Articles on trading system automation in MQL5

As we progressed, we used more and more simultaneously running instances of trading strategies in one EA. Let's try to figure out how many instances we can get to before we hit resource limitations.

The EA under development is expected to show good results when trading with different brokers. But for now we have been using quotes from a MetaQuotes demo account to perform tests. Let's see if our EA is ready to work on a trading account with different quotes compared to those used during testing and optimization.

We are going to create a matrix forecasting model based on a Markov chain. What are Markov chains, and how can we use a Markov chain for Forex trading?

Using a ready-made solution in trading without concerning yourself with the internal workings of the system may sound comforting, but this is not always the case for developers. Eventually, an upgrade, misperformance, or unexpected error will arise, and it becomes essential to trace exactly where the issue originates to diagnose and resolve it quickly. Today’s discussion focuses on uncovering what normally happens behind the scenes of a trading Expert Advisor, and on developing a custom dedicated class for displaying and logging backend processes using MQL5. This gives both developers and traders the ability to quickly locate errors, monitor behavior, and access diagnostic information specific to each EA.

The article extends a liquidity-based strategy with a simple trend constraint: trade liquidity zones only in the direction of the EMA(50). It explains filtering rules, presents a reusable TrendFilter.mqh class and EA integration in MQL5, and compares baseline versus filtered tests. Readers gain a clear directional bias, reduced overtrading in countertrend phases, and ready-to-use source files.

In order to use the data that forms the bars, we must abandon replay and start developing a simulator. We will use 1 minute bars because they offer the least amount of difficulty.

Training useful skills without an explicit reward function is one of the main challenges in hierarchical reinforcement learning. Previously, we already got acquainted with two algorithms for solving this problem. But the question of the completeness of environmental research remains open. This article demonstrates a different approach to skill training, the use of which directly depends on the current state of the system.

Build an MQL5 Expert Advisor that automates Larry Williams Hidden Smash Day reversals. It reads confirmed signals from a custom indicator, applies context filters (Supertrend alignment and optional trading‑day rules), and manages risk with stop‑loss models based on smash‑bar structure or ATR and a fixed or risk‑based position size. The result is a reproducible framework ready for testing and extension.

In this article we attempt to simplify our illustration of concepts covered in these series by dwelling on just one indicator, the most common and probably the easiest to understand. The moving average. In doing so we consider significance and possible applications of vertical natural transformations.

The article replaces hardcoded cost assumptions in triple-barrier labeling with measured inputs. An MQL5 script captures spread distribution, swap rates, and symbol metadata from your broker, and a Python model converts them into a broker-calibrated min ret you can pass to get events. Labels then reflect the actual round-trip friction for your instrument and holding period.

The article addresses the loss of temporal information in ML pipelines by encoding periodic time variables with Fourier harmonics and adding forex session structure. It implements session and overlap flags, lagged session volatility, and calendar effects, then prunes features by timeframe. The get time features function returns an index‑aligned, ML‑ready set of time features suitable for integration with price‑based signals.

The Contrastive Transformer is designed to analyze markets both at the level of individual candlesticks and based on entire patterns. This helps improve the quality of market trend modeling. Moreover, the use of contrastive learning to align representations of candlesticks and patterns fosters self-regulation and improves the accuracy of forecasts.

We continue to study timeseries forecasting algorithms. In this article, we will discuss another method: the U-shaped Transformer.