#交易策略误区 The Trap of Overfitting Quantitative Models

Many traders optimize parameters to achieve perfect backtesting results (such as a 99% win rate) using historical data, while ignoring changes in market structure. A typical case: the triangular arbitrage strategy in 2020 became ineffective after updates to the Binance API, resulting in some automated systems losing 23% in a single day. The correct approach should maintain simplicity in strategies, reserve 20% for out-of-sample testing, and regularly conduct stress tests (such as simulating the 2022 LUNA crash).