algo_betting

Functions

compute_individualized_market_brier(→ dict)

Compute per-forecaster individualized market baseline Brier scores.

Module Contents

algo_betting_.compute_individualized_market_brier(forecasts: pandas.DataFrame) → dict

Compute per-forecaster individualized market baseline Brier scores.

For each forecaster, uses the same filtered event set (dedup) and computes the Brier score as if the market prices were the predictions:

  • YES side: (yes_ask - outcome)^2

  • NO side: (no_ask - (1 - outcome))^2 [uses actual no_odds, NOT 1 - odds]

Direction is determined by the forecaster’s actual prediction vs odds. Forecaster-events with any market prediction inside the bid-ask spread are skipped.

Returns:

Dict mapping forecaster name -> individualized market baseline Brier score.