Tuesday, July 30, 2013

Rudy Gay: glass half empty, glass half full

The Raptors fanbase is torn on Rudy Gay.

On one side, the side that rushes to Rudy's aid, argue that he is a legitimate star, a player with tremendous abilities who the Raptors can build around. The other camp argues to the contrary; they believe that Rudy Gay is wholly overrated, overpaid and curse Bryan Colangelo for acquiring him in a naked and desperate attempt to save his job.

This back and forth has made me wonder; who is Rudy Gay?

This is an important question, one that I'm sure Masai Ujiri is also trying to solve. After all, Gay can potentially leave after this season if he declines his player option and opts for another monstrous deal in free agency.

To answer this question, I will first present Rudy Gay's historical performance in the context of both traditional and advanced boxscore statistics to evaluate WHAT and HOW MUCH Gay produces. I will then use synergy-based statistics to assess HOW he produces, and how his production ranks in the NBA. Knowing the what, how, and how much will allow us to quantify Gay's value. Finally, with this estimate in mind, I'll touch on the financial situation at the conclusion of this post, and approximate what the market will pay Gay if he does opt for free agency at the end of the 2013-2014 season. 

The "WHAT" and "HOW MUCH"

Raptors fans caught a brief glimpse of Gay last season. He was acquired from Memphis in exchange for Calderon's expiring contract and Ed Davis. In his 33 game tenure, he led Toronto to a record of 17-16.

Traditional boxscore statistics suggest that Rudy Gay is a star. His career per game averages of 18.0 PPG, 5.8 RPG, 2.0 APG, 1.4 SPG and 0.9 BPG are certainly impressive. However, advanced boxscore statistics disagree on Gay's productivity.

Win Scores, a statistic that boils down the boxscore into one number based on linear weights (on how much a point of each stat is worth), has Rudy Gay's career average Win Shares per 48 minutes (WS/48) at an unimpressive 0.80 (league average player: 0.100 WS/48). Similarly, Wins Produced (a model similar to Win Shares) has his career average Wins Produced per 48 minutes (WP/48) at 0.75. Even John Hollinger's Player Efficiency Rating (PER), for all of it's flaws with measuring defense and rewarding inefficient scoring, ranks Gay at an unimpressive 16.1 (league average 14). 

This discrepancy between traditional and advanced metrics is likely the root cause of the great divide in opinion regarding Rudy Gay. Fans in favor of traditional stats likely heavily value Gay's production given that only two payers averaged 18 PPG/6 RPG/1.5 SPG last season (the other being Lebron James). The sabermetrically inclined likely believe his absolute production is a mere product of minute allocation, and his production can easily be replaced at a much lower price.

We can evaluate the sabermetric claim by framing Gay's numbers in per 48 minute stats (courtesy of thenbageek.com)

Gay's per minute production is comparable to the production of the average small forward. The only significant difference is points per 48 minutes. However, the difference can easily be explained by greater usage (Gay's career usage rate: ~25% compared to 20% for average SF). His shooting statistics are merely average, as seen below (per thenbageek):

In conclusion, Gay's absolute production is impressive, but a closer look at his rate statistics shows that his rate of production is similar to that of an average small forward. Therefore, the HOW MUCH is that Gay produces a lot of absolute numbers, but the WHAT is that he rebounds and scores a lot, but his per minute production is similar to that of an average small forward. This evaluation is corroborated by advanced boxscore statistics like Win Shares, Wins Produced and PER which uniformly suggest that Gay's production has historically been approximately league average.


The "HOW"
At this point, I must admit that I am a heavy proponent of advanced statistics. However, I will also readily admit that unlike sabermetric stats in baseball, basketball stats are not soundproof and can be fooled. Statistics, no matter how simple or complex in construction, should never be substituted for analysis, but rather be a tool to help clarify and quantify the swaths of data that the human brain cannot fully account for. Therefore, when traditional and advanced stats disagree so strongly, I always look for more clues to help solve the problem.

I turned to synergy sports statistics to gain a better idea of HOW Rudy Gay produces. Boxscores are excellent at telling you WHAT and HOW MUCH a player produced, but don't quite account for the process. To some fans, the process is everything. Synergy stats are unique tools because they provide context. They offer a breakdown of how a player utilizes his offensive possessions, and how efficient he is at certain plays. Unfortunately, I cannot link to their numbers here, but they can be found if you search for Rudy Gay at mysynergysports.com. The data is below, with a full data set here.

Rudy Gay's offensive production in Toronto was something of a mixed bag. There was a lot of isolation (19.9%), along with some ball handling (15.5%) in pick and rolls, post ups (14.1%), spot ups (14.3%) and transitions (14.1%). As you can see, by league ranking, he didn't quite excel at anything other than isolation, where he produced 0.94 points per play (PPP), ranking him 29th in the league at such plays. He was pretty average at everything else, consistently ranking in the hundreds for his most frequent plays.

What this breakdown tells us is that Rudy Gay is a very versatile offensive player. He can score in a number of ways, including via the pick and roll, spotting up, posting up and isolation. Unfortunately, the synergy data also suggest that he did not excel offensively at any play outside of isolation.

I then turned to his shot chart, to see where his shots came from. Did he/the offensive scheme set him up to fail? The chart is below (courtesy of NBA stats):

Gay was significantly more effective from the right side of the floor. He shot efficiently from around the free throw line, right of the free throw line, and with three point baskets from the center of the court. Unfortunately, Gay wasn't very selective, as his attempts from the left side almost perfectly mirrors his attempts from the right side, where he was much more effective. 

One area of concern is his shooting in near the basket. He shoots excellently at the rim (~65% in the last four seasons), but shoots a paltry ~42% from 3-9 feet (numbers per Hoopdata). Given that Gay takes more shots from inside the paint than from anywhere else, his efficiency would improve significantly if he shot closer to the basket (how this can be achieved is up to Casey and Gay).

Perhaps better game-planning from Coach Dwayne Casey can improve Gay's offensive efficiency. They should design the game-plan to get Gay to shoot from his most efficient areas. However, Rudy Gay has traditionally been unselective, as shown by this shot chart from his 2011-2012 season in Memphis:

Given that his shot selection is largely the same, and that Memphis and Toronto certainly employ vastly different offensive schemes, it is likely that this lack of selectivity is a result of Gay's discretion (or lack there of) rather than a matter of coaching.

Overall, synergy stats suggest that Gay is certainly offensively versatile, as evidenced by his play usage breakdown. However, he is not exceptional at any play other than isolation. The "HOW" is that Gay produced via a number of different plays, but he simply is not very good at most of them. 

Defensive metrics are certainly flawed, but statistics from synergy stats allow us to explore a player's defensive ability beyond blocks, steals and rebounds. The following is a breakdown on how proficient Gay was at defending specific plays:

Gay was not particularly effective at defending any particular play. He is a league average defender. 

These claims may incite some disagreement and dissonance in the minds of fans who believe that his size (6'9 height, 7'3 wingspan) and per game steals and block averages (career 1.4/0.9) suggest that he is an effective defender. However, his steals and block totals are likely a reflection of minutes played, not ability (as seen with per 48 numbers), and the synergy statistics clearly show that he was merely average at defense across a variety of plays.


Who is Rudy Gay?
The data suggests that Rudy Gay is not an exceptional offensive player and an average defender. The "HOW MUCH" is that he produces a lot of absolute numbers, especially rebounds and points. The "WHAT" is that he does so because of an abundance of minutes played, not because he is exceptionally efficient or effective. This is further corroborated by the "HOW", in which synergy statistics reveal that he is certainly versatile, but is league average at executing and defending most situations.

Basically, Rudy Gay is an average player, but his raw numbers make him look like an above-average player. 


What will the market for Gay's services look like?
Despite what all the data says, Rudy Gay will undoubtedly fetch a large contract in free agency. It has been shown over, and over again that NBA general managers will pay a high price for players who are perceived to be stars, regardless of whether their productivity support that claim or not. 

Rudy Gay has the option of declining his player option and becoming an unrestricted free agent at the end of the 2013-2014 season. If he declines his option, he would forgo the 19.3 million dollars that he is slated to earn in 2014-2015. On the surface, this scenario would be unlikely, however the right offer may very well entice Gay to choose to become a free agent.

Who has money?
As it currently stands, 22 teams will be under the salary cap for the 2014 offseason (I compiled a list of how much cap space each team will have here). A number of these teams lack a quality small forward, including the Suns (~20M), Cavaliers (~22M), Bobcats/Hornets (~23M), 76ers (~37M), Jazz (~43.5M), Lakers (~46M) and Mavericks (~48M). Given that the new CBA stipulates that teams must spend a minimum of ~90% of the salary cap, there will be an abundance of money being handed out to free agents next summer. It's going to be a bonanza.

Given that the 2014 offseason will likely be a buyers market, a hefty offer may very well entice Gay to decline his player option. Remember, if Gay eschews free agency in 2014, he is betting that he can garner a similar contract in the future. However, that's a lot of risk to incur. Should he seriously injure himself, or if his performance declines, he will likely lose a significant amount of money. Gay has seriously injured himself in the past, missing 28 games in 2011 in part due to a dislocated shoulder.

Who will be available?
There's also the matter of who will be available. Should teams miss out on players like Lebron James or Carmelo Anthony (both have early termination clauses), Gay may very well benefit from the "bridesmaid effect" (everyone understands this, right? Think Stoudemire or Boozer in 2010, or just Rudy Gay in 2010). 

A full list of potential free agent small forwards can be found here, and beyond Lebron and Carmelo, the cupboard is pretty barren. Players like Butler, Leonard, Parsons, Barnes, MKG and George will definitely have their team options activated, or receive extensions. Paul Pierce will most likely retire (especially if the Nets crash and burn this season), and Andrei Kirilenko will likely continue to be underpaid as he has been all his career (career .257 WP48). Gay's main competitor will likely be Luol Deng, but given the amount of money scheduled to be on the market, there should be plenty of cash for the both of them. 

What have players of Gay's calibre historically received?
So what kind of contract offer should Rudy Gay expect? It is a proven fact that the NBA market heavily values points per game. In an excellent paper by Dr. Dave Berri et al (yes, the same co-creator of Wins Produced), he found a significant link between points per game and wage garnered, regardless of their shooting efficiency. This really shouldn't come as a surprise to any NBA fan.

Therefore, I looked at the contracts of other small forwards who score a similar number of points per game as Gay to generate a ballpark figure. The findings are below and can be found here:

Given that Gay has averaged 18 PPG for his career, he is probably looking at a ceiling of around what Josh Smith received, and a minimum of what Iguodala signed for (although Gay is significantly younger than Iggy). That puts his ballpark figure at something like 12-14 million dollars per year for four years. If my ballpark figure is correct, Gay will certainly opt out of the final year of his contract.

What should the Raptors do with Gay?
At that price, the Raptors will be wise to let Gay walk. Gay is largely average on both offense and defense, but will be paid like a star. Should Gay leave in free agency, the Raptors will clear 19.3 million off their cap, bringing their cap figure to a palatable 38 million dollars in 2014. 

Another option would be to trade Gay at the trade deadline to a contender as a rental player. However most contenders are currently capped out, meaning the Raptors would have to take back significant salary back to facilitate such a deal, or involve a third team. There's also the issue that most contending team cannot afford to shed salary because it would compromise their depth. 

The most ideal outcome for the Raptors would be if they can sell him to a "starless" team with a hole at small forward and cap space. An example would be the Suns, who will be 21 million dollars under the salary cap in the 2014 offseason. Gay would represent a significant upgrade over Michael Beasley (who is tremendously unproductive), and can offer Phoenix the clout of his "stardom". If Ujiri can manage to acquire a few picks (Phoenix will certainly hold on to their own pick, but they do have Indiana's 2014 1st rounder from the Scola deal) while taking back little in salary (perhaps swallow the remainder of the disgruntled Gortat's deal), it would be an absolute coup. 


  1. Rudy Gay = WP48 0.075
    Average SF = WP48 0.099

    Demar Derozan = WP48 0.054 in 2012-2013
    Average SG = WP48 0.099

    GEEEZ! I'm not very good at understanding stats. Kinda new at this but man can Derozan or Rudy get average WP48 next season or should we seriously look elsewhere? I like Demar, hardworking and everything but damn.

    1. It's definitely possible. I'd caution against only looking at one number. WP48 is a great summation, but a breakdown of their stats show where their weaknesses are. I won't get into specifics, but WP gives no bonus to inefficient scoring (inefficient being shooting below their positional averages) and actually punishes DD and Gay for their points scored (because of the shooting percentages)

      There's hope for Gay. He had a season where he was 0.135. Unfortunately, that was a while ago with a different team. On the other hand, DD is a lost cause. He doesn't do much except shoot, but he's bad at shooting, so...

      Look, if the game plan was to get Gay and Derozan to shoot from where they shoot well, their numbers will improve. However, they've historically been resistant to shooting from their hot zones. It's possible that Gay produces at league average. I'm not so sure about DD.

    2. Thanks once again for reply man. Hopefully Nurse can create an offence that will actually work for those two. LOL

    3. I'm always happy to respond to comments.

      I hope Nurse can design a better offense to maximize DD and Gay's effectiveness. Unfortunately, he has almost no track record, so I can't make any predictions. We'll have to wait until the season starts.