Strategy Card

Regime-Adaptive Relative Momentum on Top 100 Most Traded U.S. Stocks — strategy card

Spec ID: spec-regime-adaptive-relative-momentum-top100-us-stocks-1779642517 · Generated: 2026-05-24 17:34 UTC

One-line description

Long-only regime-adaptive relative momentum strategy on the 100 most actively traded U.S. stocks. When SPY is in a positive, low-volatility regime, it buys up to 2 top-ranked stocks with positive multi-horizon momentum and confirmed trend strength, sizing positions by realized volatility and ATR-based risk controls.

Why this trade exists

The strategy attempts to harvest cross-sectional momentum in large, liquid U.S. equities while avoiding periods when broad market conditions are unfavorable. The rationale is that recent winners can continue to outperform when supported by intermediate-term trend, positive market breadth/regime, and sufficient trend strength, but momentum crashes and correlation spikes are more likely when the market index is below long-term trend or realized volatility is elevated.

The design combines relative strength with absolute trend filters: a stock must rank well on 20/60/120-day returns, trade above key moving averages, have a positive MACD histogram, and meet a minimum ADX threshold. The SPY regime filter forces cash during broad risk-off environments, while volatility-targeted sizing and ATR stops aim to reduce single-name tail risk.

Algorithm

[code omitted from public view]

Parameters

Param	Value	Notes
Rule type	REGIME_ADAPTIVE_RELATIVE_MOMENTUM	Long-only relative momentum with regime filter
Universe	Top 100 U.S. stocks by 1-year dollar volume	Annual runtime-year selection; ADRs excluded
Benchmark / regime asset	SPY	Uses SPY&US&ETF key
Market regime	SPY close > 200d SMA and 20d realized vol <= 35%	Risk-off action is exit positions and hold cash
Momentum score	0.25 * return_20d + 0.35 * return_60d + 0.40 * return_120d	Ranked descending; minimum score 0.0
Stock trend filters	Close > SMA50, close > SMA200, EMA20 > EMA50, MACD histogram > 0, ADX14 >= 18	Requires broad and single-name trend confirmation
Max positions	2	Select top-ranked passing stocks
Position sizing	Volatility-targeted equal cap	20d realized vol; 18% annualized vol target per position
Position caps	Min 5%, max 10% per position	Max leverage parameter 4.0
ATR risk controls	ATR(20), initial stop 2.5x, trailing stop 3.0x, take profit 6.0x	Single-name risk management
Exit rank threshold	Exit if momentum rank falls below 4	Keeps holdings near the top of the relative momentum list
Indicator warmup	252 trading days	Required for SMA200, 120d momentum, MACD, ADX, ATR, realized volatility
Backtest window	2020-01-01 to 2025-10-08	Daily bars
Costs	Not specified	Metrics appear to exclude explicit commissions, fees, borrow, taxes, and slippage assumptions unless handled externally

Look-ahead audit

#	Concern	Status
1	Indicator look-ahead	Signals use trailing daily prices and require warmup; verify close-at-close execution does not use unavailable closing data for same-bar fills.
2	Universe look-ahead	Universe is selected annually by 1-year dollar volume at runtime; verify the 1-year volume window is strictly prior to selection date.
3	Survivorship bias	Universe contains current/top liquid names; confirm historical delistings and constituents are included if production-quality inference is required.
4	Corporate actions	Verify prices and volumes are adjusted consistently for splits, dividends, and symbol changes.
5	Regime filter timing	SPY SMA and realized volatility should be computed only from data available up to the signal timestamp.
6	Execution assumptions	Costs are unspecified; assess sensitivity to slippage, close auction liquidity, and turnover from 2,857 trades.

Caveats / known limitations

• Backtest does not specify explicit transaction costs or slippage; turnover is high enough that implementation frictions may materially reduce returns.
• The strategy holds at most 2 stocks, creating concentrated idiosyncratic exposure despite using liquid large-cap names.
• Universe construction may introduce survivorship or selection bias unless historical constituents and delisted securities are included.
• Close-based signal and rebalance logic must be audited carefully to avoid same-close look-ahead assumptions.
• Momentum strategies can suffer sharp reversals after crowded trend episodes or during regime transitions.
• The 2020-2025 sample includes unusual crisis, stimulus, inflation, and mega-cap-led periods; robustness outside this window is uncertain.
• Parameter count is relatively high, creating overfitting risk without walk-forward or out-of-sample validation.

Results

Over 2020-01-01 to 2025-10-08, the backtest produced a 46.39% total return with a 1.08 Sharpe, 1.50 Sortino, 7.48% annualized volatility, and -13.22% maximum drawdown. The return profile is relatively controlled for an equity momentum strategy, likely helped by the SPY regime filter and volatility-targeted sizing, but the Calmar of 0.52 and 2,857 trades suggest that drawdown efficiency and trading-friction sensitivity should be investigated before treating the result as deployable.

Backtest metrics snapshot

Metric	Value
Total Return	46.39%
Sharpe	1.08
Sortino	1.50
Calmar	0.52
Max Drawdown	-13.22%
Volatility	7.48%
Win Rate	67.68%
Profit Factor	1.51
Total Trades	2857
Symbols	100 (AAPL, ABBV, ACN, ADBE, ADI, AMAT, AMD, AMGN, AMZN, APP, +90 more)