#我的策略演变 My trading strategy has undergone three iterations: Initially relying on moving average crossovers and RSI oversold signals, I fell into the trap of overfitting, with an annualized return volatility reaching 45%; In the mid-term, I integrated fundamental cycle judgments and on-chain data to construct a multi-factor quantitative model, improving the Sharpe ratio to 1.8, but neglecting market sentiment resonance led to a liquidation during the 2024 black swan event; Currently, I am shifting to a dynamic adaptive framework, introducing LSTM to predict market sentiment curves, dynamically adjusting positions using the Kelly formula, and validating through cross-cycle stress testing, stabilizing the annualized return at 27% with a maximum drawdown controlled at 8%. The core shift is from mechanical rule obedience to dynamic risk-reward ratio balance, fundamentally evolving cognition from linear causality to nonlinear complex systems.