Gamuchirai Zororo Ndawana / Perfil

This article revisits the classic moving average crossover strategy and examines why it often fails in noisy, fast-moving markets. It presents five alternative filtering methods designed to strengthen signal quality and remove weak or unprofitable trades. The discussion highlights how statistical models can learn and correct the errors that human intuition and traditional rules miss. Readers leave with a clearer understanding of how to modernize an outdated strategy and of the pitfalls of relying solely on metrics like RMSE in financial modeling.

This article shows how to configure a black-box model to automatically uncover strong trading strategies using a data-driven approach. By using Mutual Information to prioritize the most learnable signals, we can build smarter and more adaptive models that outperform conventional methods. Readers will also learn to avoid common pitfalls like overreliance on surface-level metrics, and instead develop strategies rooted in meaningful statistical insight.

This article demonstrates how to automatically identify potentially profitable trading strategies using MetaTrader 5. White-box solutions, powered by unsupervised matrix factorization, are faster to configure, more interpretable, and provide clear guidance on which strategies to retain. Black-box solutions, while more time-consuming, are better suited for complex market conditions that white-box approaches may not capture. Join us as we discuss how our trading strategies can help us carefully identify profitable strategies under any circumstance.

All algorithmic trading strategies are difficult to set up and maintain, regardless of complexity—a challenge shared by beginners and experts alike. This article introduces an ensemble framework where supervised models and human intuition work together to overcome their shared limitations. By aligning a moving average channel strategy with a Ridge Regression model on the same indicators, we achieve centralized control, faster self-correction, and profitability from otherwise unprofitable systems.

This article helps new community members search for and discover their own candlestick patterns. Describing these patterns can be daunting, as it requires manually searching and creatively identifying improvements. Here, we introduce the engulfing candlestick pattern and show how it can be enhanced for more profitable trading applications.

In this discussion, we focus on how we can break the glass ceiling imposed by classical machine learning techniques in finance. It appears that the greatest limitation to the value we can extract from statistical models does not lie in the models themselves — neither in the data nor in the complexity of the algorithms — but rather in the methodology we use to apply them. In other words, the true bottleneck may be how we employ the model, not the model’s intrinsic capability.

Linear system identifcation may be coupled to learn to correct the error in a supervised learning algorithm. This allows us to build applications that depend on statistical modelling techniques without necessarily inheriting the fragility of the model's restrictive assumptions. Classical supervised learning algorithms have many needs that may be supplemented by pairing these models with a feedback controller that can correct the model to keep up with current market conditions.

In this discussion, we contrast the classical approach to time series cross-validation with modern alternatives that challenge its core assumptions. We expose key blind spots in the traditional method—especially its failure to account for evolving market conditions. To address these gaps, we introduce Effective Memory Cross-Validation (EMCV), a domain-aware approach that questions the long-held belief that more historical data always improves performance.

Trading strategies may be challenging to improve because we often don’t fully understand what the strategy is doing wrong. In this discussion, we introduce linear system identification, a branch of control theory. Linear feedback systems can learn from data to identify a system’s errors and guide its behavior toward intended outcomes. While these methods may not provide fully interpretable explanations, they are far more valuable than having no control system at all. Let’s explore linear system identification and observe how it may help us as algorithmic traders to maintain control over our trading applications.

In this series of articles, we look at the challenges faced by algorithmic traders when deploying machine-learning-powered trading strategies. Some challenges within our community remain unseen because they demand deeper technical understanding. Today’s discussion acts as a springboard toward examining the blind spots of cross-validation in machine learning. Although often treated as routine, this step can easily produce misleading or suboptimal results if handled carelessly. This article briefly revisits the essentials of time series cross-validation to prepare us for more in-depth insight into its hidden blind spots.

This article walks the reader through a reimagined version of the classical Bollinger Band breakout strategy. It identifies key weaknesses in the original approach, such as its well-known susceptibility to false breakouts. The article aims to introduce a possible solution: the Double Bollinger Band trading strategy. This relatively lesser known approach supplements the weaknesses of the classical version and offers a more dynamic perspective on financial markets. It helps us overcome the old limitations defined by the original rules, providing traders with a stronger and more adaptive framework.

Machine learning is often viewed through statistical or linear algebraic lenses, but this article emphasizes a geometric perspective of model predictions. It demonstrates that models do not truly approximate the target but rather map it onto a new coordinate system, creating an inherent misalignment that results in irreducible error. The article proposes that multi-step predictions, comparing the model’s forecasts across different horizons, offer a more effective approach than direct comparisons with the target. By applying this method to a trading model, the article demonstrates significant improvements in profitability and accuracy without changing the underlying model.

Preprocessing is a powerful yet quickly overlooked tuning parameter. It lives in the shadows of its bigger brothers: optimizers and shiny model architectures. Small percentage improvements here can have disproportionately large, compounding effects on profitability and risk. Too often, this largely unexplored science is boiled down to a simple routine, seen only as a means to an end, when in reality it is where signal can be directly amplified, or just as easily destroyed.

This article takes a fresh perspective on a hidden, geometric source of error that quietly shapes every prediction your models make. By rethinking how we measure and apply machine learning forecasts in trading, we reveal how this overlooked perspective can unlock sharper decisions, stronger returns, and a more intelligent way to work with models we thought we already understood.

Financial markets are unpredictable, and trading strategies that look profitable in the past often collapse in real market conditions. This happens because most strategies are fixed once deployed and cannot adapt or learn from their mistakes. By borrowing ideas from control theory, we can use feedback controllers to observe how our strategies interact with markets and adjust their behavior toward profitability. Our results show that adding a feedback controller to a simple moving average strategy improved profits, reduced risk, and increased efficiency, proving that this approach has strong potential for trading applications.

Human traders had long participated in financial markets before the rise of computers, developing rules of thumb that guided their decisions. In this article, we revisit a well-known breakout strategy to test whether such market logic, learned through experience, can hold its own against systematic methods. Our findings show that while the original strategy produced high accuracy, it suffered from instability and poor risk control. By refining the approach, we demonstrate how discretionary insights can be adapted into more robust, algorithmic trading strategies.

This article explores the powerful role of matrix factorization in algorithmic trading, specifically within MQL5 applications. From regression models to multi-target classifiers, we walk through practical examples that demonstrate how easily these techniques can be integrated using built-in MQL5 functions. Whether you're predicting price direction or modeling indicator behavior, this guide lays a strong foundation for building intelligent trading systems using matrix methods.

In this discussion, we will set the foundation for using powerful linear, algebra tools that are implemented in the MQL5 matrix and vector API. For us to make proficient use of this API, we need to have a firm understanding of the principles in linear algebra that govern intelligent use of these methods. This article aims to get the reader an intuitive level of understanding of some of the most important rules of linear algebra that we, as algorithmic traders in MQL5 need,to get started, taking advantage of this powerful library.

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, we continue our exploration of building an ensemble of trading strategies and using the MT5 genetic optimizer to tune the strategy parameters. Today, we analyzed the data in Python, showing our model could better predict which strategy would outperform, achieving higher accuracy than forecasting market returns directly. However, when we tested our application with its statistical models, our performance levels fell dismally. We subsequently discovered that the genetic optimizer unfortunately favored highly correlated strategies, prompting us to revise our method to keep vote weights fixed and focus optimization on indicator settings instead.