The thesis
momentum-code runs a simple, well-worn idea: buy the top positive movers in a 24-name large-cap universe — the usual megacap tech, financials, staples, healthcare and energy suspects — and cap exposure per position. There is no clever signal stack here; the edge, if any, comes from riding short-term relative strength and letting the position cap enforce diversification. Simplicity is a virtue for auditability, but it also means the strategy lives and dies by whether momentum persists in this particular universe.
Backtest and validation
The headline backtest is respectable: 19.76% total return over 451 days, a 10.6% CAGR, and final equity of $11,976 from a $10k base. Turnover of ~93% is moderate, and fees are negligible ($5 total). The catch is a 0.65 Sharpe against a 20.48% max drawdown — the return is real but the ride is bumpy relative to the volatility taken.
The walk-forward record is the encouraging part. All four folds came back positive (15.38%, 3.27%, 12.18%, 13.71%), with out-of-sample return of 13.71% and an out-of-sample Sharpe of 1.93 in the most recent fold. Individual folds even printed Sharpes above 2.3 with drawdowns near 6%. That consistency is a genuine mark in its favour.
And yet the validation gate reports passed: false. The reason is in the deflation math: the probabilistic Sharpe ratio is a solid 0.811, but the deflated Sharpe ratio falls to 0.338 once you account for 6 trials. In plain terms, after penalising for how many variants were tested, the strategy's edge is no longer statistically convincing. Four green folds are nice, but five trades per fold is a thin sample, and the framework is right to be sceptical.
Recent live activity
The live tape has gone quiet in a way worth flagging. The last executed trades were a cluster of buys in late May and early June — XOM, NVDA, BAC, HON and MSFT. Since then, every scheduled run from 20–27 August has printed 0 executed, 1–2 rejected, with cash pinned at $608.79 and portfolio value drifting in a narrow band around $10,460–$10,570. That is roughly break-even to modestly positive live, and a live return well short of the backtest's 19.76%.
The repeated rejections against a small cash buffer suggest the strategy is fully invested and unable to rotate into fresh momentum names without freeing capital first — a practical frictions problem the backtest does not capture.
Verdict
momentum-code has an honest, legible thesis and a genuinely consistent fold history. But the failed deflation gate, the tiny trade count, the 20% drawdown, and a stalled, cash-starved live book all argue for patience over promotion. Watch whether it can start rotating again before reading anything into the flat live curve.