The thesis
momentum-code is a plain-spoken momentum strategy: buy the top positive movers across a 24-name large-cap universe, capped per position. That universe spans mega-cap tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), staples (PG, KO, WMT) and energy (XOM, CVX) — a broad, liquid cross-section. It is a built-in, hand-written strategy, approved by the reviewer at risk 0, and it is currently live.
Recent activity
The strategy put capital to work on 29–30 October: three buys on the 29th (NKE 52 sh @ $38.38, HD 6 @ $328.57, MA 3 @ $636.60) and two more on the 30th (WMT 16 @ $123.53, XOM 10 @ $186.99). That spent it down to $177.08 in cash, and the consequence shows immediately in the log. Every scheduled run since — 2, 3 and 4 November — executed zero trades and rejected two. With only $177 left, momentum-code simply cannot act on fresh signals: it is fully invested and stuck. Total equity has drifted between $9,875 and $10,186 across those runs, essentially flat-to-marginally-up against its $10k start.
That cash-management gap is the most concrete risk on display. A momentum book that cannot rotate into new leaders is momentum in name only.
Backtest and validation
On paper the record is attractive: +19.76% total return over 451 days (10.6% CAGR), a 20.48% max drawdown, and remarkably low costs ($5 in fees, zero FX). More encouraging still, the walk-forward held up — all four out-of-sample folds were positive (15.38%, 3.27%, 12.18%, 13.71%), with an out-of-sample Sharpe of 1.93 on the final fold and a healthy PSR of 0.811.
And yet the validation verdict is passed: false. The tell is the deflated Sharpe ratio of 0.338: once you penalise the headline 0.65 Sharpe for the six trials run, confidence that the edge is real drops sharply. The weak middle fold (3.27% return, 0.46 Sharpe, 16.55% drawdown) shows the strategy's dependence on trending regimes. A reported win rate of 0 across just five trades also warrants caution — more likely a measurement quirk of so few closed round-trips than a literal zero, but it underlines how thin the trade sample is.
Verdict
momentum-code pairs a clean, cheap, consistently-positive backtest with two honest caveats: an automated gate it did not clear, and a live book that has painted itself into a corner on cash. The thesis is sound and the out-of-sample folds are genuinely reassuring; the discipline around position sizing and reinvestment is where the work remains.