Thesis
The mean-reversion strategy runs a textbook RSI rule: buy when a name is oversold (RSI below 30), sell when it is overbought (RSI above 70). It trades a 24-stock universe of large-cap US blue chips — think AAPL, MSFT, JPM, WMT, XOM — so the thesis is less about exotic edge and more about disciplined fading of short-term extremes in liquid names.
Backtest
On paper the numbers flatter it. Over 451 days the backtest returned 14.73% (7.98% CAGR), finishing at $11,473 on a $10k base, with a 70.59% win rate across 38 trades. Fees were a trivial $38 and FX cost nothing. The headline weakness is risk-adjusted return: a Sharpe of 0.58 is modest, and a 15.64% max drawdown is a lot of pain for an 8% annual return. Turnover of 879% also tells you this is an active book, not a buy-and-hold.
Validation — the part that matters
This is where enthusiasm should cool. The walk-forward validation did not pass. Across four folds, three were positive but the out-of-sample segment returned -2.84% with a -0.33 Sharpe — the strategy lost money in the most recent, unseen window (Dec 2025–May 2026), which also carried the worst fold drawdown at 14.96%. The Deflated Sharpe Ratio sits at just 0.304 against a Probabilistic Sharpe of 0.785, and the result was selected from six trials. In plain terms: the in-sample Sharpe does not survive correction for how many variants were tried, and the edge decayed exactly where we most want it to hold.
Recent activity
The last executed trades were buys in September — MCD, HD, BAC — plus a WMT position from May. Since then, nothing. The six scheduled runs from 8–15 October all report 0 executed, 0 rejected, with cash pinned at $329.71 and total equity drifting between roughly $9,620 and $9,868. So the live book is effectively fully deployed and idle: no names are hitting RSI<30 to buy, and held positions haven't reached RSI>70 to sell. Live equity is currently below the $10k mark, a reminder that the backtest's gains are not being reproduced in real time.
Verdict
Strengths are real — a high win rate, cheap execution, and a simple, explainable rule on liquid names. But the risks dominate the story: a failed validation, negative out-of-sample performance, a thin deflated Sharpe, and a live account that is both underwater versus its start and currently inert. The 70% win rate is seductive and, in this case, misleading — it coexists with a drawdown and an OOS loss that suggest the average winner is small and the occasional loser is large. We would treat this as a paper-stage strategy under review, not a validated edge.