SMA Cross in Sideways Markets: Perfect on Paper, Untested Live

We tested a classic SMA 21/50 golden cross strategy with one twist: it only enters when ADX is below 50, meaning it hunts for trend starts in sideways or weak-trending conditions. The strategy graduated from our six-gate filter with a 10.37 deflated Sharpe ratio (a risk-adjusted return metric that corrects for backtest overfitting) and a 12.3% probability of being a statistical fluke after adjusting for multiple testing. Both validation metrics matched training perfectly at 1.796% average profit per trade with a 100% win rate.

Those numbers sound too good, and they probably are. The backtest ran on 15-minute bars across ten coins (BTC, ETH, SOL, XRP, DOGE, BNB, ADA, AVAX, LINK, LTC), but we don't have a stored trade count at the gene level. A 100% win rate in both training and validation windows is a red flag for sample size: either the strategy triggered very few trades, or it got lucky in a specific regime slice. The zero live trades in forward paper testing since graduation confirm this is a low-frequency setup. It hasn't had a chance to prove or disprove itself in real market conditions.

The logic is sound in theory. Waiting for ADX to stay below 50 filters out established trends where a lagging moving average cross would enter late. The strategy wants to catch the moment a consolidation breaks into a new trend. But sideways markets with occasional breakouts don't happen every day, especially on a 15-minute timeframe. The 2% take-profit target (implied by tp2 in the gene ID) likely closed trades quickly when they worked, padding the win rate but capping upside.

We graduated this strategy because it passed our statistical gates, not because we have conviction. A 12.3% fluke probability sits well below our 50% threshold, but it's far from the single-digit confidence we see in our strongest survivors. Until this generates a meaningful live trade sample, treat it as a hypothesis that survived backtest scrutiny, not a proven edge.

Check the full survivor registry at stratproof.com/survivors to compare this against higher-frequency graduates, or bring your own low-ADX idea to stratproof.com/prove and see if it clears the same gates.