Learn

PLAYER
ARCHETYPES

Traditional NBA positions — point guard, center, forward — were designed for a game that no longer exists. APEX uses K-means clustering on six offensive efficiency variables to sort players into eight data-driven archetypes. The result is a set of peer groups that actually reflect how players function, not where they stand on the court.

These archetypes matter because APEX normalizes three of its five pillars — Shot Quality, Creation & Playmaking, and Physical Contribution — within each player's archetype peer group. A rim runner's TS% is compared to other rim runners, not to guards. An efficient wing's low usage isn't penalized against a primary creator's workload.

8
Archetypes
6
Clustering Variables
4
Seasons Mapped
K=8
K-Means Clusters
Clustering Variables
USG% TS% FTA Rate AST% TOV% REB%
Primary Creator
The rarest archetype. Very high usage combined with elite playmaking — players who both carry the offense and orchestrate it.
Jokić · Giannis · Luka
→
Shot Creator
High usage, secondary playmaking. Volume scorers who generate their own offense and carry the heaviest individual shot burden.
Tatum · Booker · Fox
→
Pure Playmaker
Elite AST% at controlled usage. Pass-first floor generals who derive most of their value from facilitating, not scoring.
Haliburton · CP3 · Garland
→
Secondary Creator
Moderate usage and moderate AST% — the versatile middle ground. Second-unit initiators, complementary stars, and scoring wings with real playmaking upside.
Maxey · Middleton · White
→
Efficient Wing
Low usage, very high TS%, minimal playmaking. 3-and-D wings and spot-up specialists who score selectively and almost never waste a possession.
Mikal Bridges · OG Anunoby
→
Role Player
Minimal offensive footprint, defensive focus. These players define their value through defense, hustle, and doing exactly what the system asks — nothing more.
Tucker · Finney-Smith · Horford
→
Versatile Big
Dominant rebounding with real passing ability and above-average efficiency. Skilled frontcourt players whose game extends beyond the paint.
Bam · Sabonis · KAT
→
Rim Runner
Very high TS% at very low usage. Pure finishing centers who generate elite efficiency through dunks, lobs, and FTA — but rarely handle the ball.
Gobert · Jarrett Allen · Capela
→
Methodology Note

K-means cluster IDs are arbitrary and reshuffle when the player pool changes each season. After each season refresh, centroids are inspected manually and archetype names are re-mapped. Cluster assignment is based on Brill et al. (2023), which independently identified 8 stable functional archetypes using K-means on 48 variables — convergent with APEX's 6-variable clustering.

Archetype labels describe a player's offensive style — not their defensive role, which is handled separately via BBall-Index defensive role groupings for the Defensive Impact pillar.