Forecast Hedging and Calibration
This work addresses forecast reliability for decision-making in fields like economics and AI, offering incremental improvements in calibration techniques.
The paper tackles the problem of forecast calibration by introducing forecast hedging, a method to ensure forecasts improve expected track records, and presents a simpler calibrated forecasting procedure for binary events.
Calibration means that forecasts and average realized frequencies are close. We develop the concept of forecast hedging, which consists of choosing the forecasts so as to guarantee that the expected track record can only improve. This yields all the calibration results by the same simple basic argument while differentiating between them by the forecast-hedging tools used: deterministic and fixed point based versus stochastic and minimax based. Additional contributions are an improved definition of continuous calibration, ensuing game dynamics that yield Nash equilibria in the long run, and a new calibrated forecasting procedure for binary events that is simpler than all known such procedures.