Research on Momentum Strategy Optimization Based on Fractal Winner-Loser Portfolio

The selection criteria for the traditional momentum strategy do not take into account that the tails of security returns follow a power-law distribution and most of them assign equal weights to the winner and loser portfolios. With the help of fractal theory, we construct a fractal momentum strategy by using fractal statistical measures as selection criteria to select the real winner and loser portfolios: fractal winner and loser portfolios; furthermore, we construct a momentum strategy based on fractal winner-loser portfolios (FPMS) by using a fractal portfolio model with constraints to assign weights to fractal winner and loser portfolios. The fractal momentum strategy does not significantly outperform the traditional momentum strategy. Even when other standard strategies underperformed, FPMS still achieved a positive excess return, a higher Sharpe-ratio and a positively skewed return. The returns of adjusted FPMS also have a high Sharp-ratio, and they are positively skewed.