Quantitative Trading Strategies: Momentum, Mean Reversion, and Beyond

Module: Fintech AI | Difficulty: Advanced

Momentum

Mean Reversion

Pairs Trading

Factor Returns

import numpy as np
from scipy.stats import zscore

class MomentumStrategy:
    def __init__(self, lookback=252, holding=21):
        self.lookback = lookback; self.holding = holding
    def signal(self, prices):
        returns = prices.pct_change(self.lookback)
        return returns.iloc[-1]
    def positions(self, signal, n_long=10, n_short=10):
        ranked = signal.rank(ascending=True)
        n = len(ranked)
        long = ranked >= n - n_long
        short = ranked <= n_short
        return long.astype(int) - short.astype(int)

class PairsTrading:
    def __init__(self, window=60, entry_z=2, exit_z=0):
        self.window = window; self.entry_z = entry_z; self.exit_z = exit_z
    def hedge_ratio(self, y, x):
        from sklearn.linear_model import LinearRegression
        reg = LinearRegression().fit(x.reshape(-1,1), y)
        return reg.coef_[0]
    def signal(self, y, x):
        beta = self.hedge_ratio(y, x)
        spread = y - beta * x
        z = (spread[-1] - spread.mean()) / spread.std()
        if z < -self.entry_z:
            return 'long_spread'
        elif z > self.entry_z:
            return 'short_spread'
        elif abs(z) < self.exit_z:
            return 'close'
        return 'hold'

Research Insight: Momentum and mean reversion are complementary strategies that perform well in different regimes. Combining them through a regime-switching model improves overall performance. The key insight is that momentum works in trending markets while mean reversion works in range-bound markets.