The thesis
momentum-code runs a simple, legible idea: buy the top positive movers in its universe, capped per position. The universe is 24 large-cap US names spanning tech, financials, healthcare, staples and energy — AAPL, MSFT, NVDA, JPM, V, JNJ, XOM and the like. There is no clever factor stack here, just trend-following with a position cap to keep any single name from dominating the book. That simplicity is a genuine strength: the behaviour is easy to audit and hard to overfit by accident.
Backtest and validation
Over 451 days the backtest returned 19.76%, ending at $11,976 on a $10,000 base — a 10.6% CAGR. Risk-adjusted, the picture is more modest: a Sharpe of 0.65 against a 20.48% max drawdown. In other words, the worst peak-to-trough loss was roughly twice the annualised return, so the ride was not smooth.
Walk-forward validation is where it gets interesting. All four folds were positive — 15.38%, 3.27%, 12.18% and 13.71% — with fold Sharpes as high as 2.47. Out-of-sample return was 13.71% at a 1.93 Sharpe, and the probabilistic Sharpe ratio (PSR) came in at a healthy 0.811. On the surface, that is an encouraging, consistent profile.
And yet validation did not pass. The deflated Sharpe ratio (DSR) sits at just 0.338, and the run logged 6 trials. Once you adjust for multiple testing, the edge thins out — the strategy cleared every individual fold but not the deflated bar meant to catch luck dressed up as skill. That is the gate doing its job, not a glitch.
A thin sample
The single biggest caveat is sample size: just 5 trades in the backtest, and the same count in each fold. A reported win rate of 0% sits oddly against a clearly positive total return — with only a handful of trades, these ratios carry almost no statistical signal and should not be read as a losing record. Turnover of 92.74% on so few trades simply confirms this is a concentrated, low-frequency book, not a diversified engine.
Recent activity: the lights are on, nobody trades
Live behaviour is the real concern. The last executed trades were in late May and early June 2026 — buys in MSFT, HON, BAC, NVDA and XOM. Since then, the last six scheduled runs (2 Oct through 9 Oct) all report 0 executed with 1–3 rejections each. Cash is stuck at $608.79 against a total of roughly $10,400–$10,550, which drifts day to day purely on mark-to-market.
The read is straightforward: the book is effectively fully invested, and the position caps plus thin cash are blocking every new buy signal. That is not a crash — total equity is holding near $10,500 — but a momentum strategy that cannot rotate into fresh movers has quietly stopped being a momentum strategy.
Verdict
Consistent folds and a strong PSR are real positives, but the failed DSR gate, a five-trade sample, and three months of rejected runs are a cluster of warnings. The engine looks sound in theory; in practice it needs its cash and cap logic unblocked before its live record means anything.