Articles on data analysis and statistics in MQL5

This article implements a path-dependent volatility model in MQL5 that keeps the sign of returns. The engine uses two-timescale EWMA kernels (four accumulators) to form activity and trend features, with closed-form coefficients and a lightweight kernel search. We provide a calibration workflow and an indicator that plots the decomposition, then test out of sample against EWMA and GARCH on multiple symbols. Results show clear gains on equity indicators and minimal gains on FX.

This article implements the Probability of Backtest Overfitting in MQL5 with Combinatorially Symmetric Cross-Validation. It shows how to build the bar-by-bar returns matrix of a parameter grid, aggregate it by blocks and rank the in-sample winner out-of-sample across every split of the history. You get a single probability that this winner ends up in the lower half of the grid out-of-sample, along with the MQL5 code to compute it on your own optimizations.

Although many people believe that we can easily embed SQL code into other code, this is usually not the case. The reason is that the SQL code is included in the executable as a string. Embedding SQL code as a string, while not a problem for small snippets, can ultimately cause us quite a few headaches.

In the previous article, we looked at an example of implementing a class in MQL5 to provide basic support. Its purpose is specifically to allow SQL code to be stored in a separate script file. This way, we won't need to write the same SQL statements as string literals in the MQL5 source code. Although this solution is functional, there are some aspects of it that we can and should improve.

The article presents CPdfBuffer and CPdfWriter, a minimal MQL5 pair that derives all offsets from the output buffer instead of hand-maintaining numbers. CPdfWriter records object starts, writes a fixed-width cross-reference table and trailer, and ensures structural consistency. The result passes Part 2's validator, and a ready-to-use script converts arbitrary text to a paginated PDF.

The article examines the architecture of an Expert Advisor based on a cellular automaton with 10,000 adaptive parameters and an independent binary predictor with a 10-bar prediction horizon. The article demonstrates a three-level online learning process, the evolution of strategies, and validation using a ring buffer and a confusion matrix. The input parameters are reduced to Magic Number, while trading settings are calculated from ATR and five genomes. The EURUSD H1 test yielded an approximate hit rate of 58%, compared to roughly 51% for the fixed MLP.

The article presents CPdfBuffer and CPdfWriter, a minimal MQL5 pair that derives all offsets from the output buffer instead of hand-maintaining numbers. CPdfWriter records object starts, writes a fixed-width cross-reference table and trailer, and ensures structural consistency. The result passes Part 2's validator, and a ready-to-use script converts arbitrary text to a paginated PDF.

Learn how to create advanced visualizations in MetaTrader 5 using a Matplotlib-like API from MQL5 while Python handles rendering. We implement key chart types—scatter, bar, horizontal bar, histogram, pie, and hexbin—and show how to control axes, legends, colorbars, and styles. The result is a consistent workflow for exploring trading and financial datasets.