The thesis
Momentum SMA20 Rider is a textbook trend-follower. It ranks a 24-name large-cap universe — the usual megacaps plus staples, financials, healthcare and energy (AAPL, MSFT, NVDA, JPM, JNJ, XOM and the like) — by how far each trades above its SMA50, then buys the leaders whose price still holds above the SMA20. Exits are mechanical: cut when RSI14 pushes past 75 (overbought) or when price closes below the SMA20 and the trend structure breaks. When the book is full, capital rotates out of laggards and into stronger ranks. It is a disciplined "let winners run, cut the broken" design, and it is currently running live.
Backtest and validation
Over 1,233 days the strategy returned 81%, lifting a notional book to $18,100, for a 12.89% CAGR at a 0.98 Sharpe and a 20.91% max drawdown. Encouragingly, it survived walk-forward scrutiny: validation passed with 3 of 4 folds positive. The most recent fold (May 2025–Aug 2026) is the standout, up 43.48% at a 2.18 Sharpe with only a 5.38% drawdown. Out-of-sample metrics carry that strength through.
The robustness statistics are a mixed but honest signal. The Probabilistic Sharpe Ratio is a confident 0.986, but the Deflated Sharpe Ratio — which penalises the 24 trials run during search — falls to 0.583. That gap is the multiple-testing tax, and it is the right number to respect before over-trusting the headline Sharpe.
The risks in the numbers
Two figures deserve caution. First, the win rate is just 37.13% across 1,427 trades — the strategy loses on most positions and relies on a minority of large winners, the classic trend-following profile that is psychologically hard to hold through drawdowns. Second, turnover is roughly 26,972%, an enormous churn that racked up 1,427 in fees. In this paper run FX cost was zero and fees were modest, but at scale or with wider spreads that turnover is a real drag.
The folds also show the regime dependence: fold 3 (Feb 2024–May 2025) essentially flatlined at −0.28% with a 23% drawdown, its worst stretch. Momentum works until it doesn't, and this strategy will chop in trendless, mean-reverting markets.
Recent live activity
The last two weeks have been quiet but active. Scheduled runs executed steadily — buying PFE, AAPL, MSFT and NVDA, while trimming MSFT, JNJ, KO, CVX and MA as ranks shifted. Rejections were common (three of three rejected on 21 Sep), consistent with a full book and tight entry filters. The live paper equity, however, has drifted down over the shown sessions, from $9,893.88 on 18 Sep to $9,714.27 on 24 Sep — about −1.8%. Cash swung widely too, from $255 fully invested to over $5,800, reflecting the high-churn rotation.
Verdict
The design is coherent and the validation is more than cosmetic. But the low win rate, heavy turnover and deflated Sharpe say this is a strategy to size carefully and judge over a full cycle — not one week of soft live returns.