The thesis
News-sentiment runs a deliberately simple rule: buy names showing positive recent news sentiment, exit when sentiment turns negative. It fishes in a 24-stock pond of mega-cap and defensive blue chips — AAPL, MSFT, GOOGL, NVDA, the big banks and payment networks, healthcare (JNJ, UNH, PFE, ABBV), staples (PG, KO, WMT, COST), and a handful of industrials and energy names. It is currently flagged live.
The appeal is clarity. The risk, as always with sentiment, is that headlines are noisy, crowded, and already priced. The data below suggests both the discipline and the fragility of that bet.
Recent activity
Live execution has been quiet to the point of inertia. Across the scheduled runs from 19–23 October, every session reports 0 executed, 1 rejected — an order is being generated each day and then bounced, worth an operational look rather than a shrug. The 16 October run was clean (0 executed, 0 rejected). Cash has sat frozen at $2,036.71 throughout, while total account value drifted between roughly $10,337 and $10,453 — modestly above a $10k baseline, but moving with the book it already holds, not with new decisions.
The last trades that did clear tell a tidier story: an AAPL buy (31 May), MSFT and BAC adds in June, a UNH position, a GOOGL buy on 1 July, and on 23 September a rotation — selling the full 35-share BAC lot at $55.96 and buying 4 NVDA at $225.34. That is a strategy that acts in bursts, then waits.
Backtest and validation
Here is where enthusiasm should cool. The backtest covers 451 days and produces a 0.19% total return (CAGR ~0.1%), a Sharpe of 0.74, and an almost invisible 0.05% max drawdown — on just 2 trades with a 0% win rate and $2 in fees. Capital preservation, yes; conviction, no.
The walk-forward validation did not pass. Of four folds, only one was positive — and folds 1–3 placed zero trades at all. The entire result lives in fold 4 (0.19% return, Sharpe 1.5, 2 trades). The headline PSR of 0.923 looks reassuring until you read the deflated Sharpe: DSR 0.551 across 6 trials, barely better than a coin flip once multiple-testing is accounted for.
Verdict
Strengths: tiny drawdowns, near-zero costs, a transparent rule, and a respectable 1.5 Sharpe in its one active window. Risks: the evidence is desperately thin — two trades, three empty folds, a failed gate, and a live book that keeps getting its orders rejected. Treat this as a prototype earning its keep on discipline, not yet on proof.