Reconstruct the option chain that existed when an opening-range signal fired, test realistic ask-to-bid returns, and compare a perfect-hindsight winner with a fixed selection rule that could have been followed at the time.
Uses logged ORB events from options_data/events.csv. Contracts are
selected from entry-time spread, DTE, delta/moneyness, volume and OI. The newest 30% of
events are reported separately as a chronological holdout.