Player Tracking Plus Minus (PT-PM) Top Thirty

Over the last couple of years I have developed my own player metric called Player Tracking Plus Minus (PT-PM), a statistical plus-minus model that uses both SportVU data and traditional box score statistics to estimate players’ contributions on both offense and defense[1. You can read about the development in some previous article here and here]. A key thing to keep in mind is that like any “one number” stat the PT-PM metric gives a general estimate of player value, in this case balancing different contributions or detrimental plays on offense and defense. PT-PM is also half of the blend I use for my preseason team wins projections, which scored well both last year and so far this year compared to other analytic models, Vegas or the ESPN Summer forecast.

There are a couple of different ways to quantify a player’s performance using PT-PM. PT-PM is a rate based metric, calculated here as a per 40 minutes of play, however, for players with limited playing time the estimate will be very noisy and in addition playing time assigned by coaches is a valuable signal of player quality. For those reasons, I include a playing time regressed version of the stat that regresses the player’s score toward a baseline adjusted for minutes player[2. The baseline for players with minimal minutes starts at approximately -3 adjusting to just above average, or 0], and a points over replacement player (PORP), which is simply the player’s total contribution over the hypothetical minimum contract player[3. Calculated as (PT-PM – 2.5) * minutes played].

With the season about half over I figured it was about time to update this year’s scores to date (through February 1). Below are the top thirty players as rated by PT-PM per 40, with the other measures alongside[4. All stats via NBA.com].

• No shocker that Steph Curry is the top player in the league his scoring rating is completely off the charts and he is basically above average at everything other than rebounding.
• Kawhi Leonard comes in second with a balanced offensive and defensive game.
• Manu Ginobili and Denver rookie Nikola Jokic emphasize the importance of playing time as their per 40 numbers put them in the top ten, but as part time players in this late stage for Ginobili or early stage for Jokic, drops each a notch or three.
• Key element on for the top thirty is performing well on both sides of the ball. Only Blake Griffin and Klay Thompson score below average on defense and no players in the top thirty score a negative on offense.
• James Harden and Isaiah Thomas narrowly miss the top thirty due to defensive rating.

I expect to follow up with a Tableau table of the whole league along with some more analysis of rookies and visualizations shortly.