YouTube's DEG Method: Our EA Reveals Its Hidden Trend Power!

A beginner-friendly summary of the verification: “YouTube’s DEG Method: Our EA Reveals Its Hidden Trend Power!”.

What’s the Idea Behind This Strategy?

We’re always on the hunt for robust trading strategies, especially those that promise to combine powerful concepts. This time, we dove into a fascinating approach that merges three giants of technical analysis: Dow Theory, Elliott Wave, and Granville’s Rules. The goal? To build an Expert Advisor (EA) that can identify strong trends and capitalize on them. Here’s how the strategy was designed to work:

Spotting Wave 1: The EA would look for the start of a new trend, which it would identify as “Wave 1.” This was signaled by a combination of factors: a break of a key trendline, a break in the existing Dow structure (meaning the market stops making higher highs/higher lows in an uptrend, or lower lows/lower highs in a downtrend), and a break of the 20-period Exponential Moving Average (EMA). The EMA gives more weight to recent prices, making it responsive.
Waiting for the Pullback: Once Wave 1 was “confirmed,” the EA wouldn’t jump in immediately. Instead, it would patiently wait for a pullback – a temporary dip against the new trend. Specifically, it targeted the 38.2% Fibonacci retracement level. Fibonacci retracements are horizontal lines indicating where support and resistance are likely to occur, and 38.2% is a common level for price to bounce.
Targeting Wave 3: The entry would happen at this 38.2% pullback, aiming to ride the market for the anticipated “Wave 3” – often the strongest and longest wave in Elliott Wave theory.
Managing Profit and Risk: Profits were targeted using a Risk-Reward (RR) ratio of 1:2 to 1:3. This means for every $1 risked, the EA aimed to make $2 or $3. In essence, the “mechanized core” of this strategy was simple: identify an upward trend using both SMA (Simple Moving Average) and EMA, wait for a Fibonacci retracement pullback, and then take profit using a predetermined Risk-Reward ratio.

How We Put It to the Test

To see if this clever combination actually worked, we subjected it to our rigorous testing process. This isn’t just about backtesting (testing on historical data); we focus heavily on forward testing. Think of it as putting the strategy through a simulated live trading environment using data it has never seen before. It’s the closest we can get to real-world performance without actually trading live money. We looked at several key metrics:

Forward Test Pass Rate: How many times did the strategy pass our forward tests? A perfect score would be 6/6.
Parameter Robustness: This is crucial! How well does the strategy perform if we slightly tweak its settings (its “parameters”)? A robust strategy should still do well even with minor changes. If it only works with one exact setting, it’s like a house of cards – easily collapsing.
Core Correlation: How similar is this strategy’s performance to our existing, proven trend-following EAs? A low correlation would mean it offers true diversification, while a high correlation means it’s essentially just another way to catch the same market moves.
Profit Factor (PF): This is calculated as Gross Profit / Gross Loss. A PF greater than 1 means the strategy is profitable. The higher the number, the better; a PF of 1.09, for example, means for every dollar lost, it makes $1.09 – razor-thin margins!

So, What Happened? (The Results!)

At first glance, one specific configuration of the strategy, which we called “DEG RR2” (using Fibonacci retracements between 23-61%, a 20-period EMA, and a 100-period SMA), looked incredibly promising! It passed 6 out of 6 forward tests, which is a fantastic initial result. It made us sit up and take notice. However, as we put it through our deeper “confirmation gates,” the strategy started to show its limitations:

Parameter Sensitivity: This was a big red flag. While DEG RR2 passed 6/6, when we started tweaking the parameters slightly, its performance dropped significantly. For instance, only 4 out of 7 other parameter configurations passed 5 out of 6 tests. Even worse, if we changed the SMA period to 150, the strategy completely collapsed, resulting in a -6.3% loss and only passing 2 out of 6 tests! This means the strategy was highly sensitive to the SMA period – a tiny change could turn a winning strategy into a losing one. Not good for long-term reliability.
Core Correlation: We found the core correlation to be +0.66. While this is still trend-aligned, it’s lower than what we see with raw Fibonacci strategies (which can be as high as 0.84). This suggests that using the Risk-Reward ratio for take-profit does reduce the correlation somewhat, but not enough to make it a truly independent or diversified strategy.
Profit Factor (PF): The strategy’s Profit Factor was 1.09. While this is technically profitable (it’s above 1), it’s a very thin margin. In other words, for every dollar the strategy lost, it only made back $1.09. This doesn’t leave much room for error or unexpected market volatility.

What We Learned (The Big Picture)

The ultimate conclusion from this experiment, like many before it, was a familiar one: when you try to mechanize complex discretionary trading concepts like Dow Theory, Elliott Wave, and Granville’s Rules, you inevitably “rediscover the trend.” What does that mean? It means that no matter how elegantly you combine these powerful ideas into an automated system, the robust edge you find almost always boils down to simply following the long-term trend. This strategy, despite its sophisticated design, ultimately acted very similarly to our existing trend-following EAs (similar to findings in our previous studies like Research 54, 71-76, 82-83, and 83 Granville). While the combination of Fibonacci pullbacks and Risk-Reward take profit can pass forward tests, its high parameter sensitivity and a core correlation of +0.66 tell us it doesn’t offer a truly new “edge” or significant diversification for our portfolio. The RR take profit did reduce the core correlation from the raw Fibonacci numbers (0.84 to 0.66), but it wasn’t enough to make it stand alone. This study reaffirms a consistent conclusion we’ve seen time and again with famous trading methods: no matter how you fuse Dow, Elliott, Granville, Fibonacci, or EMAs into a mechanized, price-only system, the reliable, robust edge always converges back to identifying and following the long-term trend. The real power and profitability in discretionary trading, where traders manually draw lines and count waves, often comes from contextual judgment – the ability to interpret nuance that simply cannot be programmed into an EA. However, not all was lost! The specific components related to FibPullback (specifically its tp_rr for take-profit risk-reward and ema_n for EMA period) are still considered “permanent assets.” This means these particular elements have proven their worth and will continue to be valuable building blocks in our ongoing research and development of robust trend-following EAs.

How this connects

This verification builds on earlier ones (what failed before and what I tried this time, comparisons between approaches).