MQL5 Programming Articles

In this article, we expand the range of tasks of our agent. The training process will include some aspects of money and risk management, which are an integral part of any trading strategy.

Algorithmic trading becomes more popular and needed, which naturally led to a demand for exotic algorithms and unusual tasks. To some extent, such complex applications are available in the Code Base or in the Market. Although traders have simple access to those apps in a couple of clicks, these apps may not satisfy all needs in full. In this case, traders look for developers who can write a desired application in the MQL5 Freelance section and assign an order.

In this article, we will discuss one of the possible solutions to what we have been trying to demonstrate—namely, how to allow an Excel user to perform an action in MetaTrader 5 without sending orders or opening or closing positions. The idea is that the user employs Excel to conduct fundamental analysis of a particular symbol. And by using only Excel, they can instruct an expert advisor running in MetaTrader 5 to open or close a specific position.

Interview with Boris Odintsov (bobsley) is the last one within the ATC Champions League project. Boris won the Automated Trading Championship 2010 - the first Championship held for the Expert Advisors in the new MQL5 language. Having appeared in the top ten already in the first week of the ATC 2010, his EA brought it to the finish and earned $77,000. This year, Boris participates in the competition with the same Expert Advisor with modified settings. Perhaps the robot would still be able to repeat its success.

The article examines an engineering approach to optimizing an Expert Advisor in MetaTrader 5: from collecting custom metrics through Optimization Frames to parameter surface analysis. A simple event-driven EMA/RSI model demonstrates CSV export, smoothing, and local stability assessment in Python. The goal is to find stable areas of configurations and validate them with forward optimization for reliable implementation.

The bet-sizing signal from Part 10 is concurrency-corrected but carries no payoff-ratio adjustment, no response to a hard drawdown budget, and no validation across combinatorial paths. This article covers three additions: a two-stage architecture in which a Kelly payoff multiplier is applied on top of get_signal, preserving the concurrency correction while incorporating win/loss asymmetry; a prop firm integration layer that calibrates the sigmoid w parameter continuously from the remaining drawdown budget under FundedNext Stellar 2-Step rules; and a CPCV backtest framework that simulates a fresh account state across all φ[N, k] paths, producing a Sharpe distribution and a PBO audit.

Fixed fractions and raw probabilities misallocate risk under overlapping labels and induce overtrading. This article delivers four AFML-compliant sizers: probability-based (z-score → CDF, active-bet averaging, discretization), forecast-price (sigmoid/power with w calibration and limit price), budget-constrained (direction-only), and reserve (mixture-CDF via EF3M). You get a signed, bounded position series with documented conditions of use.

This article will cover angle-based operations. We will look at methods for constructing angles and using them in trading.

The article presents a complete exploration of the Nelder-Mead method, explaining how the simplex (function parameter space) is modified and rearranged at each iteration to achieve an optimal solution, and describes how the method can be improved.

Trading across multiple currencies is not available by default when an expert advisor is assembled via the wizard. We examine 2 possible hacks traders can make when looking to test their ideas off more than one symbol at a time.

The article explores the revolutionary integration of large language models (LLMs) with the MetaTrader 5 trading platform, where AI does not simply predict prices but makes autonomous trading decisions by analyzing market context much like an experienced trader. The author highlights a fundamental difference between LLMs and classical machine learning models such as CatBoost — the ability to engage in metacognition and self-reflection, which allows the system to learn from its own mistakes and improve its strategy.

We continue to study time series forecasting models. In this article, we get acquainted with a complex algorithm built on the use of a pre-trained language model.

In this article we delve into the application of boostrapping techniques as a means to estimate the future performance of an automated strategy.

In this article, I will continue the development of TabControl and implement the arrangement of tab headers on all four sides of the control for all modes of setting the size of headers: Normal, Fixed and Fill To Right.

The Expert Advisor by Ge Senlin (yyy999) from China got featured in the top ten of the Automated Trading Championship 2011 in late October and hasn't left it since then. Not often participants from the PRC show good results in the Championship - Forex trading is not allowed in this country. After the poor results in the previous year ATC, Senlin has prepared a new multicurrency Expert Advisor that never closes loss positions and uses position increase instead. Let's see whether this EA will be able to rise even higher with such a risky strategy.

How to apply predictive rules of supermarket retail analytics to the real Forex market? How are purchases of cookies, milk and bread related to stock exchange transactions? The article discusses an innovative approach to algorithmic trading based on the use of association rules.

This discussion covers the updated TradeManagementPanel in our New_Admin_Panel EA. The update enhances the panel by using built-in classes to offer a user-friendly trade management interface. It includes trading buttons for opening positions and controls for managing existing trades and pending orders. A key feature is the integrated risk management that allows setting stop loss and take profit values directly in the interface. This update improves code organization for large programs and simplifies access to order management tools, which are often complex in the terminal.

Self-supervised learning can be an effective way to analyze large amounts of unlabeled data. The efficiency is provided by the adaptation of models to the specific features of financial markets, which helps improve the effectiveness of traditional methods. This article introduces an alternative attention mechanism that takes into account the relative dependencies and relationships between inputs.