The thesis
The mean-reversion strategy runs a textbook contrarian rule across a 24-name large-cap universe (AAPL, MSFT, NVDA, JPM, V, JNJ, XOM and friends): buy when RSI falls below 30, sell when it climbs above 70. The bet is that short-term price extremes in liquid blue chips tend to snap back. It is currently live.
Backtest performance
Over 451 days the strategy returned 14.73% (final equity $11,473.17), a CAGR of 7.98%, with a Sharpe of 0.58 and a maximum drawdown of 15.64%. The win rate is a flattering 70.59% across 38 trades. Fees were negligible ($38 total, no FX cost), and turnover was high at 879% — this is an active book, not a buy-and-hold.
The headline is a reminder worth repeating: a 70% win rate and a sub-0.6 Sharpe can coexist. Lots of small wins, punctuated by larger losers, is a classic mean-reversion signature, and the 15.64% drawdown shows the losing side has teeth.
Validation: the warning light
This is where the picture dims. Cross-validation did not pass. Across four walk-forward folds, three were positive — but the most recent fold (2025-12-16 to 2026-05-29) returned -2.84% with a Sharpe of -0.33 and a 14.96% drawdown. The out-of-sample return therefore lands at -2.84% with a negative OOS Sharpe.
The deflated metrics tell the same story. The Probabilistic Sharpe Ratio is a respectable 0.785, but the Deflated Sharpe Ratio is just 0.304 after accounting for 6 trials. A DSR that low means we cannot confidently reject the possibility that the backtest edge is a product of selection, not skill. The deteriorating fold sequence — strong in early 2025, weak into 2026 — suggests the edge may be regime-dependent.
Recent activity
Live behaviour has been quiet. Of the last six scheduled runs, five executed zero trades; only 2026-11-06 fired a single order — a 13-share NVDA buy at $188.24. Earlier in the window it added CAT (4 @ $622.54), GOOGL (8 @ $280.28) and BAC (48 @ $51.15) on 29-30 October. That NVDA purchase deployed most of the dry powder, dropping cash from $2,742 to $271.67.
Total equity has hovered in the $9,600-$9,870 band and sat at $9,719.55 on 2026-11-09 — modestly underwater, consistent with the soft out-of-sample read rather than the rosy backtest.
Verdict
Strengths: a clean, well-understood rule; high win rate; trivial costs; a diversified, liquid universe. Risks: the validation failure is not a footnote. A negative latest fold, a DSR of 0.30, and a live account treading water all point the same direction. The strategy is worth keeping on the live leaderboard as a learning case, but its backtest return should be read as an optimistic ceiling, not an expectation.