--- title: Nfl Milestone App2 emoji: 📉 colorFrom: gray colorTo: green sdk: gradio sdk_version: 5.45.0 app_file: app.py pinned: false short_description: Unique player distributions per stat (NFL) --- This app uses Bayesian Model Averaging (Poisson, Negative Binomial, Gamma, Log-Normal) to fit each player's historical stat distribution and estimate personalized milestone probabilities. Upload your **NFL player game logs** CSV (schema below), pick a player + stat, set a projection (historical mean or custom), hit **Submit**, and view milestone probabilities, fair odds, and vig-adjusted odds. **CSV schema (minimum):** - `player` (string) - `game_date` (date or string) - One or more stat columns; see `Supported Stats` in the app. - (Optional for outlier filtering) `actual_opportunities`, `projected_opportunities` OR `actual_snap_pct`, `projected_snap_pct` - If provided, the app can flag and filter “usage outliers” using a % deviation threshold. Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference