Analyzing wins & losses…
No data available
Run compute_winning_keys.py --team ATL to generate this team's analysis.
Key Winning Factors
Statistically significant controllable behaviors that separate wins from losses (Benjamini-Hochberg corrected, Cohen's d ≥ 0.3)
Shot Chart Comparison
Shot Zone Breakdown
Per-game averages in wins vs losses from NBA Stats API (aggregate comparison — no per-game variance available)
| Zone | W FGA | W FG% | W Freq | L FGA | L FG% | L Freq | ΔFGA | ΔFG% |
|---|
Win Correlation Ranking
Features ranked by point-biserial correlation with win/loss outcome. Color = controllability. Gray = not significant after FDR correction.
Player Splits
Per-game averages in wins vs losses from NBA Stats API. Sorted by point differential.
| Player | W PTS | W AST | W FGA | W 3PA | L PTS | L AST | L FGA | L 3PA | Δ+/- | GP (W/L) |
|---|
Touch & Movement Splits
W/L aggregate means from NBA tracking data. These are per-game averages in wins vs losses — not per-game distributions, so no t-test available.
Predictive Model
Logistic regression trained on per-game box score features with L2 regularization. Win probability = f(controllable inputs).