Stacking three uncorrelated engines in one account
Adding the Connors RSI(2) strategy to my existing core and index sleeves does not provide enough of a performance boost to justify changing my…
Connors RSI2 entry example (USDJPY daily, real data): buy the dip when price is above the 200-day SMA and RSI(2) falls below 10.
Adding the Connors RSI(2) strategy to my existing core and index sleeves does not provide enough of a performance boost to justify changing my current system.
I recently tested the Connors RSI(2) strategy, which I previously verified as a high-quality, uncorrelated edge, to see if it could act as a robust third component. I used my combine_results framework to measure how it would perform alongside my existing setup.
Performance comparison
| Metric | Core + Index | Core + Index + Connors |
|---|---|---|
| Monthly Return | 0.61% | 0.63% |
| Drawdown (DD) | -10.2% | -11.5% |
| Monte Carlo Pass Rate | 91% | 92% |
| The Connors RSI(2) strategy is impressive on its own. When traded at a 0.5% risk per trade, it delivers a 33% annual return with a 1.43 profit factor (the ratio of gross profit to gross loss) and a Sharpe ratio of 0.94. In other words, it is significantly more efficient than my core strategy, which sits at a 0.26 Sharpe ratio. | ||
| However, adding it to my current system hits a wall. While the return-to-drawdown efficiency improves slightly, the total drawdown expands beyond my 10% limit. Even when I adjust the risk allocation between the three components, the drawdown remains stubbornly high. Because the strategies overlap during poor market conditions, the net gain in monthly return is only about 0.02%. | ||
| In other words, the Monte Carlo pass rate (the probability of meeting my risk requirements based on resampled returns) is already at a ceiling of 91%. Adding a third component offers almost no room for improvement here. |
Verdict
I am keeping the current system as is. My previous research into uncorrelated edges suggested that while additional strategies are valuable, their impact diminishes as you add more to the portfolio. This test confirms that finding. I will keep the Connors RSI(2) strategy in my library as a verified backup. It is a solid candidate to integrate if the core strategy’s drawdown budget opens up in the future, perhaps after I finish the M1 intraday verification for the index sleeve. For now, my primary focus remains on the index sleeve, as it provides the most significant potential for growth.
How this connects
This verification builds on earlier ones (what failed before and what I tried this time, comparisons between approaches).