HOW NBA SIM WORKS

NBA SIM is a data-driven prediction system that analyzes coaching schemes, player archetypes, and lineup synergy to predict game spreads and over/unders. All data sourced from 2025-26 NBA season statistics via the official NBA API.

MOJO (Morello's Optimized Joint Output) — 33 TO 99

Every player gets a MOJO score from 33-99 using a 75% offense / 25% defense split plus shared impact components, blended with team synergy context.

Defensive Rating (DRtg) matters: a player with 107 DRtg earns ~20 defensive points, while 112 DRtg earns only ~7.5. Elite defenders and rim protectors get a meaningful MOJO boost.

MOJO Range shows the expected floor-to-ceiling for each player based on their score volatility. Elite players (MOJO 83+) have tighter ranges, while mid-tier players have wider variance.

Team MOJO Ranking (1-30) is the minutes-weighted average MOJO score across each team's top 10 rotation players, weighted by minutes per game.

PLAYER ARCHETYPES

Players are clustered into archetypes using K-Means on 16 statistical features per position group. The optimal number of clusters (K) per position is chosen by silhouette score, with a minimum of 3 archetypes per position. Features include per-36 rates, efficiency metrics, and impact stats — weighted differently per position (e.g., assists weighted 1.5x for PGs, blocks 1.5x for centers).

COACHING SCHEMES

Each team's offensive and defensive schemes are classified from play type distributions, pace, and shooting profiles using percentile-rank comparison.

MOJI SPREAD MODEL — 10-STEP PIPELINE

The SIM runs a 10-step pipeline to produce projected spreads and totals for every game. Real lines from sportsbooks via RotoWire. Projected lines marked (PROJ.) where real data is unavailable.

Player props marked (PROJ. LINE) are season averages adjusted for opponent defense and pace. Real player props replace projections when released by sportsbooks.

LINEUP COMBINATION ANALYSIS

We track 2-man, 3-man, 4-man, and 5-man lineup combinations across the entire season. Each combo is evaluated by net rating, minutes played, and games played together. Minimum thresholds ensure statistical reliability:

Hot Combos show lineups with the best net ratings this season — these units are outscoring opponents significantly when they share the floor. Fade Combos are the worst-performing groups — teams bleed points when these players are together.

Trend badges (🔥 HEATING UP, ⚡ ELITE FLOOR, 💀 DISASTERCLASS, etc.) use game count and net rating thresholds to flag the most notable lineup trends across the league.

PROP SCORING

Player props are scored on a normalized 0-100 scale per prop type (PTS, AST, REB, PRA, STL+BLK). Each prop is adjusted by opponent defensive rating, pace matchup, and player efficiency (TS%). OVER props trigger for favorable matchups; UNDER props trigger when facing elite defense or when a player has poor efficiency (TS% below 53%).