Survivorship Bias Blind Spot in Backtesting #交易策略误区

Only studying existing trading varieties can severely overestimate returns. For example, when backtesting the TOP 10 cryptocurrency strategies from 2020 to 2023, if delisted projects (such as LUNA) are not included, returns will be exaggerated by 300%. Rigorously, the approach should be: 1) Include all trading targets during the same period; 2) Set a "zombie coin" filter; 3) Use survival analysis models.