The thesis
Momentum SMA20 Rider is a disciplined trend-follower. It ranks a 24-name large-cap universe — think AAPL, MSFT, NVDA, JPM, V, XOM and a spread of staples and healthcare — by how far each trades above its SMA50, then buys the leaders only once price confirms above the SMA20. Exits are mechanical: trim when RSI14 pushes past 75 (overbought) or when a close breaks back below the SMA20 (structure broken), and rotate freed capital into stronger-ranked names when the book is full. It's a clean expression of a well-worn idea: let winners run, cut trend breaks quickly.
Recent activity
The strategy has traded every scheduled session over the past two weeks, usually one to two fills a day with the occasional rejection when cash runs thin. Equity sits around $9,740 and cash has swung from as little as $41 to over $2,200, so the book is frequently near fully invested. The recent tape is telling: sells in COST, V, AAPL and MSFT, buys in PG, WMT, XOM and back into V. The rotation is currently leaning toward staples and energy — a reminder that this model follows whatever is furthest above its SMA50, not a sector view.
Backtest and validation
Over 1,233 days the backtest compounded to $18,100 from the rules alone — +81% total, ~12.9% CAGR, with a Sharpe of 0.98 and a 20.9% max drawdown. Crucially, it survived walk-forward validation: 3 of 4 folds positive, a strong PSR of 0.986 and a DSR of 0.583 across 24 trials, which is the part that matters most — it says the edge isn't just the luckiest of two dozen variants. The standout is the most recent fold (May 2025–Aug 2026): +43.5% at a 2.18 Sharpe with only a 5.4% drawdown.
Strengths and risks
The strengths are real: mechanical exits, validation that accounts for multiple testing, and a recent out-of-sample stretch that looks excellent.
The risks are just as real:
- Win rate of 37%. Fewer than four trades in ten are profitable; the entire return depends on asymmetry between winners and losers. That is normal for momentum, but it means long, demoralising losing streaks are baked in.
- A dead fold. Feb 2024–May 2025 returned −0.28% at a 23% drawdown. In a choppy, trendless regime this model does nothing but churn.
- Heavy turnover. ~27,000% turnover over the window and 1,427 trades — fees are modest here ($1/trade) but in a higher-cost venue that friction bites.
Bottom line
A validated, honestly-built trend-follower with a genuine edge — provided you can stomach being wrong most of the time and sitting through flat regimes to collect the trends that pay.