Thesis
The news-sentiment strategy is straightforward by design: scan recent headlines across a 24-name universe of blue-chip and mega-cap equities, enter long positions when aggregate sentiment turns positive, and exit when it sours. The universe spans technology (AAPL, MSFT, NVDA, GOOGL), 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 (CAT, HON, DIS) — a reasonable cross-section of the S&P 500's bellwethers.
The appeal is intuitive: news moves markets, and being early to a sentiment shift should produce alpha. The execution is harder.
Recent Activity
Four trades were placed between late May and mid-June 2026 — buys in AAPL (6 shares @ $312.06, May 31), MSFT (4 @ $428.23, Jun 3), BAC (35 @ $55.93, Jun 12), and UNH (4 @ $402.85, Jun 18). Since then, the agent has run daily and found nothing to act on: six consecutive scheduled runs through June 26 returned zero executions, zero rejections.
Portfolio value has drifted between $9,606 and $9,748 over that same window, with $2,810 sitting idle in cash. The positions are moving with the market — not against it — but the strategy is essentially in standby mode.
Backtest & Validation
Over 451 days the backtest produced a +0.19% total return (annualised CAGR of ~0.1%), a Sharpe ratio of 0.74, and a maximum drawdown of only 5%. Total trading costs came to $2 — a reflection of how rarely the strategy fires, with just two completed trades in the full backtest window.
The cross-validation picture is more sobering. Across four time folds spanning August 2024 to May 2026, only the final fold generated any activity at all. Folds 1–3 (covering roughly 16 months) recorded zero trades, zero return, and zero drawdown — the signal simply never triggered. Fold 4 (December 2025–May 2026) did produce two trades, a 0.19% return, and a 1.5 Sharpe, which is respectable in isolation. But with three of four folds completely dormant, validation correctly failed: one positive fold out of four does not constitute a pattern.
The Probabilistic Sharpe Ratio sits at 0.923 — reasonable — but the Deflated Sharpe Ratio drops to 0.551 once adjusted for six backtest trials, a reminder that cherry-picked performance windows overstate confidence.
Strengths
- Low drawdown. A 5% maximum drawdown over 15 months is genuinely controlled risk, even if that low figure partly reflects inactivity.
- Clean signal when it fires. The single active fold produced a 1.5 Sharpe, suggesting the underlying logic has some discriminating power when news conditions align.
- Negligible costs. Two dollars in fees over 451 days means slippage and friction are not eating the edge.
Risks & Watch Points
- Signal drought. A strategy that doesn't trade for three of four six-month periods is effectively a cash account most of the time. Opportunity cost accumulates silently.
- Recency concentration. All observed performance lives in one recent fold. That could reflect improving news-API coverage, a noisier news cycle in late 2025–2026, or plain luck. It is too early to distinguish.
- Win rate undefined. With no completed round-trips (all current positions are buys without a recorded exit), the win rate remains at 0% — not zero wins, but zero data. That will change once positions are closed.
- Universe breadth vs. signal density trade-off. Twenty-four names is wide enough to catch sector rotations but may dilute the sharpest sentiment edges; a tighter, higher-conviction universe is worth exploring.
Bottom Line
News-sentiment is a live strategy with a coherent thesis, well-controlled downside, and an encouraging recent fold — but it has not yet earned validation. The immediate priority should be understanding why the signal stayed silent for 16 months: was news coverage sparse in those folds, were thresholds miscalibrated, or was the market simply not moving on sentiment? The answer will determine whether this is a timing issue or a structural one.