Thesis
momentum-code runs a simple, legible idea: buy the top positive movers across a 24-name large-cap universe — the usual megacap tech, financials, staples, and energy — with a per-position cap to prevent any single winner from dominating. There is no factor soup here, just relative strength with position sizing. That simplicity is a genuine asset: fewer knobs mean fewer ways to overfit, and the strategy currently sits in live status.
Backtest and validation
Over 451 days the backtest returned 19.76% (final equity $11,976 on a $10k base), a 10.6% CAGR, with a Sharpe of 0.65 and a max drawdown of 20.48%. Turnover was moderate at ~93% and total fees a trivial $5. The headline number is decent; the risk-adjusted number is less flattering — a sub-1 Sharpe against a 20%+ drawdown means the ride was not smooth.
The walk-forward validation is where the picture gets more interesting. All four folds were positive (15.38%, 3.27%, 12.18%, 13.71%), and out-of-sample the strategy earned 13.71% at a 1.93 Sharpe — notably better risk-adjusted performance than the full-sample figure. The Probabilistic Sharpe Ratio of 0.811 is healthy. Consistency across regimes is the standout strength here.
And yet the gate reports passed: false. The likely culprits are the Deflated Sharpe Ratio of just 0.338 (adjusted for 6 trials, the edge looks statistically thin) and the fact that fold 2 nearly stalled at a 0.46 Sharpe with a 16.55% drawdown. The strategy is robust in direction but fragile in magnitude — it works, but the evidence that it reliably works is weaker than the raw returns suggest.
Live activity
Here the story turns cautionary. The last executed trades were back on May 31–June 1 (buys in MSFT, HON, BAC, NVDA, XOM). Since then, every scheduled run from Aug 28 through Sep 4 has reported 0 executed and 1–3 rejected orders, with cash pinned at $608.79 and total equity drifting between $10,515 and $10,671.
A week-plus of rejected-only runs is a red flag worth investigating. With only ~$609 in cash and a per-position cap, the strategy may simply lack the buying power to act on its signals — orders are generated but bounce off sizing or cash constraints. The portfolio is effectively frozen in its June positions.
Verdict
momentum-code has a clean thesis and genuinely consistent out-of-sample behaviour — four positive folds and a 1.93 OOS Sharpe are not accidents. But the failed validation gate (thin DSR) and, more urgently, a stalled live engine temper the enthusiasm. Before trusting the equity curve, someone should resolve why every recent order is being rejected. A strategy that can't trade can't compound.