The thesis
News-sentiment is a simple, legible idea: buy names when recent news sentiment turns positive and exit when it turns negative. Its universe is 24 large-cap US blue chips — AAPL, MSFT, NVDA, JPM, V, JNJ, KO, XOM and peers — so it never lacks coverage or liquidity. The Reviewer approved it in May with a risk score of 0, flagging it as a built-in, hand-written strategy rather than an LLM-authored one.
Recent activity
The strategy has gone live and traded lightly. On 29 October it opened three positions in a single run — NVDA (9 @ $213.49), V (4 @ $410.29) and PFE (66 @ $29.92) — then sat on its hands through 30 October and 2 November with zero executions. On 3 November it added KO (21 @ $95.13). As of that run it held $2,387.18 in cash against a total of $9,915.73, leaving it fractionally below its $10k starting line. The cadence fits the thesis: it acts only when sentiment crosses a threshold, and lately it has mostly stayed in cash.
Backtest and validation
Here the picture gets uncomfortable. Over a 451-day backtest the strategy returned +0.19% (final equity $10,018.55, ~0.1% CAGR) on just two trades, with a 0.74 Sharpe and a trivial 0.05% max drawdown — figures that largely reflect a portfolio parked in cash. Win rate reads 0%, so neither recorded trade closed as a winner; the small positive return is doing a lot of quiet lifting.
Validation did not pass. Of four walk-forward folds, three (Aug 2024 through Dec 2025) produced no trades at all — the rule simply never triggered. The entire track record comes from fold 4 (Dec 2025–May 2026): +0.19%, a 1.5 Sharpe, two trades. So "1 of 4 folds positive" is really "1 of 4 folds active." The probabilistic Sharpe is a flattering 0.923, but the deflated Sharpe — which penalises the six trials run — slumps to 0.551, barely above a coin toss.
Verdict
The strength is discipline: low turnover (36.5%), negligible drawdown, and a clear, human-readable rule set applied to a high-quality universe. The risk is that we have almost no evidence it works. Two trades across one of four windows is not a track record; it is a single data point wearing a track record's clothes. The dormancy in three of four folds suggests the sentiment trigger is either too strict or starved of signal for long stretches.
Live trading will teach us more than this backtest did — but position sizing should stay modest until fold coverage and trade count fill out. Promising shape, unproven edge.