multi-factor-ranking

You rank stocks by computing per-factor Z-scores across the universe at each date and summing them into an equal-weight or user-weighted composite score. This cross-sectional standardization makes different factor scales comparable before ranking.

Cross-sectional Z-score standardization normalizes factor values across the entire stock universe at each date so different metrics like momentum and volatility become comparable. It ensures no single factor dominates the composite score due to scale differences.

You construct a TopN long-short portfolio by ranking instruments using the composite score and discretizing to positions according to top_n and bottom_n parameters. The engine assigns equal weights to selected names while keeping others at zero with rebalance handling.

Yes, the multi-factor backtesting workflow relies on pandas and numpy as its core dependencies. These libraries handle the panel data processing, Z-score calculations, and cross-sectional ranking operations required for equity strategy research.

Yes, the engine provides panel-aware support to blend Alpha Zoo factors with optional cross-sectional standardization. You can blend multiple alpha signals and discretize them into either long-only or long-short portfolio positions.

The best way is applying Z-score normalization across the universe at each date to standardize momentum, reversal, volatility, and volume ratio factors. This ensures all factor values share a comparable scale before being summed into the composite ranking score.