Behind all those advanced metrics I love — TS+, ratings, per‑75 numbers — hide the much more obscure statistics that adjust and measure a player’s impact on the floor (DARKO, LEBRON, EPM, RAPM). I thought it would be interesting to share the DARKO DPM results, explain how it works, and highlight its contextual and structural limits, obviously through the lens of the Suns.
Before anything else, it’s important to remember that all these metrics aim at the same thing: understanding a player’s or a team’s
real impact on point differential, independently of the immediate context. They’re built differently, but they all try to answer the same question: how many points does a player add per 100 possessions?
Let’s go back to the model created by Kostya Medvedovsky: DARKO DPM. It’s a projection system that estimates a player’s true future level by using all his past performances, weighted intelligently by recency, and updated every single day through a Bayesian model and a Kalman filter.
Contextual adjustments are also baked into the recipe: age, fatigue, team environment, season‑long trends, and interactions between different stats all influence the final output. In other words, DARKO DPM can be summarized in one question: What can we expect from a player based on everything he’s done so far?
These metrics can be split into two parts: offensively with ODPM, and defensively with DDPM. So what does that look like for the Suns? Offensively, Phoenix has only five players in the positive: Devin Booker (+3.0), Grayson Allen (+1.0), Collin Gillespie (+0.9), Jalen Green (+0.9), and Dillon Brooks (+0.8).
It may still feel abstract, but concretely, it means that a player improves his team’s offense compared to an average player (and yes, players with responsibilities, good efficiency, and clean decision‑making are favored).
Take Booker for example. His score is so high because he’s one of the best offensive engines in the league (and you see it immediately when he’s not there…). He’s in the 99th percentile in points created (PTS + AST), with solid efficiency on top of that.
For Dillon Brooks and Grayson Allen, it’s more about a recent evolution in their roles. Both players have seen their usage double compared to last season. And for Gillespie, it’s a complete breakout: role, usage, stats, everything is rising at the same time. In the end, this picture confirms what we see on the court: Phoenix has very few reliable creators, and without Booker, the offense lacks structure and opportunities.
Defensively, it’s a different story.
Oso Ighodaro stands out clearly with a DDPM of +2.1. His defensive profile is extremely readable statistically: mobility, verticality, clean rim protection, and immediate impact on opponent shot quality — 22.6 shots defended per 100 possessions, -3.1% Opponent eFG%, 3.4 STOP%, 1 steal, and 2.2 deflections per game.
On a team with unstable interior defense, his minutes create a sharp contrast that DARKO picks up quickly, even on lower volume. It doesn’t mean he’s already an elite NBA defender, but in his current role, he generates a clear and consistent positive defensive impact.
Oso is followed at a distance by Ryan Dunn, Collin Gillespie, and Jordan Goodwin, all at +0.4 DDPM. Three players with real impact, but one that still depends on context and role: Goodwin is a massive possession generator, Gillespie has the “good student” profile — few fouls, steals, rebounds, solid fundamentals that prevent him from being a mismatch. And Ryan Dunn has the stopper profile: good everywhere, dominant nowhere, and without any real defensive progression so far.
These three players illustrate perfectly the types of profiles this model tends to value: disruptors, stoppers, and clean defenders — who also help consistently on the glass.
Finally, let’s talk about a negative score: Dillon Brooks at -0.9. Yes, despite his reputation as a stopper, his measurable defensive impact is clearly down this season. The data shows fewer positive defensive actions (fewer contests, less presence on the boards…), more fouls and more drives allowed, and a generally negative defensive on/off (-6.5).
In a less specialized role and within a very collective defensive scheme, his contribution becomes more diffuse and less statistically visible, which explains why DARKO rates him as a below‑average defender this year.\
But let’s be careful: general impact metrics like DARKO are extremely useful to go beyond the box score and provide a global estimate of a player’s contribution. Over the long run, they remain among the best predictive tools available. But they have structural limits. They mostly value what is measurable (efficiency, volume, rebounding, box‑score stats) and struggle to capture contextual or invisible impact: tactical role, gravity, defensive adaptability, collective fit, or individual development.
And above all, they rely on the past. They project what is probable if the environment stays similar. They cannot anticipate a coaching change, a new system, a role shift, or a major personal leap.
So the goal isn’t to reject these metrics, but to put them in their proper place: they’re trend indicators, not verdicts on real impact. True understanding emerges when stats, context, and on‑court observation converge. And that’s exactly the perspective I want to bring and share.
