RSI + SMA-200 + ADX Filter: Passed, But Zero Live Trades

We tested a mean-reversion strategy that buys when RSI(14) dips below 35 while price stays above the 200-period moving average, filtering for low-ADX regimes (ADX below 30) to avoid strong trends. It passed all six gates with a 0.489% average P&L per trade, 87.5% win rate in both training and validation, and a deflated Sharpe ratio of 2.76 (a measure that accounts for the number of tests we ran and adjusts for multiple-testing bias). The strategy graduated to our survivor library.

Here's the caveat: it has logged zero live trades in forward testing. None. The backtest covered BTC, ETH, SOL, and XRP on 15-minute bars, but real-time conditions since graduation have not triggered a single entry. That could mean the regime filter is too restrictive, the market has been persistently high-ADX since we deployed it, or the specific RSI-below-35 plus above-SMA-200 combo is rarer than the backtest window suggested.

We're also flagging that the reported probability of being a fluke is 100.0%, which appears to be a data artifact or reporting bug. The deflated Sharpe of 2.76 suggests statistical significance, but that 100% fluke probability number contradicts it. We're investigating the discrepancy and will update the registry when resolved.

The strategy makes intuitive sense: buy oversold dips in coins that are still above long-term support, but only when the market is ranging rather than trending hard. The ADX filter is designed to keep us out of momentum regimes where mean-reversion gets steamrolled. The 87.5% win rate in validation is impressive, but with zero forward trades we have no evidence yet that it works outside the backtest window.

This one stays in the survivor library because it passed our gates, but we're treating it as unproven until it logs at least a dozen live trades. If you want to explore regime-aware mean-reversion ideas yourself, head to /prove and test your own ADX or volatility filters. Or browse the full survivor registry at /survivors to compare coverage and forward-trade counts across strategies.