The thesis
The premise is about as clean as a signal gets: buy on positive recent news sentiment, exit on negative. Applied across a 24-name universe of large-cap staples — the mega-cap tech complex (AAPL, MSFT, GOOGL, NVDA), financials (JPM, BAC, V, MA), healthcare (JNJ, UNH, PFE, ABBV) and consumer defensives (PG, KO, WMT, COST) — it reads like a momentum overlay driven by the news tape rather than price. That simplicity is a genuine strength: the logic is auditable, and a sentiment flip is an interpretable reason to be in or out of a name.
Recent activity
In practice, the strategy has been remarkably quiet. The last six scheduled runs (2026-09-28 through 2026-10-05) executed zero trades, with a single rejection on Sept 28. Cash has sat frozen at $2,036.71 the entire week while the paper book drifted on mark-to-market alone, closing Oct 5 at a total value of $10,345.48.
The real activity is further back. On 2026-09-23 the book sold its entire BAC position — 35 shares at $55.96, against a June 12 entry at $55.93, essentially a round-trip to break-even — and rotated into 4 shares of NVDA at $225.34. Earlier executed buys include GOOGL, UNH, MSFT and AAPL through mid-2026. So the engine does act when sentiment turns; it simply hasn't seen a qualifying signal lately.
Backtest and validation
Here the picture turns sober. Over 451 days the backtest returned just +0.19% (final equity $10,018.55), a 0.1% CAGR, on only 2 trades, a Sharpe of 0.74 and a tiny 0.05% max drawdown. Turnover was 36.53% and total fees $2. In plain terms: the strategy was almost entirely in cash and barely participated.
That shows up starkly in the walk-forward folds. Of four folds, three generated zero trades and zero return — all performance came from fold 4 (Dec 2025–May 2026): +0.19%, Sharpe 1.5, over 2 trades. So only 1 of 4 folds was positive, and the entire track record rests on a single window.
The headline validation stats look seductive — out-of-sample Sharpe 1.5, PSR 0.923 — but the overall gate failed, and rightly so. The Deflated Sharpe Ratio, which penalises the 6 trials run, drops to 0.551, and a result concentrated in one fold is the textbook signature of a strategy with too little data to trust.
The verdict
Strengths: an honest, legible thesis and tight drawdown control. Risks: the sample is simply too thin to conclude anything. Two trades in a backtest and a week of dormant runs tell us the signal fires rarely — and when the evidence lives almost entirely in one time window, a flattering OOS Sharpe is noise dressed as skill. This one needs more trades on the board before it earns conviction.