Resumos e ideias principais, escritos pelo agente de investigação da Stratmill, dos livros, artigos científicos, artigos e código consultados pelos nossos agentes de IA. Cada página inclui uma ligação para o original.
3,012 documentos
This article explains Fully Parameterized Quantile Functions (FQF), a distributional reinforcement-learning method for estimating reward distributions in Q-learning. Earlier approaches use fixed reward categories or predetermined quantiles, which require…
The article describes the Multi-Agent Self-Adaptive (MASA) framework for dynamic portfolio management. It assigns return generation to a TD3 reinforcement-learning agent, risk adjustment to a separate optimization-based controller, and market analysis to an…
This article explains how to build a C# and WPF add-on for launching Expert Advisor optimizations across multiple MetaTrader terminals on one computer. It reviews command-line launch arguments and INI configuration files, which let a terminal start with…
This installment describes an MQL5 workflow for optimizing trading strategy groups and making updated groups available to a running Expert Advisor through a separate EA database. The proposed database avoids recompiling the final EA whenever optimization…
The article explains how to build a MetaTrader 5 Expert Advisor that places a draggable risk dashboard on the chart. Using the platform’s Standard Library controls and CTrade, the panel displays floating profit and offers bulk closures for all filtered…
The article describes a live drift monitor for an ONNX trading model running in MetaTrader 5. It compares a frozen reference sample with a rolling live sample for each input feature, using the one-dimensional Wasserstein-1 distance: sort equal-sized samples,…
The article proposes a trading signal that compares normalized price slope with an indicator slope over a lookback window, then treats their difference as structural divergence. It combines this measure with a three-horizon attention proxy intended to weight…
The article explores using neural network activation functions to build custom criteria for ranking strategy parameter sets in MetaTrader optimization. It argues that returning zero for an undesirable pass can disrupt genetic search, while unrestricted…
This article surveys preparation and selection of input variables for machine learning, with examples aimed at financial data. It recommends handling missing observations, checking near-constant predictors, reducing problematic correlations where…
The article explains how Twin Delayed Deep Deterministic Policy Gradient (TD3) modifies DDPG to reduce overestimation of learned Q-values. Its three core changes are training two Critics and using the lower estimate, updating the Actor less often, and…
This document explains the MetaTrader 5 side of a three-program market-simulation setup linking Excel, a Python server, and MetaTrader through sockets. Its central design choice is to run the MQL5 component as a service, which can operate without being…
This installment explains how to connect Excel, a Python socket server, and MetaTrader 5 as parts of a market-simulation workflow. It recommends running the MetaTrader component as an MQL5 service rather than a chart-bound script, allowing it to operate…
This article presents TimeFound, a Transformer framework for forecasting time series across domains, and discusses how it can inform systematic trading. The model represents data at multiple time scales by splitting histories into patches of different…
This article explores pairing the Awesome Oscillator (AO), a momentum and trend indicator, with Envelopes, percentage bands around a moving average that can mark stretched prices. It outlines reversal ideas such as a false break beyond a band followed by a…
This tutorial demonstrates a compact market-data learning pipeline using MQL5's ALGLIB data-analysis library. It converts OHLC bars into five numerical features—one-bar return, candle-body ratio, range ratio, and two momentum measures—and labels each…
This article applies the category-theory idea of a natural transformation to time-series forecasting with differently structured datasets. One dataset contains a moving-average value; another contains the constituent lagged prices. The author treats the…
This article continues an MQL5 implementation of CATCH, a neural approach to anomaly detection in multivariate time series. CATCH transforms series with the Fourier transform, divides the spectrum into frequency patches, and uses complex-valued convolutions…
This tutorial shows how to build a basic neural-network trading Expert Advisor by connecting MetaTrader 5 with NeuroPro, a multilayer network tool. It exports EURUSD hourly closing prices to CSV, converts the data to DBF for NeuroPro, trains and tests a…
This article introduces Deep Deterministic Policy Gradient (DDPG) as a reinforcement-learning approach for trading agents that must choose continuous action values. Instead of expanding a discrete action list for every combination of direction, volume, stop…
This article examines filters as rules that suppress some of an automated strategy’s signals, dividing them into bandpass filters that select conditions over a range and discrete filters that accept signals according to a pattern. It demonstrates the process…
This article explains how to define custom optimization scores for MetaTrader 5 Expert Advisors through the tester’s OnTester mechanism. It outlines built-in genetic optimization criteria and shows how a custom score can combine outcomes such as profit,…
This guide surveys MetaTrader 4’s built-in indicators and explains their categories, display defaults, and configurable properties. It distinguishes oscillators, trend indicators, Bill Williams indicators, and volume measures, then describes how indicators…
This MQL5 cookbook chapter extends an Expert Advisor so it can retrieve position details that are not available from current position properties alone. By selecting the account's deal history for a position, the program can find the first and most recent…
The article describes a two-wave continuation setup: a main move followed by a correction, with an anticipated resumption in the main trend's direction. It identifies swing points using confirmed fractals and provisional extrema, then applies structural…