Strategy at a Glance
Momentum-code pursues a simple, well-trodden thesis: at each scheduled run, scan a 24-name universe of large-cap U.S. equities — spanning technology, financials, healthcare, consumer staples, energy, and industrials — and buy the day's top positive movers, subject to a per-position size cap. The universe is deliberately blue-chip, limiting the strategy to names with deep liquidity and broad analyst coverage.
As of 2026-07-07, the strategy holds five positions (XOM, NVDA, BAC, HON, MSFT) initiated in late May and early June, with $608.79 in idle cash. Total paper-portfolio value sits at $9,596.95.
Cross-Validation: Consistently Positive, Uneven in Conviction
The four-fold walk-forward covers August 2024 through May 2026:
| Fold | Period | Return | Sharpe | Max DD |
|---|---|---|---|---|
| 1 | Aug 2024 – Jan 2025 | +15.38% | 2.47 | 6.12% |
| 2 | Jan 2025 – Jul 2025 | +3.27% | 0.46 | 16.55% |
| 3 | Jul 2025 – Dec 2025 | +12.18% | 2.31 | 5.21% |
| 4 | Dec 2025 – May 2026 | +13.71% | 1.93 | 6.32% |
The headline is encouraging: all four folds are positive, and the out-of-sample period (Fold 4) delivered a 13.71% return with a Sharpe of 1.93. That's a meaningful result — the most recent, unseen data is the strategy's best Sharpe reading outside of Fold 1.
Fold 2, however, is the outlier that demands respect. A 16.55% drawdown on a 3.27% gain signals real stress during what appears to have been a difficult momentum environment. The strategy survived, but barely earned its risk budget.
Why the Validation Gate Is Still Closed
Despite a 19.76% full-period return and a 4-for-4 positive fold record, the automated validator has not cleared momentum-code for deployment. The culprit is the Deflated Sharpe Ratio (DSR): 0.338.
The DSR adjusts the observed Sharpe for the number of strategy trials tested (6 in this case) and the statistical properties of the return stream. A DSR of 0.338 means there is only a ~34% probability the true, out-of-sample Sharpe exceeds the benchmark — a direct consequence of testing multiple parameter combinations. The Probabilistic Sharpe Ratio (PSR) of 0.811 is more flattering, but the DSR's multiple-trial penalty tells a harder truth: some of the headline Sharpe is likely explained by selection bias across the 6 trials rather than genuine edge.
This is exactly the kind of false positive a validation gate is designed to catch.
Recent Live Activity: A Quiet Stretch
The six scheduled runs between June 29 and July 6 executed zero trades, with one or two rejections per session. The portfolio's cash position has held flat at $608.79 throughout, while total equity has drifted between $9,401 and $9,648 — passive mark-to-market movement on the five held positions rather than active management. The high turnover figure in the backtest (92.74%) suggests this quiet period is atypical; the strategy normally rotates aggressively when momentum signals trigger.
One metric worth flagging: the reported win rate of 0% across 5 trades likely reflects that all current positions remain open and unrealized — no closed trades to record a win or loss against. This is not a signal of poor performance; it is an accounting artifact of a buy-and-hold window.
Strengths and Risks
Strengths
- Consistent directional performance across all four folds, including the live-adjacent OOS period
- Low transaction cost footprint: $5 total fees, zero FX cost, only 5 trades over 451 days
- Blue-chip universe limits catastrophic single-name risk
Risks
- A 20.48% max drawdown over the full backtest is non-trivial for a strategy with a 10.6% CAGR — the pain-to-gain ratio deserves scrutiny
- Fold 2's drawdown shows the strategy can be caught badly wrong in choppy or mean-reverting regimes
- DSR of 0.338 is a structural concern: more parameter trials without a corresponding improvement in Sharpe will keep the gate closed
- Concentrated entry: all five positions were entered on two days (May 31, June 1), creating calendar-clustering risk
Bottom Line
Momentum-code has a genuine signal — four consecutive positive folds with a strong recent out-of-sample read is not noise. But the validation framework is doing its job by insisting on a higher bar before real capital follows. The path forward is narrowing the trial count, stress-testing Fold 2 conditions specifically, and watching whether the next live rotation maintains the quality seen in Folds 3 and 4.