Forecasts created by quantitative models—not box scores
Each analysis sheet is designed to report estimated future outcomes: model win probability, expected score, margin, total, player projections when supported, and uncertainty. These examples demonstrate the format.
Brooklyn Nets at Cleveland Cavaliers
Player projection example: Donovan Mitchell — 27.4 projected points, with a supplied illustrative range of 15–41; 5.6 assists and 3.2 made threes.
These are forecast estimates, not a player's season averages or recorded game results. Ranges indicate uncertainty, not guaranteed bounds.
Example City Hawks at Example Bay Wolves
Player projection format: Starting quarterback — projected passing yards, attempts and touchdowns; running back — projected rushing yards and attempts, where supported.
All NFL team names and figures above are fictional placeholders. Real NFL figures will only be published when generated and approved by the FAT SPORTS model pipeline.
The underlying models may incorporate historical team performance, opponent strength, statistical estimation and uncertainty calibration. Specific inputs and model readiness vary by sport. Forecasts may be withheld when inputs are incomplete or reliability is insufficient.