The thesis
momentum-code runs a simple, legible idea: buy the top positive movers in a 24-name large-cap universe, capped per position. The universe is deliberately vanilla — AAPL, MSFT, NVDA, the big banks, staples like PG, KO and WMT, plus energy in XOM and CVX. There is no clever factor stack here, just trend-chasing with a position cap for risk control. That simplicity is a feature: the strategy is easy to reason about and hard to overfit by accident.
The backtest
Over 451 trading days the strategy returned 19.76%, lifting a notional book to $11,976.21 — a CAGR of 10.6%. It did so cheaply: just $5 in fees, zero FX cost, and only 5 trades despite 92.74% turnover. The headline blemishes are a Sharpe of 0.65 and a 20.48% max drawdown — a lot of pain for a single-digit-to-low-double-digit annualised return. The reported 0% win rate reflects that no round-trips have closed; the five positions are all buys, so there are no realised winners to count yet rather than five outright losers.
Validation: where it stumbles
This is the most interesting part of the record. Across four walk-forward folds the strategy was positive in all four (15.38%, 3.27%, 12.18%, 13.71%), with the most recent out-of-sample window returning 13.71% at a 1.93 Sharpe. The probabilistic Sharpe ratio (PSR) of 0.811 is encouraging. And yet the validation verdict is passed: false.
The culprit is the deflated Sharpe ratio of 0.338. With 6 trials in the search, the DSR penalises the Sharpe for multiple-testing selection bias, and on that adjusted basis the edge is not statistically convincing. In other words: four green folds look great, but once you account for how many variants were tried to find them, the signal thins out. That is exactly the kind of honesty an auto-deploy gate should enforce, and here it did.
Recent activity: a stalled engine
The live paper account tells a quieter story. The last executed trades were back in late May and early June 2026 — XOM, NVDA, BAC, HON and MSFT. Every scheduled run since (Oct 12 through Oct 19) logged 0 executed and 1–3 rejected, with cash pinned at $608.79 and total equity drifting in a narrow $10,399–$10,591 band. The book is effectively idle: fully invested, with too little cash to act on new signals, so momentum-code is now a holder rather than a trader.
Verdict
Strengths: a clean, interpretable thesis, consistent fold-level profitability, and negligible trading costs. Risks: a weak risk-adjusted profile (0.65 Sharpe, 20% drawdown), a failed validation gate driven by trial-count deflation, and a live account that has been unable to execute for over a week. The strategy earns a watch, not a green light.