FX Revolution! How Stock Market Signals Unlocked a New Edge!
## What if the Stock Market Held a Secret for Your FX Trades?
A beginner-friendly summary of the verification: “FX Revolution! How Stock Market Signals Unlocked a New Edge!”.
What if the Stock Market Held a Secret for Your FX Trades?
We’ve always been on the hunt for new ways to improve our algorithmic FX trading strategies (EAs). Today, we’re diving into a fascinating new approach: using the stock market as a kind of “weather forecast” for our FX trades! This is the first time we’ve tried using an external index as a signal, and the results are pretty exciting.
The Big Idea: Using Inter-Market Signals
Our core hypothesis was simple: what if the big drawdowns (DD) – those painful drops in equity – that our FX trend-following strategies experience often coincide with “risk-off” periods in the stock market? And what if the stock market could act as a leading indicator, giving us a heads-up? To test this, we looked at the US500 stock index. We defined “risk-on” as when the US500 was trading above its Simple Moving Average (SMA), suggesting a generally positive market sentiment. “Risk-off” was the opposite. Our idea was to dynamically adjust the leverage of our core FX trend strategy based on this US500 sentiment. If the stock market was in a risk-off phase, we’d dial down the leverage on our FX trades, hoping to protect ourselves from potential FX drawdowns.
How We Put It to the Test
We developed a system to inject this “leverage series” into our FX strategy. Crucially, we ensured there was no “look-ahead bias.” This means the decision to adjust leverage for today’s FX trades was based only on yesterday’s US500 sentiment. We didn’t cheat by using future information!
First Results: A Breakthrough for Our Core FX Strategy
We started by applying this stock market filter to our core FX strategy on its own, without any other bells and whistles. The improvements were significant:
- Calmar Ratio (a measure of risk-adjusted return, where higher is better) jumped from 0.22 to 0.31. In other words, for every unit of risk taken, the return was notably better.
- Maximum Drawdown (DD), which is the largest peak-to-trough decline in our equity curve, improved dramatically from -44% to -25%. That’s a huge reduction in potential portfolio pain! This filter particularly helped during known crisis years:
- In 2015, the drawdown improved from -23% to -13%.
- In 2018, it went from -18% to -13%. However, it wasn’t a silver bullet for every tough period. For example, 2022, a year marked by both stock market declines and a strong yen depreciation, was a bit of a sacrifice for this strategy. It’s important to remember that no strategy is perfect for all market conditions.
Even Better: Combining with Volatility Targeting
What’s more, we found that this stock market signal was complementary even when combined with our existing “vol-target” approach (where we adjust position sizes based on the asset’s own historical volatility). The stock market signal, being a leading indicator, offered a different perspective than the asset’s own volatility, which tends to be a lagging indicator. This means these two methods work together, each bringing a unique benefit to the table.
Applying the Magic to Our Full EA: Version 1.3.1 Gets an Upgrade!
Encouraged by the standalone results, we applied this inter-market signal to our entire real Core v1.3.1 strategy. The overall improvements were fantastic:
- Overall Maximum Drawdown (DD) dropped from -9.5% to -7.9%. Less risk, yay!
- Profit Factor (PF) increased from 1.57 to 1.63. The Profit Factor is simply gross profit divided by gross loss; a PF greater than 1 means the strategy is profitable, and a higher number means it’s more profitable per unit of loss. So, we’re making more money relative to our losses.
- Calmar Ratio improved from 1.02 to 1.14.
- The strategy’s overall robustness (how well it performs under various simulated market conditions) remained stable.
The Bottom Line: What This Means for Your Trading
These improvements are significant! If we were to re-leverage the strategy to achieve a similar drawdown level as before (say, -9.7%), this new approach would boost our monthly profit rate from 0.79% to 0.85%—that’s a 7.6% increase in monthly returns for essentially the same risk! This confirms that the new approach is a genuine step forward, and we’re officially rolling it out as v1.4.0 of our Core strategy.
Beyond the Ceiling: Why This Discovery Matters
For a while, it felt like we’d hit a “ceiling” in our research, limited by what we could extract from price data within a single asset. As some of our users pointed out, we were perhaps reaching the limits of our exploration range. But this breakthrough proves otherwise! By tapping into information from a different asset (stocks) and using it as an “unused axis” in our analysis, we’ve managed to push past that ceiling. It shows that even when previous attempts (like certain allocation strategies we explored in Research 95) didn’t work out, there are still completely unexplored avenues out there. We’re continuing to explore other promising areas, such as regime detection for leverage adjustments, different bar types for analysis, and even sequence-based machine learning. The journey to better EAs is far from over!
How this connects
This verification builds on earlier ones (what failed before and what I tried this time, comparisons between approaches).