Articles on trading system automation in MQL5
Read articles on the trading systems with a wide variety of ideas at the core. Learn how to use statistical methods and patterns on candlestick charts, how to filter signals and where to use semaphore indicators.
The MQL5 Wizard will help you create robots without programming to quickly check your trading ideas. Use the Wizard to learn about genetic algorithms.
Add a new article
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Neural Networks in Trading: Robust Trading Signals in Any Market Regime (ST-Expert)
In this article, we will explore the ST-Expert framework, which ensures the robustness of forecasts under market uncertainty by taking local and global dependencies in time series into account. Its flexible architecture promotes model adaptability and improves the accuracy of predictions.
Intrinsic Time: From the Directional-Change Scaling Laws to the Alpha Engine
The article implements intrinsic-time analysis in MQL5: an event-based directional-change operator that splits ticks into directional-change and overshoot sections. We reproduce the core scaling laws on 17.8 million live EUR/USD ticks and compare them to a random-walk baseline. Finally, we build a hedging-account Expert Advisor that trades the Alpha Engine with limit orders, detailing thresholds, inventory skew, and liquidity control for practical reuse.
Neural Networks in Trading: A Unified View of Space and Time (Conclusion)
The Extralonger framework demonstrates a unique ability to integrate spatial and temporal factors into a single model, ensuring high forecast accuracy. Its architecture allows it to adapt to different planning horizons and financial instruments while maintaining the system's transparency and manageability.
Self-Exciting Markets: Building a Hawkes Process from Scratch
Volatility arrives in clusters: one large move makes the next large move more likely, and quiet spells stay quiet. This article builds a Hawkes self-exciting point process in pure MQL5 to measure that effect directly, ending in a single number, the branching ratio, that says how reflexive a market currently is. You get a small, tested library, an indicator that plots the fitted intensity live, and a demonstration Expert Advisor, along with the honest limits of all three.
Architecture for Collective Trading Decisions by AI Agents
The article describes the architecture of a multi-agent trading system based on the grok-4-fast language model, in which, instead of a single system prompt, four independent analysts with fundamentally different roles operate: a bull, a bear, a risk manager, and an arbiter. Three analysts run in parallel using a ThreadPoolExecutor and, within 3–5 seconds, formulate well-reasoned positions based on the same market data; after that, a deterministic judge renders a final verdict according to strict rules.
Encoding Candlestick Patterns (Part 5): Expanding Taxonomy of Candlestick for General Pattern Frequency Analysis
This article extends the candlestick encoding framework by separating previously unclassified bullish and bearish candlesticks into distinct symbols. Using an MQL5 script, historical data were automatically encoded, sequential patterns were extracted, and frequency statistics were generated. The results show that these newly identified candles represent a significant portion of market activity, revealing structural information previously hidden by the generic underscore notation. The expanded taxonomy improves the completeness of the symbolic representation while preserving an objective, quantitative framework for large-scale market structure analysis.
Neural Networks in Trading: A Unified View of Space and Time (Global-Local Attention)
We are continuing our work on implementing the approaches proposed by the authors of the Extralonger framework. This time, we will focus on building a Global-Local Spatial Attention module using MQL5, examining both its structure and its practical integration into the overall computational process.
Working with ONNX Models in MQL5 (Part 2): Drawing the Model Graph on an Interactive Chart Panel
We turn the ONNX reader into an interactive graph viewer on a MetaTrader 5 chart. The parser is extended to carry tensor shapes and node attributes, and a canvas with anti-aliased drawing renders nodes and labeled connections. The layout arranges layers by depth, while zoom, pan, and clicks let you inspect operators and the tensors they read. This makes a model's structure and data flow visible directly in MQL5.
Position Management: Deriving a Self-Calibrating Exit Ladder From Historical MFE in MQL5
We implement three MQL5 classes that replace fixed 1R/2R/3R targets with data-driven scale-out levels. CExcursionTracker records each closed trade's maximum favorable excursion in R, CExitLadderCalibrator derives runs from distribution percentiles with lookback and minimum-sample controls, and CLadderExecutor executes them on open positions. The ladder recalibrates as trades accumulate and uses a fallback until enough samples exist.
Neural Networks in Trading: A Unified View of Space and Time (Extralonger)
The Extralonger framework demonstrates an approach to integrating spatial and temporal factors into a single model, which makes it possible to account for both local patterns and long-term cycles simultaneously. This architecture makes time series forecasting more resilient to market noise and enables data analysis across different time horizons. The article takes a detailed look at how these ideas are put into practice using OpenCL and MQL5.
Testing for Residual Autocorrelation with the Ljung-Box Portmanteau Test in MQL5
A complete MQL5 implementation of the Ljung-Box test helps verify independence in trading data and fitted-model residuals. It computes sample autocorrelations, the Q statistic over selected horizons, degrees of freedom with user-controlled adjustments, and right-tail p-values via the regularized incomplete gamma function. Run it on returns, deal outcomes, or external residuals and review decisions directly in the Experts tab.
Neural Networks in Trading: The Adaptive Graph Diffusion Model (Conclusion)
In this article, we conclude our work on building the SAGDFN framework using MQL5, summarizing the development process and presenting the results of its practical testing. Let's combine the modules we've already implemented into a single system, highlight the strengths of this approach, point out its weaknesses, and discuss possible ways to improve it.
Building AI-Powered Trading Systems in MQL5 (Part 11): Optimizing the UI with Frame Throttling and Partial Rendering
We optimize an MQL5 canvas interface to stay responsive under rapid input without changing its appearance. The article adds a direct-buffer canvas for block region copies, caches text widths and glyph coverage, caps repaints at 60 fps (16 ms), and limits drawing to panes and regions that actually changed. As a result, hover, scroll, and popups render smoothly without full-panel redraws.
Practical Modules from Other Languages in MQL5 (Part 07): The OS Module from Python
This article introduces a lightweight OS-like helper for MQL5 that streamlines file and path operations using a Python-inspired interface. We implement getcwd, listdir, scandir with DirEntry, remove, rmdir, rename, mkdir, stat, and an os.path subset (exists, isfile, isdir, join, split, pardir). You will learn how to work consistently within the terminal Files/common sandbox and simplify everyday filesystem tasks.
Automating Trading Strategies in MQL5 (Part 53): Double Top and Double Bottom Reversal Model
We build an MQL5 program that detects and trades Double Top and Double Bottom patterns from confirmed swing pivots with rule-based logic. The detector matches peaks within a tolerance, derives the neckline, and applies leg-balance and spacing filters to reject weak shapes. It supports entry on a neckline break or an optional pullback, with a stop beyond the extreme and targets by measured move or reward-to-risk.
Neural Networks in Trading: The Adaptive Graph Diffusion Model (Attention Module)
In this article, we will take a detailed look at the practical implementation of the key components of the SAGDFN framework. We will show how sparse attention and the selection of significant neighbors are organized for time series forecasting. The approaches presented strike a balance between forecast accuracy and computational efficiency.
How to Connect an LLM to an MQL5 Expert Advisor via a Python Server
The article examines three key obstacles to integrating LLMs with MetaTrader 5: the lack of direct access, strict rate limits, and API key security given the architectural limitations of MQL5. A configuration is proposed that uses a local Python server as a bridge between the Expert Advisor and OpenRouter. The article covers WebSocket and fallback to TCP, storing the key on the server, batch processing of multiple symbols, and constructing a technical prompt. Readers get a ready-made architecture that reduces latency and costs.
Replay and Market Simulation: The Grand Finale
I know that many of you may have thought I would publish a few more articles to explain other aspects of this system. The missing elements are easy to implement. Even so, developing them will let you see how prepared you really are.
Market Simulation: Unity Is Strength (III)
In this article, I will present our system for simulating market operations. Although everything is practically finished, there are still a few things to implement and a few changes to make. However, I have to admit that, after everything we've already developed, I'm tired of still being stuck on implementing this system.
Developing a Multi-Currency Expert Advisor (Part 32): Secrets of the Optimization Project Creation Step (II)
The article discusses the parameters of the second stage of the automatic optimization pipeline for a multi-currency Expert Advisor. We analyze the criteria for filtering first-stage passes and the rules for forming groups of trading strategies. The article demonstrates how settings affect optimization results, discusses aspects of process reliability, and examines the balance between selection strictness and having enough candidates for the algorithm.
Python-MetaTrader 5 Strategy Tester (Part 06): MQL5-Style Backtesting for Python Expert Advisors
Code and build Python-based trading robots just like MQL5 Expert Advisors (EAs). In this article, we develop a Python-based replica of the MetaTrader 5 Python package, providing methods that closely resemble those of MetaTrader 5 during simulation. This allows us to backtest Python EAs in a simplified environment, using an approach similar to developing and testing Expert Advisors in MQL5.
Building AI-Powered Trading Systems in MQL5 (Part 10): A Resolution-Independent Vector Icon System
We replace the embedded bitmap icons in our MQL5 canvas interface with a resolution-independent vector icon system. A small set of anti-aliased primitives (strokes, discs, rings, rounded rectangles, and polygons) draws every logo and sidebar glyph procedurally, then plugs into the header, sidebar, and theme toggle. You get smaller builds, theme-aware recoloring, and icons that stay sharp at any size.
Developing a Reusable Dynamic Volatility Trailing Stop Engine in MQL5
This article presents a modular, object-oriented volatility trailing stop engine for MQL5 packaged as a reusable include class, it calculates dynamic stop-loss levels using a True Range average filter on closed bars, supports step-based trailing, includes a visual diagnostic indicator, and provides an Expert Advisor execution template.
Defining your Edge (Part 6): Harnessing Fourier Transform and a Spiking Neural Network in an Expert Advisor
Article revisits Trading Robot that merged Discrete Fourier Transform with Leaky Integrate-and-Fire Spiking Neural Network. Evaluation is made over the seven operating modes with different symbols, timeframes, and test windows. We use two-thirds of the test window to optimize while the one-third does a forward walk run. This study tries to detail parameter interactions, input scaling, gating effects, and outlines practical checks that include rolling windows, frozen inputs, and others. The goal remains identifying settings worth further testing or paper trading.
Market Simulation: Position View (XIX)
One of the issues that bothered me the most was that the `C_ElementsTrade` class contains code for accessing positions. Don't take this as a mistake, because it really isn't one. However, this increases the risk of errors in some of the tasks we will need to handle later. All work on implementing the position indicator was carried out with a view to its use in the replay/simulation service. However, when running in this environment, we will have no access to actual positions. Consequently, any call to the MQL5 library intended to retrieve position data will have no effect in this environment.
Neural Networks in Trading: Heterogeneity-Informed Meta-Parameter Learning (Conclusion)
The article describes a practical implementation of the HimNet framework based on MQL5, ready for integration into automated trading. We demonstrate how heterogeneity-adapted meta-parameters transform the model into a universal tool capable of handling fluctuating volatility.
Neural Networks in Trading: Heterogeneity-Informed Meta-Parameter Learning (Key Components)
In this article, we take a detailed look at the algorithms used to implement the key components of the HimNet framework. We demonstrate how, with a minimal number of trainable components, a high degree of consistency and controllability can be achieved throughout the entire system. The presented implementation is compact and transparent, which makes it easier to adapt to real-world market tasks.
Building a News Filter Engine in MQL5 Using a Local Economic Calendar File
A file-based news filter for MQL5 reads a pre-downloaded Forex Factory CSV from MQL5/Files, avoiding fragile web scraping and paid APIs. It provides a modular CNewsFilter with a quote-aware CSV parser, suffix-robust currency extraction, an inclusive time-window checker with clear block reasons, and chart zones for today's events. A demo EA and assertion tests help you integrate and verify offline filtering around scheduled releases.
Automating Classic Market Methods in MQL5 (Part 8): Ed Seykota's Trend Following System
The article presents a full MQL5 implementation of a multi-symbol trend system: dual EMA crossovers for entries, ADX to avoid ranges, ATR to normalize position size, and a heat monitor to cap total portfolio risk. We explain the architecture, calculation details, and entry/exit logic on daily bars. The result is a practical EA template for systematic, risk-aware portfolio trading.
Developing a Multi-Currency Expert Advisor (Part 31): Secrets of the Optimization Project Creation Step (I)
The article examines two practical aspects of the Adwizard-based optimization pipeline: diagnostics and recovery after failures when generating the final Expert Advisor database, as well as preliminary selection of strategy parameter ranges before project creation. It is shown how analyzing the stages/jobs/tasks tables in SQLite and restarting stages based on their statuses help restore the process, while trial optimization narrows the search space, eliminates redundant parameters, and reduces the risk of getting stuck at local maxima.
Neural Networks in Trading: Heterogeneity-Informed Meta-Parameter Learning (HimNet)
We invite you to explore the HimNet framework, which combines the flexibility of spatio-temporal adaptation with high computational efficiency, enabling accurate and stable forecasts for financial time series. The article explains in detail how its key components interact with one another, transforming complex algorithms into a manageable architecture.
Machine Learning in Pure MQL5 (Part 1): Logistic Regression from Scratch with SGD
The series develops machine learning in 100% native MQL5 with no external dependencies. Part 1 delivers logistic regression from first principles: a CLogReg class with standardization, a stable sigmoid, SGD training, and model persistence, plus a script that builds ATR-normalized features, labels the next bar, and tests out-of-sample against a baseline. Readers get a compact include file and a clear template for leakage-free evaluation.
Automating Classic Market Methods in MQL5 (Part 7): The Nicolas Darvas Box System
This article implements the Darvas Box method as a complete MQL5 Expert Advisor. We code box detection with a three-session hold, volume contraction during consolidation, and volume-confirmed breakouts, plus a staircase pyramid with a shared, rolling stop at the latest box floor. The EA uses a state machine to run box scanning and trade management in parallel, providing a ready-to-compile system with configurable inputs and clear on-chart diagnostics.
Designing a Multi-EA Communication Bus Using Named Pipes in MQL5
This article implements a typed message bus over Windows named pipes to replace MetaTrader's untyped GlobalVariables for inter‑EA communication. A broker EA manages the server and registry, serves multiple slave EAs, and responds with a live, per‑symbol‑attributed portfolio risk measure. It also explains the non-blocking accept pattern that preserves terminal responsiveness, and includes a dashboard and a test script.
Neural Networks in Trading: The Temporal Query Model (Conclusion)
We are pleased to present the final stage of the TQNet framework’s development and testing, where theory meets real-world trading practice. We will move from historical training to a stress test using recent market data, evaluating the model's robustness and accuracy. The final results are not just dry statistics, but also a clear demonstration of the practical value of the proposed approach.
Development and Forward Testing of an Autonomous LLM Agent for Trading with SEAL
A hybrid architecture based on Llama 3.2 and SEAL is being tested on eight currency pairs (M15), with forward-period data isolation and information leakage control. The methodology combines adversarial self-play, curriculum learning, and class balancing to ensure stable training. The experiments confirm the gap between forecast accuracy and actual returns, providing readers with practical guidelines for testing strategies and accurately assessing their generalizability.
Neural Networks in Trading: The Temporal Query Model (TQNet)
The TQNet framework opens up new possibilities for modeling and forecasting financial time series by combining modularity, flexibility, and high performance. The article explores the possibility of implementing complex mechanisms for handling global correlations, including advanced parameter initialization methods.
Developing a Multi-Currency Expert Advisor (Part 30): From Trading Strategy to Launching a Multi-Currency Expert Advisor
The article outlines the complete process of creating a multi-currency Expert Advisor using the Adwizard library for MetaTrader 5: from setting up the environment for creating optimization projects to obtaining the final multi-currency Expert Advisors, which combine multiple instances of a simple trading strategy. We will walk through setting up the necessary input parameters, conventions for convenient file names, and launching three instances of the final Expert Advisors on different trading accounts with different parameters.
Measuring What Matters (Part 4): Reading the Spectrum — What Eigenvalues Tell You About Risk
We turn eigenvalues from a covariance matrix into a normalized spectral‑entropy score that measures how evenly variance is spread across factors. SpectralEntropyCalculator.mq5 compares two portfolios in one run, using native vector summation, ArraySort()-based ordering, element‑wise division, and the Shannon entropy formula. The report makes dominant factors visible and enables quick, repeatable checks of diversification quality.
Building Your Personal Expert Advisor (Part 6): Risk Management V — Portfolio and Correlated Risk
This part implements PortfolioRisk.mqh, a shared library that shifts risk management to the account level. It scans positions and pending orders, computes margin and floating results, counts symbols, and decomposes pairs into currencies to detect concentration, then validates each new trade against portfolio limits. The Series EA example illustrates configuring scope (account-wide or magic-filtered), registering magics, and integrating the pre-trade gate.