The thesis
Mean-reversion is one of the oldest ideas in systematic trading, and this strategy keeps it deliberately plain: buy when RSI falls below 30 (oversold), sell when RSI climbs above 70 (overbought). It runs across a 24-name large-cap universe — the usual megacap suspects (AAPL, MSFT, NVDA), financials (JPM, BAC, V, MA), healthcare (JNJ, UNH, PFE), and staples/industrials (PG, KO, WMT, CAT, XOM). It is a built-in, hand-written strategy, approved by the reviewer at risk 0, and is now flagged live.
Recent activity
The scheduled run on 2026-10-29 executed a single order and rejected none: a buy of 48 shares of BAC at $51.15 — a textbook oversold entry. After that run the book stood at $7,521.07 cash and $9,976.27 total. That total is the number to sit with: live equity is currently a hair below the $10,000 starting line, so the strategy has yet to prove itself with real (paper) money. One trade is not a trend, but it is an honest starting point.
Backtest performance
Over 451 days the backtest returned 14.73% (final equity $11,473.17), a 7.98% CAGR, with a 70.59% win rate across 38 trades. On paper that is attractive — most trades worked. But two numbers temper the enthusiasm: a Sharpe of only 0.58 and a 15.64% max drawdown. A high win rate paired with a middling Sharpe and a double-digit drawdown is the classic mean-reversion signature — many small wins, punctuated by painful losers. Turnover was heavy at 879%, though fees stayed trivial ($38 total, no FX cost).
Validation — where it stumbles
This is the part that matters, and it is unflattering: validation did not pass. Walk-forward testing split the history into four folds; three were positive, but the most recent fold (2025-12-16 → 2026-05-29) lost 2.84% with a negative Sharpe of -0.33 and a 14.96% drawdown. The out-of-sample read mirrors that single weak fold almost exactly (OOS return -2.84%, OOS Sharpe -0.33), suggesting the headline 14.73% leans on the strong middle period (fold 2: +11.1%, Sharpe 1.32) rather than consistent edge.
The deflated statistics agree. Against 6 trials, the Probabilistic Sharpe Ratio sits at 0.785 — decent — but the Deflated Sharpe Ratio collapses to 0.304, below any reasonable confidence bar. In plain terms: once you account for how many variants were tried, the evidence that this strategy's Sharpe is genuinely above zero is weak.
Verdict
Mean-reversion is a sound, well-understood idea executed cleanly, and its high win rate and low fees are real strengths. But the deteriorating final fold and a sub-threshold Deflated Sharpe are warnings, not footnotes — the recent regime has not rewarded buying the dip in these names. Running it live at small size is a reasonable way to gather honest out-of-sample evidence; scaling it up before that evidence arrives would be a mistake. Watch the live Sharpe and the drawdown, not the win rate.