Skew-normal from the model's own per-alliance spread, centred to match the published odds

Dotted labels are illustrative: we store each alliance's auto, teleop, endgame and foul points, not the game-piece detail behind them, so the itemised rows are a plausible spread that adds back up to the totals we do have. Field names match what the engine already parses.