Rethinking Synthetic Scenario Realism: Compatibility, Not Fidelity, Drives Hedging Performance
This work investigates how synthetic price paths for deep hedging should be evaluated. It moves beyond simple statistical realism by introducing 'compatibility,' which measures risk discrepancies between synthetic and real markets for specific hedging strategies. Findings show that statistical realism alone is insufficient; generator suitability depends on the interaction between the generator, the hedger class, and the hedging task.