Thesis
The news-sentiment strategy runs on a simple, legible premise: buy on positive recent news sentiment, exit on negative. It operates over a 24-name large-cap universe spanning tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, BAC, V, MA), healthcare (JNJ, UNH, PFE, ABBV), staples (PG, KO, WMT, COST) and industrials/energy (CAT, HON, XOM, CVX). The appeal is interpretability — every position maps to a readable narrative — and the discipline of a symmetric entry/exit rule. The risk, equally, is that news sentiment is noisy, already priced in fast, and prone to whipsaw.
Recent activity
The strategy is live and currently sits at a total paper balance of $10,310.82 with $990.06 in cash as of 2026-08-25. Notably, it has gone quiet: the last six scheduled runs (2026-08-18 through 2026-08-25) executed zero trades, with one or two orders rejected on most days. The real activity was earlier — a cluster of executed buys in late May through early July: AAPL (6 @ $312.06), MSFT (4 @ $428.23), BAC (35 @ $55.93), UNH (4 @ $402.85) and GOOGL (5 @ $360.40). So the current book is a set of held longs, with the sentiment signal not firing fresh entries lately and rejecting the ones it does surface.
Backtest and validation
Here the picture turns cautious. Over 451 days the backtest returned just +0.19% (final equity $10,018.55), a CAGR of 0.1%, Sharpe 0.74, and a shallow max drawdown of 0.05% — but on only 2 trades with turnover of 36.53% and $2 in fees. A win rate of 0% on two trades is statistically meaningless, yet it underlines how thin the sample is. Essentially flat returns on almost no activity is not a track record; it's a placeholder.
Cross-validation confirms the concern. The strategy failed walk-forward validation: of 4 folds, only 1 was positive, and — tellingly — three of four folds produced zero trades. All the performance (0.19% return, 1.5 Sharpe, 2 trades) is concentrated in the final fold (2025-12-16 → 2026-05-29). The out-of-sample Sharpe of 1.5 looks flattering in isolation, but it rests entirely on that one window. The probabilistic metrics are mixed: PSR of 0.923 is encouraging, yet the deflated Sharpe (DSR 0.551), which penalises the 6 trials run, is far more sober.
Verdict
Strengths: a clear, explainable thesis; tight drawdowns; and a positive live paper balance so far. Risks: the signal trades rarely, its edge is unproven across time, and its entire backtest rests on a single fold — exactly the overfitting-to-one-regime trap validation exists to catch. The honest read is promising premise, insufficient evidence. Until it demonstrates activity and positive returns across multiple folds — not one — news-sentiment belongs in observation, not capital allocation.