Thesis
momentum-code runs one of the oldest ideas in systematic trading: buy the strongest recent movers and let winners run, with a per-position cap to avoid concentration. Its universe is 24 US large-caps spanning tech, financials, healthcare, staples and energy — AAPL, MSFT, NVDA, JPM, V, JNJ, XOM and peers. It is a simple, transparent rule, which is exactly what makes it a useful benchmark for the more elaborate agents in the lab.
Backtest and validation
On the full sample the strategy returned 19.76% (final equity $11,976 on a $10k base), a 10.6% CAGR over 451 days, with a Sharpe of 0.65 and a max drawdown of 20.48%. Turnover was moderate at 92.74% and fees negligible ($5). The headline number is respectable; the Sharpe and the one-fifth drawdown are the honest counterweights.
The walk-forward picture 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 and single-digit drawdowns in three of four windows. Out-of-sample return averaged 13.71% with an OOS Sharpe of 1.93, meaningfully better than the aggregate. That consistency is the strongest thing on the sheet.
And yet validation did not pass. The probabilistic Sharpe (PSR) is a healthy 0.811, but the deflated Sharpe (DSR) is only 0.338 across 6 trials. In plain terms: once you penalise for how many variants were tested to arrive here, the edge is no longer statistically convincing. The gate is doing its job — refusing to certify a result that could be selection luck dressed up as skill.
Recent activity
The live book tells a cautionary story. The five executed trades — buys in XOM, NVDA, BAC, HON and MSFT — all fired at the end of May and start of June 2026, and none have been closed. That is why the reported win rate is 0%: there are no completed round-trips to score, not a run of losers. Read it as "not yet measurable," not "never wins."
Since mid-August, every scheduled run has executed zero trades and logged rejections (one or two per day). With cash pinned at $608.79, the strategy is effectively fully deployed and cannot act on new signals. Meanwhile total equity has drifted down from $10,804 (Aug 14) to $10,500 (Aug 21) — roughly a 3% slide in a week while it sits unable to rotate.
Verdict
Strengths: a clean, interpretable thesis and four-for-four positive folds with strong OOS Sharpe. Risks: a 20% historical drawdown, a full-sample Sharpe under 0.7, a failed deflated-Sharpe gate, and a live book that is capital-constrained and drifting lower. The signal quality looks real; the statistical confidence and the execution plumbing do not yet justify scaling. Keep it live as a monitored benchmark, but let the validation gate — not the pretty fold table — govern any capital decision.