Ranging Low-Vol BB+RSI Oversold on 15-Minute Bars

We tested a mean-reversion strategy that buys when price touches the lower Bollinger Band and RSI confirms oversold conditions, but only when ADX signals a ranging market. It graduated with 1.7% average profit per trade, a 90% win rate, and a deflated Sharpe ratio (a metric that corrects for multiple testing and overfitting) of 23.41. The probability this is a statistical fluke rounds to zero.

The strategy runs on 15-minute bars across BTC, ETH, SOL, XRP, DOGE, BNB, ADA, AVAX, LINK, and LTC. Entry requires price to close below the 20-period lower Bollinger Band and RSI(14) to drop below 30. The regime filter is the critical piece: ADX must stay below 50, killing entries during strong trends where mean reversion bleeds. Exit is by net profit target at 2.5%. No short side.

Walk-forward validation matched training performance exactly at 1.7% per trade and 90% win rate, which is cleaner than we typically see and suggests the edge is stable across time splits. The deflated Sharpe of 23.41 is unusually high, a product of the tight regime filter and the high win rate on small moves.

Where to be skeptical: this strategy has zero forward test trades so far. Graduation means it survived six statistical gates on historical data, not that it has proven itself in live conditions. The 15-minute timeframe and ADX filter may produce fewer entries than a pure breakout system, so trade frequency in production could be low. The identical training and validation numbers, while encouraging, also mean we have less variance to study. Regime coverage is not reported at gene level, but ADX below 50 will naturally exclude the volatile breakout periods that dominate certain coins and months.

Check the full survivor registry at stratproof.com/survivors to compare this against other graduates, or head to stratproof.com/prove to test your own mean-reversion idea with regime awareness baked in.