Thesis and Approach
momentum-code follows a straightforward premise: at each rebalance, buy the top positive price movers within a curated 24-stock universe spanning mega-cap tech, financials, healthcare, consumer staples, energy, and industrials. Position sizes are capped to limit concentration. The logic is deliberate in its simplicity — momentum as a factor has decades of academic support, and constraining the universe to liquid large-caps keeps execution risk low.
Backtest Overview
Over 451 calendar days (covering roughly August 2024 through late May 2026), the strategy returned 19.76%, growing a hypothetical $10,000 to $11,976. The annualised CAGR of 10.6% is respectable for a long-only large-cap strategy, though the overall Sharpe of 0.65 and a 20.48% maximum drawdown suggest the ride was not entirely smooth. Turnover came in at 92.74% — high, but expected from a momentum rotator — and total fees were a nominal $5 across 5 backtest trades.
One number to flag: the reported win rate is 0%. With only 5 backtest trades logged, this metric carries almost no statistical weight and should not be read as a genuine signal about trade quality.
Walk-Forward Validation
The four-fold walk-forward tells a more nuanced story:
| 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% |
All four folds are positive — a clean result. The out-of-sample return of 13.71% with a Sharpe of 1.93 (Fold 4, the true holdout period) is the most credible data point here and is genuinely encouraging. Folds 1 and 3 also show strong risk-adjusted returns.
Fold 2 is the outlier: a 16.55% drawdown and Sharpe of just 0.46. That period (January–July 2025) coincided with elevated macro uncertainty, and momentum strategies are notoriously exposed to sharp reversals. It's a realistic stress-test, not an anomaly to dismiss.
Why Validation Did Not Pass
Despite the positive fold record, the strategy did not clear the platform's validation gate. The key constraint is the Deflated Sharpe Ratio (DSR) of 0.338 — well below the 0.50 threshold typically required. The DSR penalises strategies tested across multiple parameter trials (6 here) to correct for selection bias. With a Probabilistic Sharpe Ratio (PSR) of 0.811 and an in-sample Sharpe of 0.648, the raw numbers look reasonable — but the trial-count penalty pulls the DSR down sharply. This is the platform doing its job: a hand-written strategy that was evaluated six times before deployment carries meaningful overfitting risk, and the DSR surfaces that.
Live Activity
momentum-code went live on 31 May 2026 with a $10,000 deployment. Its opening trades were MSFT (4 shares @ $450.24), HON (8 @ $237.86), and BAC (38 @ $51.60). On 1 June it added XOM (13 @ $148.69) and NVDA (8 @ $224.20). Since then, the daily scheduled runs on 2 and 3 June executed zero trades and rejected several signals — consistent with a momentum filter that only acts when candidates clear the positive-mover threshold. Portfolio value has drifted from $10,033 on 1 June to $9,850 on 3 June, a modest short-term dip as the initial positions settle.
Risk Considerations
- Momentum crashes: factor reversal events (e.g., Fold 2) can be steep and fast.
- Concentration: five holdings across a 24-stock universe means idiosyncratic risk is real.
- Low trade count: both backtest and live history are thin; statistical conclusions are premature.
- DSR overfitting flag: the validation failure is a genuine caution, not a technicality.
The out-of-sample Fold 4 result is the most honest signal available. Watch it across the next two to three rebalance cycles before drawing conclusions about edge durability.