Most traders reach for RSI when they need an oscillator filter. I ran 4.5 million tests to see if that is actually right.
It is not.
RSI hit a 50% recommendation rate across our test set - 1,000 strategies, 1,000 market conditions. Good enough to use, not great, not robust.
CCI hit 100%.
The condition that made the difference was not the standard CCI above or below zero. It was this: take the highest CCI value over the last N bars. If that value is above 100, meaning the market touched momentum territory recently, even once - that is your filter.
Period 30. Lookback 12 bars. Average trade improved by 65% across 650 strategies. Maximum drawdown cut from $40,000 to $30,000. Bounce index- the percentage of losing strategies that recovered up 60%.
The optimization map showed consistent green across almost every parameter combination we tested. You can pick almost any value within the valid range and you are not overfitting.
I have 110% confidence in this filter because I did not test it on one strategy. I tested it across 1,000 strategies and 1,000 market conditions.
RSI: 50%. CCI: 100%. 4.5 million iterations.