Evgeniy Scherbina / Profil

What's up, folks! Another month is wrapping up, and "LSTM Ensemble" kept delivering — up +10% this month. Right now the EA is making a big bet on JPY strengthening, so let's see how March plays out!

In the coming days I'll be dropping updates for both "LSTM Ensemble" and "Edge Ensemble" — and there's a lot to cover. The headline news is the switch to ONNX-based computations. This makes neural network inference significantly faster and cleaner, and most importantly, adding Transformer models to the ensemble is no longer a headache.

Speaking of which — Transformer models are now joining the party! More diverse architectures mean a more robust ensemble, and that's always a win.

And here's the one you've been waiting for: next month "Edge Ensemble" is finally getting a live real-money signal. I've been building toward this for a while, and it's almost here. Stay tuned — March is going to be a big one!

Hi everyone! In the tests I’m still running, I’m coming to a simple conclusion: the "LSTM Ensemble" outperforms the "Edge Ensemble". In theory, a more complex ensemble architecture should deliver better results — but in practice, that doesn’t seem to be the case. Sometimes complexity adds noise instead of value. Here’s the rule I’m following: if a simple model performs as well as or better than a complex one, choose the simple model. Clarity beats complication. Even more exciting, in live mode the "LSTM Ensemble" looks very convincing. In less than two months, it has delivered nearly 30% growth. That’s the kind of performance that speaks for itself.

Hi everyone! I’m back again with new charts — and who knows, maybe not the last ones. I’m still carefully comparing two ensemble approaches: the LSTM+TCN ensemble and the LSTM-only ensemble. As you can see from the images below, the LSTM+TCN model changes direction more frequently and adapts faster to shifting conditions, which helps it reduce drawdowns more quickly.

However, this same responsiveness also results in slightly lower total profit during the August–December 2025 test period. The LSTM-only ensemble, on the other hand, makes adjustments less often, commits to its positions longer, and therefore carries a bit more drawdown. Yet this steadier behavior allowed it to accumulate more profit overall — about 21% more (699 vs. 578).

So what does this mean in practice? Honestly, it might mean absolutely nothing — because only live trading can truly validate these differences. And what are we seeing right now? At the moment, the "LSTM Ensemble" is indeed “hammering in” more profits, while the "Edge Ensemble" (LSTM+TCN) continues searching for higher-quality opportunities with a more adaptive approach.

Very soon these latest updates will be published, and real trading performance will give us the final answer. Stay tuned — the most important results are always the live ones.

What if the key to better results lies not just in the model, but in how we test it? My latest experiment proves just that.

I'm excited to share my best-performing model to date: the LSTM+TCN ensemble. It has achieved the highest returns yet in a rigorous out-of-sample test (Aug-Sep 2025 period, starting balance: $1,000). See the equity curve below.

What changed? I moved away from simplified "maximum exposure" backtesting to a realistic trading simulation. This methodological shift was the game-changer.

I must also highlight the standalone power of a 12-model Transformer ensemble. Yet, the quest to successfully hybridize it with the LSTM+TCN core continues.

In short: the current LSTM+TCN architecture has proven its high effectiveness. An upgraded version will be implemented in the trading advisors shortly.

Hi all, just a quick note: I'm still working on the updates, specifically porting the new structures from Python to MQL5. It's a bit of technical work, but going smoothly. I'll publish everything as promised very soon—possibly even today. Cheers, and thanks for your patience!

They hesitate to write a good review. Only bad ones. Well then, I will write a good review myself. The "LSTM Ensemble" EA continues to earn steadily, closing this calendar month at +22%.

Its robust performance is driven by an ensemble of 12 models of the same class, significantly aided by frequent signal changes. I have now devised a way to improve this ensemble variant further by adding a second timeframe. The primary one is H6, and I am adding H8 this weekend.

I know the rule: if it works well, don't touch it. However, all tests indicate the "LSTM Ensemble" can perform even better with two timeframes simultaneously.

This weekend, I will also update another advisor, the "Edge Ensemble," to include two timeframes alongside its three model classes. Furthermore, advisors for trading Gold and Bitcoin are next in line.

After one month of live trading, the "LSTM Ensemble" EA is showing steady and confident equity growth.

So far, a fixed profit of 4.5% has been recorded, and the current equity is above the balance by the same amount. This early performance already allows for some preliminary conclusions.

It appears that competition between models of the same class, when combined into a single ensemble, may provide tangible advantages. Building an ensemble from models of one class seems to be more predictable and easier to stabilize at early stages.

In contrast, an ensemble that combines models from three different classes clearly requires more extensive testing and longer validation. However, this does not mean that such an approach is inferior. The "Edge Ensemble" EA is also operating in positive territory and remains profitable. It is possible that the "Edge Ensemble" is designed to wait for larger, more significant market movements rather than reacting to smaller fluctuations.

In any case, ensemble-based trading systems are still a very new concept in retail algorithmic trading.

At the moment, no one except me is selling EAs based on neural model ensembles, which is why the price is still relatively low — but I am confident this will change very soon.

Why an ensemble really matters 🧠📊

An ensemble approach means combining multiple neural network models instead of relying on just one 🔗.

Right now, I’m the only developer actually selling a true neural ensemble in trading — and that makes a real difference 🚀.

Currently available:
“LSTM Ensemble” — combines 12 LSTM models
“Edge Ensemble” — combines 12 models of two different types: LSTM and TCN

Each model is trained on different folds of the data, so they learn different market behaviors instead of repeating the same mistakes 🎯.

I’m currently working with LSTM, TCN, and Transformer models, each representing a different class of memory and pattern recognition 🧩. TCN and especially Transformer are mathematically complex models, and that complexity allows them to capture structures that simpler models may miss 🧮.

Because each model “remembers” the market differently, the ensemble becomes more stable, adaptive, and resilient over time 🛡️.

Below, you can see a chart showing how this ensemble performs on a true out-of-sample test period, far beyond the training data 📈.

Selecting the winning model within its own fold is a critical step — this diversity is the real power behind the ensemble ⚡. The result is honestly stunning, unlike anything I’ve seen before ✨.

And yes — strategies of this class, powered by neural ensembles, are available only here 🔥.

For years, the industry has blindly followed the traditional 80/20 training split ⚙️ — and in reality, this approach is flawed twice over.
When you train on 15 years of data, the most recent 3 years are thrown away ❌ and never used for learning.
Even worse, those final years cannot represent the full diversity of market behavior that came before them 📉📈.

A far more robust solution is to use historical folds 🧩, where each period trains its own dedicated model.
Instead of one fragile model trying to learn everything, I found a solution in an ENSEMBLE ARCHITECTURE, where each model specializes in its own market regime 🔍.
This approach mirrors how real markets work: they evolve, rotate, and repeat — but never stay the same 🔄.

By combining proven architectures like LSTM 🧠, known for their stability, with modern innovations like Temporal Convolutional Networks (TCN) ⚡, a true breakthrough in time-series forecasting emerges.
TCN is rapidly becoming a new standard for sequential data, including complex financial markets 💹, and its synergy with LSTM is remarkably powerful 🤝.

Today, there is not a single marketplace strategy built on a true 12-model ensemble 🚀.
This is not a cosmetic improvement — it is A FUNDAMENTAL SHIFT in how trading systems are designed and validated 🛠️.

What is most striking in technical testing is the consistency 📊: even with extremely aggressive settings where the history is almost memorized, the strategy still manages to push the equity curve upward on completely unseen data 📈✨.
That level of robustness is not an accident — it is the direct result of THE ENSEMBLE PHILOSOPHY 🎯.

Over the next year, I plan to release several more strategies, all based on this ensemble-driven approach 🧪.
Let’s test a new way of trading together — one that finally respects market history instead of discarding it 🤖💡.

Happy Holidays and Happy New Year!

Wishing you a profitable year ahead, steady growth, and confident trading decisions in the Forex market.

To celebrate the holiday season, I’m offering 35% discounts on my selected best products:
LSTM Ensemble — new price $210 USD
Chanlun Master — new price $210 USD

May the new year bring strong trends, clean signals, and plenty of profits.
Trade smart, manage risk, and let the market work in your favor.

I am thrilled to announce 📣 that I will soon be releasing the Ensemble Trader, a new advisor that integrates 12 distinct models with stunning results 🚀💥.

Traditional machine learning often relies on a standard 80/20 data split 🤖, but it is strange to think 🤔 that the final 20% can truly "verify" the entire learning process. Over a long history ⏳, the market transitions through vastly different environments; while Brexit and the COVID crash 📉 are the most obvious examples, there are countless other specific trading regimes. A single rigid model simply cannot adapt to all of them effectively 🧊❌.

This adaptive strategy employs an ensemble approach where each sub-model 🧩 is responsible for covering a specific regime found in historical data. Instead of relying on a single logic, the models weigh their inputs ⚖️ to reach a robust consensus decision 🤝✅.

To be clear, this strategy still does not draw a perfectly straight equity line 🚫📈, because that is theoretically impossible in real markets. However, the Ensemble Trader solves many classic stability issues ⚓🛠️. It finally provides solid grounds to believe that earning with a fully automated Forex strategy is absolutely possible 💰🎯.

Here are the first steps of my newest strategy, Chanlun Master.
You can already check out a live demo signal on myfxbook.com.
You’ll notice that the strategy expects large moves and usually opens trades early in the day.

One of the key differences between this strategy and my previous ones is the way it’s trained. It’s based on time-series learning, which usually means you split the data by time — the “right” part of the data, the most recent one, is excluded from training. But after thinking about it, I realized this approach is both illogical and counter-productive.

So I changed it. I split the training into folds, where each fold has its own training and validation sets. Tests show that this method is more stable and reliable. Of course, the real results will become clear in a few months — fingers crossed!

Hi everyone! 👋

My new EA Chanlun Master blends the Eastern logic of Chanlun 🏯 with modern neural networks 🤖 to give you a clean and intuitive trading style.
It focuses on spotting Super Turning Points ⚡ — those moments when a trend finally runs out of steam and a strong reversal is likely.
Smart filtering keeps out the noise, and the system works smoothly across several symbols on higher timeframes 📈.

What really makes it shine is consistency: historical tests show stable accuracy, loss, and—most importantly—profitability 🚀 even on unseen data beyond the training period.
This kind of stability means the model isn’t overfitting, but actually learning real market behavior.

Chanlun Master is simple, focused, and surprisingly powerful.

The EurUsd has been moving sideways for months now, locked in a massive consolidation that started back in the summer. Anyone who's watched this pair long enough knows what that usually means — a big move is coming. The current range looks almost identical to the ones we saw in late 2024 and early 2025, and even to that long, painful consolidation at the start of 2024. History tends to rhyme, and right now it's whispering "trend breakout."

Most likely, that breakout will be to the upside, with potential targets around 1.20 to 1.40. It won't happen overnight, but the setup is already there. So what's been holding EurUsd back? Mainly politics and uncertainty out of France, where the financial and political crises are still dragging on. Germany already went through a similar phase, and its right-centrist forces managed to stabilize things. France, on the other hand, is still trying to figure out what's next — and Macron, being a republican rather than a socialist, has limited room to maneuver.

Once that cloud lifts, the euro could easily regain momentum. Whether it's by the end of the year or even in the next month, a strong rally in EurUsd looks more like a question of when, not if.

My Latte EA operates in a trend-following mode — it's built to catch big moves, not minor noise. And right now, it's been waiting far too long. The market's been quiet, but that silence won't last forever. What's clear is that Latte's cumulative position remains against the USD — and that's the right side to be on. When the breakout finally comes, Latte will already be where it needs to be.