Using stock-market signals to de-risk FX: a new axis that worked
Using stock market indices as a signal for FX trading has finally allowed me to break through a performance ceiling I had previously hit.
Using stock market indices as a signal for FX trading has finally allowed me to break through a performance ceiling I had previously hit. My hypothesis was that severe drawdowns in FX trends often coincide with stock market risk-off phases. Since indices act as leading indicators, I could use the US500’s position relative to its simple moving average to adjust my leverage. By shifting the index data by one period to avoid hindsight bias, I tested how this market sentiment affects FX position sizing. The results of applying this filtering to the core FX strategy were encouraging:
| Metric | FX Core Only | With Stock Filter |
|---|---|---|
| Calmar Ratio | 0.22 | 0.31 |
| Max Drawdown | -44% | -25% |
| In other words, the filter significantly improved resilience during crisis years such as 2015 and 2018. While 2022 was a difficult year due to the unique combination of falling stocks and a weakening yen, the filter proved complementary to my existing volatility-targeting methods. While my volatility-target approach reacts to realized price changes, the stock index filter acts as a leading indicator, providing a different layer of protection. | ||
| When I applied this logic to the full version of my Core v1.3.1 strategy, the improvements were consistent across the board: |
- Profit Factor (PF): Increased from 1.57 to 1.63. (PF is the ratio of gross profit to gross loss; a number above 1 means the strategy is profitable).
- Max Drawdown (DD): Reduced from -9.5% to -7.9%.
- Calmar Ratio: Improved from 1.02 to 1.14.
- Performance: By re-leveraging the strategy to return to the original -9.7% drawdown level, the monthly return increased from 0.79% to 0.85%, representing a 7.6% boost in performance. This upgrade is now confirmed as v1.4.0. For a long time, I felt I had reached the limit of my search parameters, but this experiment confirms that looking at external assets like stocks provides an “unused axis” for optimization. While my previous attempt at asset allocation did not yield results, this intermarket approach proved effective. It is a good reminder that there are still unexplored methods available (such as regime-detection leverage or alternative bar-type analysis) which I will continue to investigate.
How this connects
This verification builds on earlier ones (what failed before and what I tried this time, comparisons between approaches).