Trend-Based Machine Learning Crypto Strategy
Log in to collectAcademic paper
Trend-based Forecast of Cryptocurrency Returns
Xilong Tan; Yubo Tao
- MOUniversity of Macau
- ?University of Macau - Department of Economics
Strategy in a nutshell
This strategy predicts cryptocurrency returns using 24 technical indicators based on momentum, volume, and moving averages. The nonlinear scaled sufficient forecasting (sSUFF) method processes these signals to guide weekly portfolio allocation.
Economic rationale
Trend-based factors have proven predictive power, and sSUFF’s nonlinear dimension reduction enhances signal extraction. This improves forecast accuracy and allows robust, risk-adjusted allocations across the cryptocurrency market.
Backtest performance
Annualised return12.6%
Volatility18.26%
Sharpe ratio0.69