The thesis
Momentum SMA20 Rider is a straightforward trend-continuation play. It ranks a 24-name large-cap universe by distance above the SMA50, buys the leaders once price confirms above the SMA20, and rotates capital into stronger ranks when the book fills up. Exits are rule-bound rather than discretionary: a position is cut when RSI14 pushes past 75 (overbought) or when price closes back below the SMA20 (trend structure broken). It is a clean expression of "ride confirmed uptrends, get out when the confirmation fails."
Backtest and validation
The headline backtest is respectable: 81% total return over 1,233 days, a 12.89% CAGR, a Sharpe of 0.98, and a 20.91% max drawdown. More important than the raw number is that the strategy cleared walk-forward validation across four folds, three of which were positive, with a Probabilistic Sharpe Ratio of 0.986 and a Deflated Sharpe Ratio of 0.583 across 24 trials. The Deflated Sharpe — which discounts for the number of configurations tried — is the honest figure to anchor on, and at 0.58 it says the edge is more likely real than not, without being emphatic.
The fold-by-fold picture is where the character shows. Fold 4 (May 2025 to Aug 2026) is the standout: 43.48% return, a 2.18 Sharpe, and a shallow 5.38% drawdown. Fold 2 was similarly strong. But fold 3 (Feb 2024 to May 2025) essentially flatlined at -0.28% with a 23% drawdown — the worst of the run — and fold 1 eked out under 5%. Momentum strategies live and die on regime, and this one clearly earns its keep in trending markets while treading water in choppy ones.
The uncomfortable numbers
Two figures deserve a hard look. First, the 37.13% win rate: this is a system that is wrong on most individual trades and relies on letting winners run far past its many small losers. That is a legitimate momentum profile, but it means long stretches of losing trades are structurally normal, not a malfunction. Second, turnover of 26,971% across 1,427 trades. Fees totalled $1,427 in the backtest — roughly $1 per trade — but the churn is a real-world execution and slippage risk that a paper backtest understates.
Recent live activity
On 6 August the strategy put money to work, buying JPM, BAC, V, MA, and ABBV — a distinctly financials-heavy tilt. Since then the scheduled runs from 7 through 14 August have all reported 0 executed, 0 rejected, with cash parked at $626.50 and total equity drifting between $9,832 and $9,982. In other words, the book is invested and quietly holding, sitting modestly below the $10,000 mark. No RSI or SMA20 exits have fired, and no rotations have triggered — consistent with a full book in a calm tape.
Verdict
This is a genuinely validated trend-follower with an honest edge and honest weaknesses. The out-of-sample Sharpe and PSR are encouraging; the low win rate, high turnover, and regime-dependent fold 3 are the price of admission. Watch it through the next choppy stretch — that is where the thesis will actually be tested.