Articles with examples of trading robots developed in MQL5

One of the key problems within reinforcement learning is environmental exploration. Previously, we have already seen the research method based on Intrinsic Curiosity. Today I propose to look at another algorithm: Exploration via Disagreement.

In the previous article, we introduced the multi-agent self-adaptive framework MASA, which combines reinforcement learning approaches and self-adaptive strategies, providing a harmonious balance between profitability and risk in turbulent market conditions. We have built the functionality of individual agents within this framework. In this article, we will continue the work we started, bringing it to its logical conclusion.

In this series of articles, we revisit classic strategies to see if we can improve them using AI. In today's article, we will examine the popular strategy of multiple time-frame analysis to judge if the strategy would be enhanced with AI.

In this article we are implementing ARIMA forecasting indicator in MQL5. It examines how the ARIMA model generates forecasts, its applicability to the Forex market and the stock market in general. It also explains what AR autoregression is, how autoregressive models are used for forecasting, and how the autoregression mechanism works.

The article describes a variant of options emulation through an underlying asset implemented in the MQL5 programming language. The pros and cons of the chosen approach are compared with real exchange options using the example of the FORTS futures market of the MOEX Moscow exchange and the Bybit crypto exchange.

The use of anisotropic diffusion processes for encoding the initial data in a hyperbolic latent space, as proposed in the HypDIff framework, assists in preserving the topological features of the current market situation and improves the quality of its analysis. In the previous article, we started implementing the proposed approaches using MQL5. Today we will continue the work we started and will bring it to its logical conclusion.

The problem of reinforcement learning lies in the need to define a reward function. It can be complex or difficult to formalize. To address this problem, activity-based and environment-based approaches are being explored to learn skills without an explicit reward function.

How does portfolio trading work on Forex? How can Markowitz portfolio theory for portfolio proportion optimization and VaR model for portfolio risk optimization be synthesized? We create a code based on portfolio theory, where, on the one hand, we will get low risk, and on the other, acceptable long-term profitability.

In this article, we prepare our MQL5 trading system for strategy testing by embedding economic calendar data as a resource for non-live analysis. We implement event loading and filtering for time, currency, and impact, then validate it in the Strategy Tester. This enables effective backtesting of news-driven strategies.

In the previous article, we implemented the Soft Actor-Critic algorithm, but were unable to train a profitable model. Here we will optimize the previously created model to obtain the desired results.

This article replaces binary swing validation with a volatility‑normalized pullback model. Retracement depth is measured as a ratio of the prior impulse and calibrated to a rolling ATR regime, while entries require a minimum quality score and confirmation by structure or liquidity signals. The five‑layer design integrates detection, validation, liquidity mapping, regime‑aware scoring, and execution, helping you filter weak corrections and size stops dynamically to current conditions.

In the previous article, we introduced the DDPG method, which allows training models in a continuous action space. However, like other Q-learning methods, DDPG is prone to overestimating Q-function values. This problem often results in training an agent with a suboptimal strategy. In this article, we will look at some approaches to overcome the mentioned issue.

SAMformer offers a solution to the key drawbacks of Transformer models in long-term time series forecasting, such as training complexity and poor generalization on small datasets. Its shallow architecture and sharpness-aware optimization help avoid suboptimal local minima. In this article, we will continue to implement approaches using MQL5 and evaluate their practical value.

The previously developed risk manager contained only basic functionality. Let's try to consider possible ways of its development, allowing us to improve trading results without interfering with the logic of trading strategies.

In this article, we explore how previously invalidated orderblocks can be reused as mitigation blocks within Smart Money Concepts (SMC). These zones reveal where institutional traders re-enter the market after a failed orderblock, providing high-probability areas for trade continuation in the dominant trend.

We have already made some progress in developing a multi-currency EA with several strategies working in parallel. Considering the accumulated experience, let's review the architecture of our solution and try to improve it before we go too far ahead.

In this series article, we will empirically analyze classic trading strategies to see if we can improve them using AI. In today's discussion, we tried to predict higher highs and lower lows using the Linear Discriminant Analysis model.

In this article, we continue the implementation of the approaches of the ATFNet model, which adaptively combines the results of 2 blocks (frequency and time) within time series forecasting.

We continue to implement approaches proposed vy the authors of the DUET framework, which offers an innovative approach to time series analysis, combining temporal and channel clustering to uncover hidden patterns in the analyzed data.

This article enhances our Economic Calendar dashboard by implementing real-time news updates to keep market information current and actionable. We integrate live data fetching techniques in MQL5 to update events on the dashboard continuously, improving the responsiveness of the interface. This update ensures that we can access the latest economic news directly from the dashboard, optimizing trading decisions based on the freshest data.

Factorization is a mathematical process used to gain insights into the attributes of data. When we apply factorization to large sets of market data — organized in rows and columns — we can uncover patterns and characteristics of the market. Factorization is a powerful tool, and this article will show how you can use it within the MetaTrader 5 terminal, through the MQL5 API, to gain more profound insights into your market data.

In this article, triangular arbitrage is presented as a problem of finding cycles in a directed graph, where the vertices are currencies and the edges are currency pairs with weight rates. Profitable cycle: product of weights >1. Our Floyd-Warshall and DFS algorithms find optimal currency exchange paths that return to the starting point with a profit.

We continue studying the environment in reinforcement learning models. And in this article we will look at another algorithm – Go-Explore, which allows you to effectively explore the environment at the model training stage.

The authors of the FreDF method experimentally confirmed the advantage of combined forecasting in the frequency and time domains. However, the use of the weight hyperparameter is not optimal for non-stationary time series. In this article, we will get acquainted with the method of adaptive combination of forecasts in frequency and time domains.

A large number of the models we have reviewed so far are based on the Transformer architecture. However, they may be inefficient when dealing with long sequences. And in this article, we will get acquainted with an alternative direction of time series forecasting based on state space models.

Today, we delve into incorporating useful trading metrics within a specialized window integrated into the Admin Panel EA. This discussion focuses on the implementation of MQL5 to develop an Analytics Panel and highlights the value of the data it provides to trading administrators. The impact is largely educational, as valuable lessons are drawn from the development process, benefiting both upcoming and experienced developers. This feature demonstrates the limitless opportunities this development series offers in equipping trade managers with advanced software tools. Additionally, we'll explore the implementation of the PieChart and ChartCanvas classes as part of the continued expansion of the Trading Administrator panel’s capabilities.

A multi-task learning framework based on ResNeXt optimizes the analysis of financial data, taking into account its high dimensionality, nonlinearity, and time dependencies. The use of group convolution and specialized heads allows the model to effectively extract key features from the input data.