APEX (Advanced Plus-Minus Experience) measures how much a player actually moves the scoreboard — adjusted for teammates, opponents, game leverage, and recency.
Most advanced stats either hide their methodology or rely heavily on box-score inputs that miss half the game. APEX takes a different approach:
No box-score shortcuts. APEX uses only on-court stint data — who was playing and what happened to the score.
See every adjustment step: raw +/-, leverage, recency, and teammate/opponent effects. Nothing hidden.
Ratings recalculate every morning during the season with the latest game data.
Raw plus-minus can't answer that question. Role players on great teams look like superstars, and stars on bad teams look average. APEX isolates individual impact from context by adjusting for every teammate and opponent sharing the court.
APEX starts with raw plus-minus data and applies three adjustments to isolate true player value:
Points scored minus points allowed, per 36 minutes
Normalizes for playing time so starters and bench players can be compared fairly.
Close games count more
A +10 in a 2-point game matters more than a +10 in a 30-point blowout. APEX weights every possession by how much the game outcome was still in doubt — rewarding clutch play over garbage-time stats.
Current form weighted over historical peaks
A player's 2025 performance tells us more about their ability today than their 2018 numbers. APEX applies per-player recency weighting so recent seasons count more.
Isolating individual impact from lineup context
The core of APEX. Using an iterative solver across every lineup combination in the dataset, APEX mathematically separates each player's contribution from their teammates and opponents. That Warriors bench player's +8.0? Their true impact is closer to +1.5.
True individual impact per 36 minutes
Positive means the player helps their team outscore opponents. Negative means the opposite. The higher the number, the greater the impact.
Tier cutoffs are computed dynamically from 331 qualified active players with 2,000+ career minutes. Players with fewer than 2,000 minutes at a given position render in italic gray as Unqualified — shown for context but not ranked alongside qualified players.
APEX is the only public metric that pairs pure stint-level plus-minus (no box-score prior) with leverage weighting, per-player recency, and teammate/opponent strength.
PER · Win Shares · BPM · VORP · FIC · WPA
Box-score ratings grade the stat sheet, so they miss defense, spacing, and off-ball value and reward volume over efficiency — which is why NBA front offices broadly distrust them. APEX never reads the box score; it reads what happened to the scoreboard while you were on the floor.
On/Off Net Rating · raw plus-minus (NBA.com · Cleaning the Glass · PBP Stats)
On/off and raw +/- are uncontrolled: they inflate value due to strong teammates, weak benches, and garbage time. APEX runs a simultaneous teammate/opponent solve to isolate your individual signal, then weights each minute by game leverage.
Regularized Adjusted Plus-Minus — the backbone every hybrid regresses toward
RAPM shares APEX's no-box-score-prior philosophy, which is why front offices trust it. APEX builds on that foundation with two things plain RAPM lacks: leverage weighting and a per-player recency taper.
EPM · DARKO/DPM · LEBRON · RAPTOR · RPM · PIPM
The catch-alls front offices trust most — but each blends a box-score prior for stability. APEX keeps the plus-minus core and drops the prior, adding leverage weighting.
What data does APEX use?
Stint-level data from every regular season game since 2003-04 — over 20 years of play-by-play data, millions of possessions, and thousands of unique lineup combinations. Players are tracked at each position they've played.
How often is it updated?
Daily. The pipeline runs every morning during the season, collecting new game data and recalculating all ratings.
How is a player's position decided?
Each player is rated at exactly one position: the one with the highest APEX where they have 2,000+ career minutes. If no position reaches that threshold, the highest-minutes position is used. Position is assigned by on-court body size within each lineup (smallest = Guard, largest = Center, the middle three = Wing), not by NBA.com listings.
How does APEX differ from EPM, DARKO, LEBRON, RAPM, and BPM?
EPM, DARKO, LEBRON, RAPTOR and RPM all blend box-score priors into their models, and BPM is a pure box-score regression. APEX — like RAPM and raw on/off — uses no box-score prior at all, but unlike them it weights each minute by game leverage and tapers to every player's current form.
Which metrics do NBA front offices actually use?
Surveys of NBA executives (e.g. HoopsHype) consistently find box-score metrics like PER and Win Shares are distrusted, while adjusted plus-minus models with priors — EPM, DARKO, LEBRON, and pure RAPM — are preferred. APEX sits in that adjusted-plus-minus family but improves on RAPM's no-box-score-prior philosophy further, adding leverage weighting and per-player recency.
Explore every player's APEX rating — the full bridge-metric breakdown shows each adjustment step, free for everyone.
