevin Love trade have been Klay Thompson and Chandler Parsons. The general consensus seems to be that both players will get max or near-max deals next year. So for a player making 15 million per year, you would expect a WS per 48 of 0.136. Here are Parson’s and Thompson’s WS per 48 over the past few seasons:
Rookie |
Yr 2 |
Yr 3 |
Pos. Avg. Value at Max Deal |
|
Chandler Parsons |
0.9 |
0.121 |
0.131 |
0.114(SF) |
Klay Thompson |
0.05 |
0.07 |
0.112 |
0.125(SG) |
You can see that being an average producer for a 15 million dollar contract is well within reach for Chandler Parsons, certainly if you consider average production for a max contract SF. If you squint your eyes, hope, and extrapolate, Klay Thompson could even be on his way to being an average producer at that salary. Klay Thompson is younger and the pundits around the league seem to think he has a lot of potential, so maybe 0.136 (from 0.112) isn’t so out of reach. Certainly average production for a max contract SG (0.125) is attainable***.
***However, "average" here is just that. It involves taking all the Eric Gordons and all the James Hardens along with the players who actually perform at an average rate. So if an average includes some really really nasty contracts belonging to players that vastly underperform, perhaps we shouldn’t be shooting for average. A deeper analysis could give us production "goals" for an average, good, or great player at a given salary (but I bet we'd run into sample size issues).
One interesting name that has come up recently is Terrence Jones. Jones is largely beloved by every advanced stat not named RPM and has the youth and versatility to be something really special. By the value metrics here, Jones is the second most valuable player on the rockets (to James Harden) relative to his position (PF, +0.074) and the most valuable player on the Rockets relative to all players (+0.087). If he can maintain his performance from 2013-2014 (0.169; 2012-2013 – 0.128), he could easily become worth a max contract if he migrates to the Wolves. Even if he regresses to his rookie year stats, he’s still a borderline (-0.003 PF, -0.008 Overall) average value at a max contract, which likely wouldn't happen if he regressed.
Success for us post-Love will largely depend on how many players like Jones (and not Parsons or Thompson) we can accumulate and how discerning we are in whether they deserve big contracts after they are no longer cheap.
To sum up this section; I could see one best case scenario and two realistic best case scenarios arising from the potential outcomes of dealing with Love:
1. We keep him and his value (best case scenario by far, we all know this).
2. We trade him to Houston (or Chicago I could get excited about Jimmy Butler and Taj Gibson). The best case scenario is that Parsons continues to steadily improve and Terrence Jones is who the numbers say he is.
3. We let him walk.
I don’t think option 3 is the worst possible outcome. I do think the worst possible outcomes involve taking on contracts like what Klay (or David Lee who is even +0.017 on the average PF making 15M) will receive to, more than likely, play average basketball.
Bonus Content: Player Value by Team… AKA: Spurs.
Looking at player value by teams is also pretty interesting (bold are significantly different from average, value = 0):
Team |
Mean Value |
SAS |
0.0491 |
TOR |
0.0423 |
LAC |
0.0422 |
GSW |
0.0339 |
IND |
0.0269 |
HOU |
0.0258 |
OKC |
0.0252 |
DAL |
0.0245 |
PHO |
0.0186 |
POR |
0.0180 |
MEM |
0.0133 |
MIA |
0.0128 |
MIN |
0.0067 |
CHI |
0.0055 |
CHA |
0.0050 |
ATL |
-0.0054 |
NOP |
-0.0106 |
WAS |
-0.0111 |
CLE |
-0.0113 |
DEN |
-0.0115 |
BRK |
-0.0116 |
NYK |
-0.0124 |
ORL |
-0.0139 |
SAC |
-0.0150 |
DET |
-0.0228 |
LAL |
-0.0292 |
BOS |
-0.0304 |
MIL |
-0.0532 |
UTA |
-0.0550 |
PHI |
-0.0630 |
|
|
Player |
Team |
Value Overall |
Value Position |
Patrick Mills |
SAS |
0.090 |
0.114 |
Kawhi Leonard |
SAS |
0.105 |
0.114 |
Manu Ginobili |
SAS |
0.072 |
0.080 |
Cory Joseph |
SAS |
0.064 |
0.069 |
Marco Belinelli |
SAS |
0.053 |
0.061 |
Danny Green |
SAS |
0.030 |
0.036 |
Tiago Splitter |
SAS |
0.050 |
0.035 |
Tim Duncan |
SAS |
0.048 |
0.035 |
Jeff Ayres |
SAS |
0.038 |
0.026 |
SAS |
0.017 |
0.007 |
|
Tony Parker |
SAS |
0.015 |
0.002 |
Boris Diaw |
SAS |
0.008 |
-0.001 |