Search results

Instead of every model trading its own position, all enabled models are polled on each bar and vote. The majority direction opens one shared position; a tie (or no signal) means no entry — so the outcome never depends on which model was asked first. Set InpVoteLog to see the tally per bar.

This resource has been removed and is no longer available.

1) Be careful with models that perform poorly in the second test window, such as Leopard 1(m3617) 2)If you have to choose, opt for models with a high number of trades rather than a low one, even if they appear less appealing. The larger the sample size of market entries, the more reliable the...

This resource is for researchers only. Research result from 07.2026. In this research, 200,000 models were trained, of which up to 160,000 were rejected by the overfit filter, and about 100 models were manually selected based on metrics. All these models are presented in this EA ZOO2. Training...

📌 Idea The idea is — it seems to me that exactly at 10:00 EURUSD always goes up. How can this be verified? This indicator analyzes historical data at the same hour of the day over the past N days and shows: ProbUp — the probability that after a given number of bars (HorizonBars) the price will...

leontolstoy submitted a new resource: MAD TURTLE - #XAUUSD Read more about this resource...

hi, there is no source code here, the source code is = trained model + running the model on mql5. If I give you the model itself in the ONNX wrapper, will it give you that? The only thing I can give you is the FEATURES on which the model was trained.

The model was trained to trade in the direction of the trend. It performs well in an uptrending market but struggles in flat or reversal conditions. In a falling market, the model hardly trades at all — the reason for this behavior during training is currently unclear. It is recommended to use...

leontolstoy submitted a new resource: Trend UP (MIMIC model) - Mimic model Read more about this resource...

The Problem of Model Overfitting or Robots don't understand us.. What happens when machine learning ends up in inexperienced hands? It leads to garbage generation. It results in hundreds of “cool” expert advisors that show perfect backtests — but quickly turn into scams in real trading...

leontolstoy submitted a new resource: DeMark ML (MIMIC model) - Mimic model Read more about this resource...

leontolstoy submitted a new resource: ML - XAUUSD Read more about this resource...

Machine Learning in Trading (AI) If you open articles on MQL5 about machine learning, you'll likely be amazed at how deeply algorithmic traders have advanced in this area. Hidden Markov models, deep neural networks, Gaussian processes, Bayesian optimization, clustering... But why do we...

Please update to version 3.0 Attention! no trading features have been changed! This is the same Frank P. Changes were made in the number of connections to the server, the load was removed from the server, which made it impossible to remove the provider's lock. -- In more detail: recently a DDoS...

Franc Pacific is a CHF scalper that trades between 17 and 22 hours, this watch is the most low-volatility and is well suited for counter-trend scalpers. Intended for EURCHF only. Please note that all strategies of the Franc (CHF) type are poorly reproducible and testing significantly differs...