Comparing the Cost & Production of Ed Davis and Tristan Thompson

Earlier this summer the Portland Trail Blazers signed Ed Davis to a 3-year contract worth $20 million. Meanwhile, in Cleveland, Tristan Thompson has demanded a 3-year, $53 million contract from the Cavaliers. Thompson will probably settle for less than $53 million, but his final salary will almost certainly be more than double the $20 million Davis earns in Portland.

So, who would you rather have: Ed Davis for $20 million or Tristan Thompson for $40+ million?

On the surface, the answer may seem obvious - Thompson is a rising potential star while Davis was practically run out of Memphis and played for a minimum contract last season. The NBA is a star-based league so teams, especially contenders, should pay the players who have star potential.

The problem with that logic is that Davis and Thompson are remarkably similar on the court, despite the differences in how they're perceived and paid. Both players excel at scoring around the rim and at rebounding, but cannot hit a jump shot. They both struggle as primary "backstops" on defense, but use their athleticism and mobility to bother opposing players in open space.

Examining their traditional statistics from last season highlights the similarities between the two players. Here are their per game numbers:

Davis was a worse free throw shooter but had a substantial lead in assists and blocks. The two players were nearly identical in all other categories.

Thompson did play about three and a half more minutes per game than Davis, though, so these stats may not be capturing differences in their true on-court production. To normalize for minutes played, consider per 36 minute stats:

Note: Per 36 numbers are limited in utility when comparing players who play short minutes in a reserve role as the level of production is generally not sustainable. Thomas Robinson's 19.6 points,17.8 rebounds, and 7.6 fouls per 36 minutes during the 2013-2014 season being a classic example. But Davis and Thompson both saw significant playing time and were deployed regularly as part of their teams' primary rotations so there should be sufficient common ground to use per-36 numbers for comparison.

In this case, Davis averaged 1.4 more points, one more defensive rebound, three times as many assists, and almost twice as many steals and blocks as Thompson. For per minute production, Davis has a clear slight advantage.

Davis and Thompson are similar not just in the total stats they produce, but also in how they're producing.  Consider their scoring "heat charts," representing where they make the most field goals:

Davis (source):

Thompson (source):

The heat charts are remarkably similar; both players struggle to convert shots further than five feet from the basket. Their field goal percentages are also almost identical for every distance, with each player shooting over 65% at the rim and then plummeting to under 40% if they step away from the basket. Interestingly, Thompson only shot 33% on tip-ins, while Davis shot 55%. Thompson is a slightly better offensive rebounder, according to the numbers, but that advantage seems to be mitigated by an inability to convert tips on missed shots. In short, the shot charts and numbers suggest that they are virtually identical as offensive players.

Next, consider the advanced stats:

For these stats, Davis trounces Thompson, leading in virtually every category. Notably, his overall lead in production is captured by a higher PER, while leads in TRB%, AST%, STL%, and BLK% suggest that Davis is having a more tangible impact on games than Thompson.

The box plus/minus stats are also heavily in Davis' favor. These numbers are especially relevant as they control for the quality of team, unlike traditional plus/minus. In other words, Davis would have a poor traditional plus/minus because the Lakers were not very good, but box plus/minus adjusts for team success and converts to a per 100 possessions scale so Thompson and Davis can be fairly compared despite the different success level of their teams. Remarkably, Thompson has negative scores suggesting that his contribution is actually worse than an average player, while Davis has a net positive 6.0!

Who's negotiating for $53 million, again?

Despite the differences in box plus/minus they do fill similar roles for their teams on defense and seem to be about equally effective. Specifically, Davis lacks strength and relies on quickness to make himself known on defense. He can be overpowered on the low block, but performs adequately when he is out in space and can face up to an offensive player. He also has a nose for seeking out weakside blocks. Here's a compilation of some of these skills in Memphis:

Simlarly, Cleveland has utilized Thompson's athleticism and length to disrupt pick-and-rolls, or switch onto opposing perimeter players. While Davis' primary defensive weakness is defending post-ups, Thompson struggles when asked to be a primary rim protector. When possible, the Cavaliers hide that weakness by relying on Timofey Mozgov to defend the lane.

Statistically, both players have allowed opponents to shoot a relatively high percentage at the rim but Thompson only gives up .68 points per possession on pick-and-rolls whereas Davis gives up .95. Davis does average significantly more blocks and steals per game, but those numbers may be deceiving. Davis was often the Lakers' primary weakside and/or help defender at the rim, whereas the Cavaliers relied on Mozgov for more deep protection, so Davis' block numbers may have been slightly exaggerated. Similarly, many of Davis' steals are the result of hustle plays after loose balls are created in his area via poor passes or ballhandling mistakes, rather than solid defensive playmaking.

In summary, empirical analyses of the statistics and a limited eye test do not reveal any decisive differences between Davis and Thompson. Davis does have the lead with some advanced stats, but an eye test of their style of play suggests that those numbers may be exaggerated by the roles the two play for their teams.

With no clear on-court difference to explain the salary disparity, there must be some intangible explanation for the salary difference between Davis and Thompson. One possible key difference is the perception of the development of the two players. Thompson came into the NBA as a raw but athletic project player. Since his rookie season he has made strides by maintaining elite rebounding skills, while significantly improving his ability to finish around the basket and play defense. Thus, the narrative around Thompson is that he is a hard worker who has improved his game significantly.

Davis, on the other hand, has garnered a reputation as a player who has not significantly improved since being drafted. As recently as one season ago, his offensive skills were essentially the same as when he joined the league and he had not added much-needed strength to his frame. Reports that "coaches have questioned both his IQ and motor" have also dogged Davis and may have led to hesitancy on the part of GMs to sign him.

Thompson and Davis have played under very different circumstances, as well, which may have affected how they are viewed. An often overlooked aspect of success in the NBA is the impact that a good or bad fit between a young player and team can have on his career. Rajon Rondo's fall from grace after leaving Doc Rivers and Boston, or Brandon Jennings' pre-injury ascension last season in Detroit both illustrate that point.

In a way, Thompson hit the circumstances jackpot when he was drafted by Cleveland. Early in his career the Cavs were not a good team and were able to play Thompson big minutes as the only notable players in front of him were long-in-the-tooth Antawn Jamison, perpetually injured Anderson Varejao, and, briefly, Andrew Bynum. When Lebron James returned to Cleveland Thompson's skills at cleaning up rebounds and acting as an offensive bailout when James is double teamed were a perfect fit. As noted above, Thompson's defensive skills also mesh well with Mozgov and James. Lastly, the Cavs are serious championship contenders; most teams cannot afford to pay $10 million+ to a player who just averaged 8/8 for a season, but the Cavaliers are already over the salary cap and Thompson may be the piece that propels them to a championship, theoretically justifying his overpayment.

Davis, on the other hand, was originally drafted by the Toronto Raptors in 2010. He saw his first major opportunity in 2012 when he was asked to assume the starting position vacated by an injured Andrea Bargnani. Despite playing the best basketball of his career as a starter, he was quickly traded to Memphis for Rudy Gay. The trade left Davis temporarily stunned, and he struggled to find consistent playing time for the Grizzlies behind veterans Marc Gasol and Zach Randolph. After a tenure in Memphis best described as "disappointing," Davis signed with the Lakers with the hope of earning more playing time. Davis did perform well in Los Angeles and reportedly wanted to return, but the Lakers preferred to keep their salary cap space intact rather than re-signing him early in the free agency period. Suffice it to say, Davis has not found a solid niche since being traded from Toronto.

But that may change for Davis now that he is in Portland; His strengths will potentially pair well with Meyers Leonard's skills, and he may also be able to fill some of the holes left by Robin Lopez. The hope for both parties is that Davis' inside scoring, rebounding, and length/energy/hustle on defense will help him find a home with the Blazers. Meanwhile, Thompson will continue to fill holes around James, Mozgov, and Kyrie Irving with his inside scoring, redounding and length/energy/hustle on defense. In many ways, despite the far different aspirations of their teams, Thompson and Davis will most likely continue to be carbon copies of each other for the 2015-16 season. Except one copy will cost $50 million, and the other will cost $20 million.

h/t to Blazer's Edge reader Ecsa for pointing out this similarity in a fanshot. Grantland's Zach Lowe briefly compared Thompson and Davis in April of this year, as well; that column was also linked earlier in this article.