The thesis
The news-sentiment strategy runs on one of the most intuitive ideas in the book: buy when recent news sentiment turns positive, exit when it turns negative. It trades a 24-name large-cap universe spanning tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), healthcare (JNJ, PFE), and staples (KO, PG, WMT). It is a hand-written, reviewer-approved strategy — approved at risk level 0 back on 2026-05-31 — and it is currently live.
Recent activity
The live book came alive on 2026-10-29, when the scheduled run executed three buys and rejected none: PFE (66 shares @ $29.92), NVDA (9 @ $213.49), and V (4 @ $410.29). That left roughly $4,405 in cash against a total of $9,942.53. The very next day the strategy did nothing — zero executed, zero rejected — and the book drifted down to $9,914.81. Two sessions in, the pattern is already clear: this strategy acts in short bursts and then waits.
Backtest and validation
Over a 451-day backtest the strategy returned just +0.19%, ending at $10,018.55 — a CAGR of about 0.1%. The risk profile looks clean on paper: a Sharpe of 0.74, a maximum drawdown of only 0.05%, modest 36.5% turnover, and a trivial $2 in fees with no FX cost. But those numbers rest on just two trades, with a win rate of 0%.
Validation tells the real story, and it did not pass. Across four walk-forward folds, only one was positive — fold 4 (Dec 2025–May 2026), which posted +0.19% and a 1.5 Sharpe on two trades. Folds 1 through 3 each produced zero trades and zero return: the sentiment signal simply never triggered. The probabilistic Sharpe (PSR) reads a flattering 0.923, but the deflated Sharpe (DSR), which accounts for the six trials run, drops to 0.551.
Strengths and risks
Strengths: tiny drawdown, negligible costs, low turnover, and a genuinely respectable out-of-sample Sharpe where it did trade. It is cheap to run and well-behaved.
Risks: the headline stats are an artifact of almost no activity. A 0% win rate, two backtest trades, and three of four folds with no fills mean we have almost no evidence the edge is real — which is exactly why validation flagged it. The signal is so selective it may be miscalibrated or starved of qualifying news.
Verdict: promising premise, unproven in practice. Watch whether the live burst of three buys actually converts into realized gains before reading anything into that clean Sharpe.