The thesis
momentum-code runs one of the oldest ideas in the book: buy the top positive movers in a fixed large-cap universe, capped per position. The universe is 24 mega- and large-cap US names spanning tech, financials, healthcare, staples and energy — AAPL, MSFT, NVDA, JPM, UNH, XOM and the like. No shorting, no exotic sizing; just chase strength and limit concentration. It is currently flagged live.
What the backtest says
Over 451 days the strategy turned a starting book into $11,976 in final equity — a 19.76% total return, 10.6% CAGR. Sharpe came in at a modest 0.65, with a 20.48% max drawdown and turnover of 92.74%. Fees were trivial ($5 total, no FX cost). The headline return is respectable, but the Sharpe tells you it was a bumpy ride: you gave back roughly a fifth of equity at the worst point to earn that ~20%.
One caveat on the stats: winRate reads 0 across 5 trades. With only executed buys on record and positions still open, that figure looks like an artifact of no closed round-trips rather than five losers — treat it as uninformative, not damning.
Why validation failed
This is the interesting part. Walk-forward testing produced four folds, all four positive (+15.38%, +3.27%, +12.18%, +13.71%), with out-of-sample return of 13.71% and a punchy OOS Sharpe of 1.93. On the surface that is exactly what you want to see. Yet the gate reports passed: false.
The reason sits in the deflated metrics. PSR is a healthy 0.811, but the Deflated Sharpe Ratio is just 0.338 after accounting for 6 trials. In plain terms: once you penalise for how many variants were tried before this one, the evidence that the full-sample Sharpe of ~0.65 is genuinely above zero is weak. Four clean folds are encouraging, but the gate is right to be sceptical — a 0.65 Sharpe strategy selected from a search is precisely the profile that overfits.
Recent activity: silence
The live book has gone quiet. The last executed trades were a cluster of buys at the end of May and start of June 2026 — XOM, NVDA, BAC, HON, MSFT. Since then, every scheduled run from 19 to 26 October executed zero and rejected one to two orders, with cash frozen at $608.79 and total value drifting between ~$10,375 and ~$10,533. The pattern is consistent with a fully-invested book that cannot afford the next top mover: it wants to trade, the orders bounce, and the portfolio simply rides its existing positions.
Verdict
momentum-code is a plausible, cheap-to-run momentum sleeve with genuinely consistent fold-level returns. But two flags temper the enthusiasm: a failed validation gate driven by a weak DSR, and a live book that has been effectively idle for months on rejected orders. Before trusting it with more capital, the sizing/cash logic that keeps rejecting trades needs a look, and the overfitting risk the DSR is warning about deserves respect.