Fractals Evolved: Can Combining Indicators Create the Ultimate EA?
## What's the idea?
A beginner-friendly summary of the verification: “Fractals Evolved: Can Combining Indicators Create the Ultimate EA?”.
What’s the idea?
We’re always on the hunt for robust trading strategies, especially those based on common technical indicators. This time, we wanted to see if we could take a “FractalBreakout” strategy – which essentially looks for price breaking above or below a recent fractal (a specific high or low point) – and make it more reliable by adding other popular indicators as filters. Think of it like this: a fractal breakout can give you a signal, but maybe it’s often a false alarm. Could we use other indicators to confirm that signal, like a second opinion, before entering a trade? We chose three well-known indicators to combine with the fractal breakout:
- Simple Moving Average (SMA): A basic trend-following indicator, smoothing out price data. We might use it to ensure we’re only trading in the direction of the trend.
- Relative Strength Index (RSI): A momentum oscillator that tells us if an asset is overbought or oversold. We could use it to avoid trades when momentum is extreme or to confirm a reversal.
- Average Directional Index (ADX): This indicator measures the strength of a trend. We might use it to ensure we’re only trading when there’s a clear, strong trend in place.
Our hypothesis was that by adding filters like
rsi_min,rsi_max(to keep RSI within a certain range), oradx_min(to ensure a minimum trend strength), we could weed out the weaker fractal signals and create a truly robust trading edge.
How I tested it
To put this idea to the test, I built an Expert Advisor (EA) that combined the FractalBreakout logic with these additional filters. I created 8 different variations, testing each indicator individually and in various combinations.
The most crucial part of this test was using forward testing. This is different from simply backtesting over a historical period. Forward testing simulates how a strategy would perform after its parameters have been optimized on an earlier period, much like how a real trader would deploy an EA. It’s a much tougher, more realistic way to judge a strategy’s true potential. If a strategy looks good in backtests but fails in forward tests, it often means it was “over-optimized” – tweaked to look perfect on past data, but not robust enough for the future.
I ran these tests on two popular timeframes: D1 (daily charts) and H4 (4-hour charts).
To determine if a strategy was truly robust and had a real edge, I set a strict robustness criterion: the strategy had to show an overall profit and be profitable in at least 5 out of 7 years of the forward test period. This helps ensure consistency, not just a lucky run.
What happened?
Here’s the honest truth: the results were disappointing. Even for combinations that looked quite promising in initial fixed in-sample/out-of-sample (IS/OOS) tests – for instance, a “Fractal + RSI < 70” combination that showed over 50% profit in the latter half of its out-of-sample period – none of them met our strict robustness criteria during forward testing. In other words, while some strategies might have teased us with good performance on a specific historical segment, they just couldn’t maintain that profitability when faced with new, unseen market conditions. The best performer among all 8 combinations, found on the H4 timeframe, was profitable in only 4 out of 7 years, and its overall profit was virtually zero. That’s not a real edge, just breaking even over a long period. I also observed something interesting: as I added more filters to the fractal breakout strategy, the number of trades naturally decreased. This makes sense; more conditions mean fewer opportunities. But crucially, this also “flattened” the impressive numbers we sometimes saw in the fixed OOS tests. It seems that while adding filters can reduce the effects of over-optimization (making the strategy less “tuned” to past data), it also seemed to eliminate any perceived edge the base fractal strategy might have had. It’s like trying to filter out impurities from water that has no nutrients to begin with – you just end up with less water.
What I learned
The clear conclusion from this research is that combining standard indicators like SMA, RSI, or ADX with a fractal breakout strategy does not create a robust trading edge. This reinforces a principle we’ve seen before: if the individual components of a strategy don’t have a robust edge themselves, simply combining them isn’t likely to magically create one. It’s like trying to build a profitable business by combining two unprofitable ventures – you’re still left with no profit. This study adds “fractal-based systems” to our growing list of standard technical analysis approaches that, despite their popularity, haven’t shown a consistent, robust edge when subjected to rigorous forward testing. Our overall findings continue to suggest that with price-based strategies, the most consistent edge we’ve identified so far remains firmly in the realm of long-term trend following. So, while fractals and these indicators are useful tools for analysis, relying on their combinations alone for a robust trading system might lead to more frustration than profit. Keep searching, keep testing!
How this connects
This verification builds on earlier ones (what failed before and what I tried this time, comparisons between approaches).