Thesis
The news-sentiment strategy is simple by design: scan recent headlines across a 24-ticker universe of large-cap US equities, buy into names showing positive sentiment momentum, and exit when the sentiment turns negative. No earnings models, no macro overlays — just the signal embedded in the news cycle.
The universe spans mega-cap tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, BAC, V, MA), healthcare (JNJ, UNH, PFE, ABBV), consumer staples and discretionary (PG, KO, WMT, COST, MCD, NKE, HD), energy (XOM, CVX), and industrials/media (CAT, HON, DIS). Coverage is intentionally broad to let the agent find signal wherever the news cycle happens to be running hot.
Recent Activity
The strategy has been running daily scheduled sweeps since at least early June. Of the six most recent runs logged, only one produced a trade: a 4-share buy of MSFT at $428.23 on June 3. Combined with a 6-share AAPL buy at $312.06 on May 31, those are the only two executed trades in the strategy's entire live history.
As of June 9, the portfolio holds approximately $6,415 in cash against a total value of $9,779 — meaning a meaningful portion of capital remains undeployed. The total value has drifted slightly downward over the week ($10,016 on June 2 → $9,779 on June 9), which tracks normal mark-to-market movement in the two open positions rather than any new activity.
The strategy is running, but it is mostly watching.
Backtest & Validation
Over 451 days of backtested history the strategy produced a +0.19% total return on two trades, a 0.74 Sharpe, and a shallow 5% maximum drawdown. Annualised CAGR sits at 0.10%, fees were negligible at $2 total, and there were no FX costs.
The walk-forward validation is where the picture gets complicated. Across four sequential folds covering August 2024 through May 2026:
| Fold | Period | Return | Trades |
|---|---|---|---|
| 1 | Aug 2024 – Jan 2025 | 0% | 0 |
| 2 | Jan 2025 – Jul 2025 | 0% | 0 |
| 3 | Jul 2025 – Dec 2025 | 0% | 0 |
| 4 | Dec 2025 – May 2026 | +0.19% | 2 |
Only 1 of 4 folds was positive — and that sole positive fold accounts for the entire return. The Probabilistic Sharpe Ratio (PSR) of 0.923 is encouraging in isolation, suggesting the observed Sharpe is likely above zero. However, the Deflated Sharpe Ratio (DSR) of 0.551 — which adjusts for the number of trials tested (6 in this case) — tells a more cautious story: once you account for the search over multiple configurations, the edge is harder to confirm.
The validation status is formally not passed.
Strengths
- Low drawdown: a 5% max drawdown is genuinely conservative and suggests the signal, when it fires, isn't chasing crowded momentum.
- Out-of-sample Sharpe of 1.5 in fold 4 is an attractive number, and the PSR implies some real signal is present.
- Minimal friction: two total trades, $2 in fees, zero FX cost — the strategy isn't grinding capital away in churn (36.5% turnover is moderate).
Risks
- Three dormant folds out of four is a red flag for regime-dependence. The strategy may only activate — and only work — under specific news-cycle conditions that happened to coincide with fold 4.
- Win rate is recorded as 0%, which is an artefact of only holding open positions (no closed winning trades yet), but it means there is no completed-trade track record to evaluate.
- Concentration of signal: with $6,415 sitting in cash and daily runs consistently returning zero executions, the strategy is either very selective or the news-sentiment threshold is rarely met.
Outlook
News-sentiment is a coherent thesis with a plausible mechanism. The current evidence base is thin — two trades, one productive fold — and the DSR flags that the observed results may not yet clear the multiple-comparison bar. The next meaningful test is whether fold 4's conditions repeat: if the strategy begins firing regularly and those trades close profitably, the validation picture improves materially. For now, it warrants continued paper trading with close attention to what news conditions triggered the two executions.