Next Season Forecast
Who is the model expecting to improve or decline?
Mean reversion-based forecast for each team's next season, ranked by predicted points percentage. Wide confidence intervals reflect how much of next season's outcome the model cannot explain.
How to read this forecast
This forecast is based on mean reversion alone - the strongest signal in the research. The model explains roughly 19% of the variation in team performance change, which means about 80% of what actually happens next season comes from factors not in the model: trades, internal development, injuries, coaching changes, and noise. Treat the prediction as a baseline expectation, not a confident bet.
League forecast table
Model Accuracy shows how close each prediction was as a ratio. Statistical consistency is evaluated separately by whether the actual result falls within the model's 95% prediction interval.
| Loading forecast... |
How accurate is the model?
95% CI flag: Yes means the outcome is statistically consistent with model uncertainty; No means the miss is larger than expected model noise.
Calculating historical calibration...