FanPost

What is the Value of each Phoenix Sun?

Christian Petersen

Screenshot2014-04-28at111723AM.0.png Statistics are incredibly important in the NBA. They are important in talent evaluation, contract negotiations, but maybe most important to the casual fan in supporting or denying a particular argument. Attempts have been made to develop more sophisticated units of measurement, but they are inevitably discounted if they don't fit the particular narrative of the respondent.

WP48, PER, RAPM, WARP, EWA - are we talking basketball, or the stock exchange? All of these metrics have been developed to tell a more complete story of what goes on within the confines of the court. But have you ever noticed that if you want salary data, it's almost always separate from the other statistics associated with that player?

Value is a concept that is largely ignored in discussions about any given player. I have spent the summer, with the help of St. More, trying to come up with a repeatable, comprehensive database of every stat that could have an impact on player evaluation, and I thought I would start the discussion with a look at the value of the Suns players. ( We hope to make the entire database available to all before the end of the summer.)

When talking about value, it's pretty much a given that Lebron James and Kevin Durant are underpaid, even at the max. This is by design, and is something that is sanctioned by the CBA, which was agreed to by both the owners and the players union. It is a kind of protection - for the owners, to keep them from their own tendencies to do something stupid, and from the players point of view, to more equitably distribute the available revenues across the maximum number of players. But beyond the elite players in the league, it becomes much harder to justify a max contract for players who might produce less.

Rather than spark an endless discussion about which advanced stat is "right", I have incorporated Wins Produced, Estimated Wins Added, and Wins Above Replacement Player.

Wins Produced is a comprehensive stat developed by the good folks at Wages of Wins. It uses an offensive and defensive efficiency formula based on common box score stats which is adjusted for position and also incorporates team stats. It has a very strong correlation (about 94%) to what actually happened, therefore it's the one I personally prefer. Some criticisms of this stat claim that it overvalues defense. As with any stat, they will point to outliers to make their case. Let me say that due to the comprehensive nature of this stat, sample size has a huge net effect on the results. WP is really only truly accurate after the season is complete, and it's most accurate for the players who have logged 2000 minutes plus throughout the season. You can read about how it's calculated here - spend some time reading it before commenting on its value.

The next metric I will use is EWA - Estimated Wins Added. This is based on John Hollinger's PER stat. This is a much simpler stat, and tends to favor scoring. Critics say it doesn't evaluate defensive impact enough, and it can reward inefficiency. I can't actually find an article explaining how it is calculated, so I hope someone will post it in the comments section.

The third metric I will use in this discussion is WARP - Wins Above Replacement Player. This is a stat derived from the "sabermetrics" pioneered in baseball, and seeks to evaluate players in the context of a team made up of them and four completely average players. You can read how it's calculated here.

All of these metrics are based heavily on comparisons with an average player. The theory is, that if you put a completely average team on the floor, it will win an average number of games - (41 games). This is already a very flawed concept - otherwise, it should be pretty easy to put a team on the floor that could win half its games - yet many don't. 14 teams didn't win 41 games last year. Given that all teams operate under the same basic salary constraints, caps, minimums, and talent pool, there must be some factors that contribute to this imbalance. Sure, there are injuries, there is strategic positioning (otherwise known as tanking), and there are breaks - but the fact is, it's hard to post a winning record, particularly in the Western Conference.

Let me also say this about stats - we, the fans, are operating in the Dark Ages. We are worshipping the Sun, while actual NBA Front Offices are evaluating the latest astrophysical data from the SportsVU cameras - something that we have extremely limited access to. This pisses me off. I was reading an article about Kyle Korver on Grantland, and I came across this:

The gurus at Stats LLC, the company behind the SportVU cameras, have developed two previously unreleased metrics designed to measure the amount of attention an offensive player gets from defenders when he doesn't have the ball.

The first, dubbed "gravity score," measures how often defenders are really guarding a particular player away from the ball. Korver had the fourth-highest score, behind only Kevin Durant, Carmelo Anthony, and Paul George. The second - "distraction score" - is a related attempt to measure how often a player's defender strays away from him to patrol the on-ball action. Korver had the lowest such score in the league.1

Yes, that's right. They are inventing new stats, and aren't telling us about them. I can only guess what they are, but this dynamic, positional, situational method of evaluating a player's actions, and the effects of these actions, is so much more valuable than what we fans have to work with, and I say it's not fair.

GRAVITY SCORE??? DISTRACTION SCORE?!?!?!?!

Oh, well.

On to value.

The three sets of stats have this advantage - they each result in a single number that represents, to the best of their ability, "wins". Rather than comment on the preferability of one stat over another, we can simply throw all three out there, and leave it open for discussion.

GORAN DRAGIC

The Dragon has to be close to the heart of every Suns fan. In spite of the ill-advised trade that sent him and a first round pick to Houston, he returned, and won our hearts. He signed a four year deal at $7.5 million a year, and will likely opt out in that 4th year (player option). But for the last year, and the coming year, he's ours, and he's a very good value.

Using the Wins Produced number, Goran produced 12.4 wins., with a WP48 of .224 (which ranked 21st in the league) At his salary, that equates to $604,838 per win, an incredibly nice number.

Using Estimated Wins Added, he produced 13.8 wins, off of a PER of 21.4 (which ranked 20th in the league). This translates to an even better per win number of $543,298.

Calculating Wins Above Replacement Player, Goran produced 9.75 wins, which translates to $769,230 per win.

Looking at other players across the league, Goran's value in all three metrics is quite consistent, and it's apparent that he is a top 20 player, in spite of ranking 40th in RPM. He definitely earned his all-NBA selection, he deserved an all-star berth, and he's just an all-round great guy to have on the team. Even if he opts out, and gets more than twice what he makes now, he'll still be in the general range of a per win cost with the top players in the league.

PJ TUCKER

PJ's strengths lie on the defensive end of the floor. As such, he is valued much less by the EWA and WARP stats. Even so, he was one of the best values in the league.

WP- PJ, at his incredibly low salary of $884,293, cost $85,028 per win. He would be a bargain at 10 times that price. Fortunately, next year, his salary will be increasing by a factor of approximately 5, which means that his 10.4 wins will still be a bargain at $548,076 per win, should he produce exactly what he did last year. Given that the year before, he produced a team high 7.38 wins, there is some hope that he can increase production.

EWA - This metric, based on PER, does not value PJ as highly - largely due to the fact that he scores a completely average amount of points at his position. Never mind that he's one of the best rebounding and possession getting SF's in the game, he doesn't score a lot of points. Even so, EWA valued him at 3.5 wins, or $254,938 per win. At his salary of $5.7 mil next year, that will come out to $1,628,571 per win - a completely average number. It's a good thing that PJ is not a completely average player.

WARP - This stat credits PJ with 3.41 wins, very similar to the PER based EWA. You can make your own judgments about which stat is more accurate in regard to Mr. Tucker - I think (personally) that WP has captured the essence of PJ much better than the other two.

GERALD GREEN

Gerald, according to WP, was our third most productive player, producing 5.21 wins. He cost $673,076 per win, according to this calculation, still a very good number. I think most would agree that the bulk of his production came on the offensive side of the floor. WP48 ranked him at a .107, just slightly above average, but there ain't much that's average about a Green highlight reel dunk.

EWA - This stat ranks Gerald even higher, crediting him with 7 wins, at a cost of $503,218 per win. His PER was 16.5, so this makes some sense.

WARP - This is our first trip to the silly side. According to the RAPM (Real Adjusted Plus/Minus) figures, Gerald was worse than your average NBA Joe. Throw just about anyone out there, and you'd get more production than Green gave us. He had a negative (-.44) ORPM, a negative (-1.12) DRPM, and an overall negative (-1.56) RPM. This translates to 1.17 wins throughout the season. Did they not watch the season? Anyway, his cost per win balloons to $2,991,452 per win using this metric. Pish, I say.

MILES PLUMLEE

Miles is, for some reason, controversial. Not sure why. According to WP numbers, he was our 4th most productive player, giving us 5.1 wins at a salary of $1,121.520, equating to $219,905 per win. Given that he's slated to be on his rookie contract for two more years, I would consider that an encouraging sign. He posted a nice WP48 of .124, which is pretty great for what was essentially a rookie season.

EWA - This metric also values Miles highly, crediting him 3.9 wins, at a cost of $286,946 per win.

WARP - Once again, this stat is the outlier, crediting Miles only .71 wins throughout the season, at a cost of $1,579,605. Even that is an OK number, but c'mon. He got us more wins than that in just the matchup against Portland.

MARCUS MORRIS

WP

This engendered some incredulous comments when I brought it up earlier. How could Marcus be more valuable to us than Markieff? Apparently, a lot of people didn't notice that he was actually on the floor this year, and by the way, doing pretty well. He only played 200 minutes less than Kieff, he scored a lot more 3's, and put up a respectable .127 WP48, which equates to 4.8 wins, but he did give us slightly less value than Kieff, translating to a $414,025 per win cost.

EWA

Marcus did well in this calculation, posting a PER of 17, and an EWA of 5.8 wins, at a cost of $341,411.

WARP

Once again, the "REAL" Adjusted Plus/Minus seems to be the outlier. While $845,668 per win is a REALLY good number, it seems to value Marcus half as much as the other stats do.

ERIC BLEDSOE

While y'all try to figure out if Eric REALLY, REALLY, REALLY wants to be a Phoenix Sun, I'll confine my comments to what he did while he was one, and what he will (likely) do this coming year.

WP

At his rookie salary of $2,626,473, his 4.7 wins (with a WP48 of .159) was $558,824 per win. It's difficult to project what that will be next year, considering that he played less than half a season, and we don't know what he'll make next year, but suffice it to say, he will likely make a lot more. Hopefully, he'll play a lot more, too, and will still be a good value. For instance, if he played about what Goran played last year, and made $14 mil, he would still cost about $1,458,333 per win, or about half of what LaMarcus Aldridge cost per win (according to WP). The hope would be that increased playing time would lead to increased production, but I'll leave that argument for the comments section.

EWA

.6.1 wins (PER of 19.62) at a cost of $430,169 per win.

WARP

Crediting Eric with 5.77 wins, and a very nice Defensive Real Plus/Minus of 3.96, this translates to a cost of $455,194 per win.

MARKIEFF MORRIS

WP

This stat doesn't capture all the wonder of our 6th MOY candidate. Likely because of suppressed 3 pt and TS% numbers, it credits the older Morrii with 4.39 wins, at a cost of $475,118 per win (with a .098 WP48).

EWA

With a PER of 18.4, Kieff fares much better here, getting credit for 7.4 wins, at a cost of $283,029 per win.

WARP

While not as appreciative of Kieff as the other two stats, he still did pretty well, getting credit for 3.21 wins (better offensively than defensively) at a cost of $651,663 per win.

Something tells me that there are some SportsVU numbers that tell McDonough and Hornacek that Markieff has value in the way that he scores (as does Marcus). He scored a lot of points off of contested shots, and shot a pretty good average. Contesting those kinds of shots should have cut his average by about 10% - but it didn't. I wish they'd share.

CHANNING FRYE

WP

Ah, yes. Everybody's darling. But not, apparently, at Wages of Wins. Channing is credited with a meager 2.2 wins, which, given his $6.4 million salary, translated to $2,909,090 per win. With a well below average of .045 WP48, it's hard to use this metric to justify Channing's existence.

EWA

This one is even more of a Frye hater, giving him credit for only 2 wins, at a cost of $3,272,949 per win. WOW.

WARP

Whatever you want to say about individual stats, when Channing was on the floor, the Suns did a lot, lot better. They were better offensively, they were better defensively. This leads to credit for 10.81 wins, at a cost of $591,497 per win. This should lead to some fun discussions. ;-)

ISH SMITH

WP

Everybody's phavorite pharaoh. Ish actually produced 1.87 wins, although I distinctly remember him turning more games than that around with his hustle. That worked out to $500,770 per win.

EWA

With a PER of 11.7, Ish only gets credit for .4 wins, at an astronomical cost of $2,715,763 per win. Ish.

WARP

Well, I never found ish's WARP. Since he's the definition of an NBA replacement player, let's just leave it at that, shall we?

THE REST

Shavlik Randolph, Dionte Christmas, Archie Goodwin, Leandro Barbosa, and Alex Len all combined to contribute less than one win, total. While LB and Dionte are gone, it's not unreasonable to expect a few additional wins from Shavlik (with a full year to contribute), Archie, and Alex, not to mention our two newest additions of Warren and Ennis. We hope, anyway.

ISAIAH THOMAS

WP

Replacing Ish with Thomas should give us an additional 5 wins next year. Isaiah was one of the bargains of the league last year, clocking in at $128,158 per win (on a 28 win team). At his new salary of $7.2 mil, he will still be a bargain at last year's production (about $1 mil per win), but I would be surprised if he didn't fare a bit better on his new team.

EWA

11.8 wins, with a PER of 20.5. $74,929 per win. Even at his new salary, he would be at $610,169 per win. That is AWESOME!

WARP

Even this stat likes the little one. The numbers are quite close to WP. Thomas should make us a much better team next year.

ANTHONY TOLLIVER

WP

Last year, Tolliver made just over one mil, and WP credited him with 3.2 wins, at a cost of $321,070. It's worth noting that he only played about 1300 minutes, so with his WP48 of .117, he might be expected to do better this coming year. But, while making $3 mil a year, even that puts him as a $1 mil per win player.

EWA

With a PER of 11, Anthony didn't fare so well here, and he was only credited .3 wins. That would translate to about $10 mil per win this year, so let's hope that isn't the case.

WARP

Tolliver's defensive +/- numbers are quite good, and this translates to a nice figure of $290,232 per win, or around $900,000 under his new salary. And, as much as WARP liked Frye, it likes Tolliver better.

So, there it is. A final note - if you figure out how many wins WP actually credited the Suns with, it comes out to a slightly higher than real 51.55 wins. The discrepancy may have something to do with the late Michael Beasley, who actually cost us 3.6 wins last year, so sending him away actually only cost us $1,296,296 per win. It was a bargain!

If you calculate the number of wins EWA credits the Suns with, (including the 3.8 wins that it inexplicably credits Archie with) it comes out to about 49 wins. WARP only comes up with about 37 wins.

Overall, the players on the Suns who actually played (cough Okafor cough) and contributed, ran about $32,521,145, or $677,523 per win. We were probably the best team per dollar in the league. This coming year, we will be adding Bledsoe, Tolliver, and Thomas's salary, subtracting Okafor and Frye, so we will be spending more per win, but I think we can expect more wins. 5 more from Bledsoe, 5 more from Thomas, a couple from Len and Archie, and BAM! We're a 62 win team!

If only it were that easy. Wins above the mean are more expensive, and our salary structure will be going up in the next couple of years with the Dragic and Morrii, so not only will the wins be more expensive, there will be less money to purchase the extra wins with. However, the salary structure of the league is likely to change radically in the next few years as well - I would encourage you to read this piece about the expected influx of "Basketball Related Income" from the new TV deal. By the time the end of the decade rolls around, the salary cap could be as high as $100 million.

There's a lot to discuss here. Go!