The thesis
"Momentum SMA20 Rider" is an unapologetic trend-follower. It ranks a 24-name large-cap universe — the usual megacap tech, financials, staples, and energy suspects like AAPL, NVDA, JPM, XOM, and PG — by distance above the SMA50, then buys the leaders whose price is still holding above the SMA20. Exits are mechanical: it cuts a position when RSI14 pushes past 75 (overbought) or price closes back below the SMA20 (trend structure broken), rotating freed capital into higher-ranked names when the book is full. It is a clean, rules-based expression of "ride confirmed uptrends, don't argue with them."
Recent activity
The strategy is live and trading on schedule. Over the six sessions from 18–25 September, the paper book drifted from a total of $9,893.88 down to $9,715.59 — a soft ~1.8% pullback, with cash swinging widely (from as little as $254.75 fully invested to $5,802.65 after a heavy sell day on 21 September). Recent fills show the rotation in action: it dumped KO, CVX, and MA on the 21st, then rebuilt into NVDA (8 shares at $229.36), AAPL, and MSFT. MSFT in particular saw some churn — sold at $497.53 on the 24th, bought back at $516.46 the next day — a reminder that whipsaw is the tax this style pays.
Backtest and validation
The headline numbers are genuinely good. The backtest returns 81% total (a 12.89% CAGR over 1,233 days), finishing at $18,100.23 with a Sharpe of 0.98 and a max drawdown of 20.91%. Crucially, it survives scrutiny: walk-forward validation passed with 3 of 4 folds positive, and the deflated Sharpe (DSR) of 0.583 — which penalizes the 24 trials run — stays comfortably above zero. The most recent fold (May 2025–Aug 2026) is the standout: +43.48% at a 2.18 Sharpe with just a 5.38% drawdown.
The risks
Balance demands the caveats. First, the win rate is only 37.13% across 1,427 trades — this strategy is wrong most of the time and relies entirely on letting winners run, so a stretch without a strong trend will bleed. Fold 3 (Feb 2024–May 2025) proves the point: essentially flat at −0.28% with a 23% drawdown, its worst regime. Second, turnover is a staggering 26,971%, and while fees are modest ($1,427, roughly $1/trade) in the model, real-world slippage on that churn is a live risk the MSFT round-trip hints at. Third, the gap between an in-sample-flattering PSR of 0.986 and the more sober 0.583 DSR is the honest measure of how much luck to discount.
Verdict
A credible, well-validated trend engine with a strong recent tailwind — but one whose returns are lumpy, regime-dependent, and turnover-heavy. Worth riding while trends persist; worth watching closely the moment they don't.