Articles on data analysis and statistics in MQL5

A market-neutral trading strategy based on the empirical return distribution offers an alternative to traditional technical analysis methods, replacing price direction forecasting with the statistical placement of orders at levels the price is likely to reach. This article provides a detailed analysis of the mathematical framework for calculating percentiles, algorithms for weighting position sizes based on the probability of an order being triggered, and mechanisms for adapting to changing market conditions through grid expiration. A complete implementation in MQL5 is provided.

The Crow Search Algorithm (CSA) is an elegant metaheuristic inspired by crows’ ability to hide food and find other crows' caches, solving optimization problems by balancing following successful solutions with random exploration of the search space. Let's find out how well the algorithm performs.

Learn to assemble an MT5 Expert Advisor that hosts a chart management dashboard written in MQL5. The guide walks through shared definitions, symbol acquisition and filtering, chart lifecycle functions, and a UI panel with search, scrolling, and state indicators, all driven by events and a timer. The result is a reproducible tool that reduces clicks and accelerates multi-symbol analysis.

This article demonstrates a practical approach to researching trading ideas using a range breakout strategy as an example. We will go through the entire process, from formalizing trading rules and building a baseline model to parameter optimization, forward testing, and evaluating the robustness of the results. The main goal of the article is to develop an understanding of how statistics and testing can be used to identify, validate, and evaluate trading hypotheses.

Stop-loss and take-profit placement is usually the least-measured decision in a trading system. This Expert Advisor reads your closed history, replays M1 price between each entry and exit to measure Maximum Adverse and Favorable Excursion per trade, and splits winners from losers. From the distributions and trade efficiency it derives data-driven stop and target levels - measured from your own account, not a rule of thumb. Analysis only; it does not trade.

Symbolic Aggregate approXimation (SAX) encodes price windows as short words to enable fast, sound similarity search on history. We implement SAX in pure MQL5, including Gaussian breakpoints, PAA, and the lower-bounding MINDIST, and validate it with a test harness. An indicator applies a no-lookahead, two-stage search, summarizes forward paths in ATR units, and draws a forecast fan, explicitly indicating when the sample shows no edge.

A statistically significant backtest is not proof of a robust edge. This article presents a three-part validation battery in Python that consumes an MQL5 trade-level CSV export. A sign-randomization permutation test evaluates whether the Sortino reflects real directional skill, bootstrap BCa intervals assess metric stability, and Monte Carlo trade-order shuffling tests sequence dependence of drawdowns. The results feed a five-condition framework for deployment decisions.

A new optimization method, the Bison Algorithm (BIA), uses two strategies, inspired by the behavior of bison, for solving continuous problems with a single objective function. The key features of BIA are two fundamental principles borrowed from the behavior of bison: the ability to move dynamically and a defensive strategy.

In this article we make the case for merging Divergence Mapping with a Temporal Fusion Proxy in a Trading Robot. Rather than depending on lagging price confirmations, the Divergence Mapping's thesis is that acting like a structural sensor can help identify hidden momentum shifts from price action and indicator anomalies. To establish how these anomalies are interpreted over time we use a Temporal Fusion Transformer proxy. This network incorporates historical context to weigh developing trends such that merging it with Divergence Mapping should set us up to spot shifts in accumulation and distribution before price breakouts.

The article introduces the Mapper pipeline in MQL5 by implementing the two fundamental components: CTDAMapperFilter (lens) and CTDAMapperCover (overlapping intervals). It explains three lens options—eccentricity, density, and coordinate—plus cover parameters (resolution and gain), and demonstrates how a price point cloud is reduced to one value per point and interval memberships. Readers obtain ready inputs for subsequent clustering and graph construction.

The article describes an interesting approach to algorithmic trading based on symbolic mathematical equations instead of traditional machine learning "black boxes". The author demonstrates how to transform opaque neural networks into readable mathematical equations using the SymPy library and polynomial regression, allowing for a full understanding of the logic behind trading decisions. The approach combines the computational power of ML with the transparency of classical methods, giving traders the ability to analyze, adjust, and adapt models in real time.

In this article we make the case for pairing the Discrete Fourier Transform with a Spiking Neural Network in a Trading Robot. The Fourier Transform helps represent data as oscillations instead of its raw values. To govern how we interpret these cycles, we engage a Spiking Neural Network that unlike regular networks, uses time dependent electrical charges to accumulate potential and only "spike" when a target threshold is met. Combining these two engines allows us better control on the timing of discrete market movements, that in theory should give us entry signals with rigorous mathematical confirmation.

The custom modification of the Dingo algorithm presented in the article has raised the bar for finding the best optimization algorithm. Are even better results possible?

The article quantifies correlation and portfolio risk in MetaTrader 5: from time-aligned returns to a covariance matrix, true portfolio variance against the independent-sum assumption, and position-level risk attribution. A MetaTrader 5 service runs in the background, shows the metrics on a small chart panel, and pushes alerts when risk thresholds are crossed. Source code is provided for an example script, a reusable risk engine class, and the service.

Backtests often understate spread, commission, and slippage. This MQL5 analyzer loads closing deals and simulates rising execution costs to measure robustness. It computes the breakeven cost per deal, the cushion over an assumed cost, the net profit and profit factor at that cost, and how many winners turn into losers, then summarizes the result with an A+ to F grade and targeted guidance.

The article presents a step-by-step development of a multi-threaded trading robot with machine learning in Python and MetaTrader 5. The system architecture is considered – from data collection and creation of technical indicators to training XGBoost models with portfolio risk management. The implementation of data augmentation, feature clustering via Gaussian Mixture Models, and flow coordination for parallel trading of multiple currency pairs is described in detail.

This article presents CSymbolMetaCache, an MQL5 layer that preloads contract specifications and trading-session schedules for monitored symbols at EA startup and then serves typed getters from memory. It explains which properties are safe to cache versus dynamic ones, including the semi-dynamic tick value on cross-currency pairs, and implements an in-memory IsMarketOpen() evaluator. A benchmark quantifies latency reduction across a set of twenty symbols.

An MQL5 script reconstructs closed trades from raw deal history and replays them on the chart bar by bar, drawing entry, exit, stop, target, and an annotation with per‑trade statistics. Four classes separate concerns: a trade data record, history reconstruction with a two‑pass SL/TP lookup and partial‑close aggregation, chart rendering, and a controller with polling‑based keyboard navigation. This enables consistent, fast visual review of each trade in its original candlestick context.

The article presents a new metaheuristic method based on the hunting strategies of Australian dingoes: group attack, chase, and scavenging. Let's see how the Dingo Optimization Algorithm (DOA) performs algorithmically.

Net profit and win rate tell you how much a strategy made, not how the result is distributed. This article builds a native MQL5 script that reads your closed trades and measures profit concentration: the top-N trade share, the Gini coefficient of the winners, an outlier-dependence stress test that removes the best few winners, and the largest day against a prop-firm consistency limit. It combines these into one A+ to F score with recommendations, running inside MetaTrader 5.

This article describes a prototype reusable market structure framework for MQL5, built with a clean modular architecture and an internal event queue. It shows how to detect swing points, classify break-of-structure and change-of-character events, maintain a deterministic market state, and persist data to CSV. The focus is entirely on software engineering, component separation, and extensibility, not on trading signals. The prototype is a foundation for further development, not a production-ready library.

A reproducible, read-only Python audit for MetaTrader 5 that verifies history quality before any backtest. It exports M5 data from multiple terminals, detects gaps and synthetic bars by timestamp spacing, and reports coverage per year. The same deterministic strategy then runs on three broker feeds over a common window to quantify result drift and decompose it into spread, data/price, and trade effects.

Part 8 adds bar-by-bar micro-trend scoring for NQ M1. GetMicroTrendStrength() builds a continuous [-1, +1] composite from EMA alignment, ATR‑normalized price position, slope consistency, and volume, with a contradiction penalty to suppress alignment/price conflicts. Session-adaptive thresholds scale by Part 7 confidence to modulate signal frequency across regimes. Outputs include a seven-state label, a binary signal, and a persistence check, calibrated on 514 New York sessions (May 2024–May 2026).

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 describe an MQL5 framework that aligns entries with session rhythms and scheduled clock events. A script identifies the broker's time zone and DST by detecting NFP spikes on EURUSD and matching them to EU/US/AU transition dates, producing EA‑ready settings. Session-to-broker time conversion and 15-minute marks constrain execution. A multi‑timeframe AMA signal aggregates trends for strategy selection and optimization.

Despite what was shown in the previous article, all of this may seem simple at first. In reality, several problems remain, along with many tasks that still need to be completed. You, dear reader, may imagine that everything is easy and straightforward. Out of inexperience, you may simply accept whatever is presented to you. And that is a mistake you should try to avoid. Even worse is trying to use something without truly understanding what exactly you are using. Beginners often pass through a copy-and-paste stage. If you do not want to remain stuck at that stage forever, you should learn how to use certain tools. One of the tools most often used by programmers is documentation. The second is testing, supported by log files. Here we will see how to do this.

The article presents a complete implementation of TimeGPT, a specialized Transformer-based architecture for forecasting financial time series on the MetaTrader 5 platform. Adaptation of the attention mechanism to financial data, selective tokenization of price changes, hardware-aware optimizations, and advanced learning techniques are discussed. Included are practical testing results showing 87% forecast accuracy over a 24-bar horizon with a training time of 15 minutes on the CPU. We also present a ready-made trading EA with automatic retraining.