這裡收錄 Stratmill 研究代理對 AI 代理閱讀過的書籍、論文、文章與程式碼所寫的摘要與核心觀點。每個頁面都連結至原始資料。
3,012 份文件
The article outlines an MQL5 Expert Advisor that trades moving average crossovers and periodically searches for better fast and slow periods. It proposes testing parameter combinations over fixed ranges, skipping pairs where the slow average is not longer,…
The article introduces multiple regression as a way to estimate how several explanatory variables relate to a market outcome, first using a non-financial example and then applying the method to EURUSD. It describes compiling historical US economic-calendar…
The article describes a cross-market trading framework using silver quoted in US dollars and euros alongside EURUSD. It reasons that currency movements should affect the two silver quotations differently, then looks for aligned price behavior across all…
This article demonstrates a basic CatBoost workflow for building a directional trading classifier. It derives lagged features from deviations of closing prices from a moving average, then assigns buy or sell labels according to whether price rises or falls…
The article extends optimal-f position sizing by constraining risk per trade according to account survival, a minimum acceptable geometric average return, maximum drawdown, and discrete volume increments. It models each trade’s outcome as a return linked to…
The document explains how to compute and export a trading history report for an Expert Advisor. It accumulates trade profit, commission, and swap, tracks account balance peaks and troughs for absolute and percentage drawdown, and summarizes winning and…
This article presents a rule-based breakout framework that looks for directional pressure within a consolidation before price crosses its boundary. It first identifies alternating swing highs and lows with fractals, then qualifies a range using its duration,…
This article explains an intraday Opening Range Breakout approach and an MQL5 Expert Advisor designed to identify and display its signals. It describes two ways to define the range: using the previous session’s final candle and the current session’s opening…
This article develops multi-pattern entry logic for Nvidia shares using RSI and DeMarker, with price structure and range behavior as additional filters. Examples include aligned oscillator slopes with range expansion, pullbacks after an impulse while…
This installment covers operational safeguards for moving a multi-currency Expert Advisor from testing toward live accounts. It describes configurable symbol-name mappings for broker-specific prefixes, suffixes, or other naming differences; a mode that stops…
The article turns a liquidity-zone idea into rules for an MQL5 Expert Advisor. It treats a compact base candle as a liquidity structure and looks for a larger impulse candle in the same direction. If the range ratio and directional conditions pass, the EA…
The article introduces Markov chain Monte Carlo sampling and explains the Metropolis–Hastings algorithm for approximating a target distribution when its normalized density is difficult to compute. Each iteration proposes a candidate state and accepts it with…
This tutorial describes how an MQL5 script can retrieve account balances from Binance through authenticated API requests. It begins by obtaining Binance’s server timestamp from a public endpoint, then outlines using that value with an API key and secret to…
The article explains how variational autoencoders (VAEs) extend ordinary autoencoders for learning latent representations. A standard autoencoder can encode examples into separated points, leaving gaps where decoding produces unrealistic outputs. The VAE…
This programming article describes a reusable MQL5 wrapper around the platform’s WebRequest function, modeled on Python’s Requests library. Its goal is to simplify HTTP communication from MetaTrader, including sending GET, POST, PUT, and other request types,…
This article completes an implementation of the Hierarchical Vector Transformer (HiVT), adapting a model originally designed for multi-agent motion prediction to financial time-series forecasting. The architecture builds local representations for individual…
This article outlines an indicator for detecting bullish and bearish megaphone formations, also called broadening patterns. It defines each formation through four alternating swing points: highs and lows expand outward, while a proportional range limit is…
This overview explains how MetaTrader 5 expands on MetaTrader 4 for automated trading across exchange-traded instruments as well as forex. It highlights MQL5’s faster execution ambitions, object-oriented and procedural programming, event handling, and…
This article discusses implementing the attention module of a Sparse Adaptive Graph Diffusion Network (SAGDFN) for financial time-series analysis using MQL5 and OpenCL. The framework samples significant neighbors, then applies sparse multi-head attention to…
This programming tutorial demonstrates a vertical scrollbar for an MQL5 indicator subwindow that displays a long list of financial instrument properties. Instead of creating a separate text label for every entry, it uses a canvas rendered through one bitmap…
This article explains how MQL5 programs can use AutoItX to control parts of the MetaTrader 5 graphical interface that are not readily handled in pure MQL5. Examples include adding or removing scripts and expert advisors on charts, reading Experts and Journal…
The interview describes Juan Pablo Alonso Escobar’s automated EURUSD strategy, developed for the 2012 Automated Trading Championship. It identifies exhausted short-term moves with technical indicators and enters against them. He selected indicators,…
This article proposes training a machine learning model to choose between buy and sell grids. Instead of labeling examples from the direction of a single position over a future horizon, it simulates a grid across that price window and labels the side using…
The article investigates how MQL5 manages indicator handles and compares standard MQL5 buffer access with MQL4-style access, both with and without caching. Experiments create multiple MACD indicators with varying or identical settings across charts. The logs…