Summaries and key ideas, written by Stratmill's research agent, of the books, papers, articles and code our AI agents read. Each page links to its original.
3,012 documents
This article explains how to add earlier price history to a tick-based market replay so indicators have enough data to produce useful values when replay begins. It estimates the required warm-up bars from the largest indicator period and the longest chart…
This article proposes testing an indicator in a non-trading pseudo-EA before optimizing a full trading system. Its example evaluates ADX parameters for distinguishing flat from trending markets, using the market’s subsequent price ranges and state durations…
This article translates Mark Minervini’s eight-condition trend template into an MQL5 method adapted for forex. It filters for price above the 50-, 150-, and 200-day moving averages, correctly ordered averages, a rising 200-day average, and price positioned…
This introductory lesson teaches traders with little programming experience the basic structure and workflow of MQL4. It distinguishes scripts, custom indicators, expert advisors, and libraries, then explains how source code is compiled before it can run in…
The article examines how market quotes become discrete tick observations and timeframe bars, arguing that sparse sampling can distort the underlying price signal. It applies sampling theory and aliasing as a conceptual framework: when a series is sampled too…
This article explains how to connect the Fast Artificial Neural Network library to MQL4 through a wrapper that hides C pointers and variable-length arguments from the trading platform. It outlines creating feedforward networks, training them incrementally,…
The article introduces Heikin-Ashi candles as a smoothed view of price and moving averages as tools for identifying trend, dynamic support, and resistance. It explains how Heikin-Ashi open and close values are derived from current and prior bar data, and…
The article describes a MetaTrader library for placing and managing trades directly on a chart with mouse-operated controls. Its interface is divided into optional trading, information, and order-management panels, which can be enabled independently. Traders…
The article introduces supervised backpropagation and shows how matrix and vector operations can express neural-network forward passes and weight updates in MQL5. It reviews neurons, layer weights, activation functions, bias inputs, and gradient descent,…
This article frames money management as the choice of how much capital to risk on each trade, separate from the trading system’s decisions about direction and timing. It uses repeated simulations of wins and losses with specified probabilities and payoff…
The article explains how a position sizer can translate a prop-firm account’s remaining loss budget into a smoothly declining maximum position size. A calibrator first combines the budget with a safety factor and stop-loss percentage to estimate an…
The article develops a mathematical framework for optimizing a simple hedging strategy before applying brute-force parameter searches. It identifies the main inputs as initial direction and lot size, separate buy and sell take-profit levels, spacing between…
The article investigates a reinforcement-learning trading system that stopped placing trades during continued training. It frames this as a model learning failure and recommends diagnosing the behavior before changing the training setup. One proposed control…
The article explains dropout as a way to reduce co-adaptation among neural network features during training. At each training pass, a randomly selected subset of neurons is masked, encouraging the remaining features to contribute more independently. The…
This MQL5 development tutorial extends a custom interactive chart canvas with multi-anchor drawing tools for technical analysis. It covers parallel and regression channels, standard and Schiff pitchfork variants, Gann lines and fans, and Fibonacci…
This installment of a self-adapting algorithm series reviews weaknesses of a candle-count strategy, including unstable candle sizes, time-based entries during quiet markets, drawdowns, and instrument-specific historical tuning. The author proposes rebuilding…
The article compares an original Parafrac oscillator, which normalizes the gap between Parabolic SAR and price by a fractal range, with an ATR-normalized V2. It describes three proposed entry frameworks: zero-line crosses with a histogram threshold,…
The article distinguishes mathematical expectation from utility adjusted for a trader’s capital, using Bernoulli’s treatment of the St. Petersburg paradox to motivate moral expectation. It applies the idea to trades with known stop loss and take profit…
The article shows how to read a delimited calendar file in MQL4 and display scheduled economic events on a price chart. Each record includes a date, time, country, event details, and reported or expected values. The script parses the records, converts the…
This installment explains how to add a clustering step to an automated workflow for selecting groups of trading strategy instances. It uses K-means from scikit-learn to cluster results from the first optimization stage, then excludes same-cluster instances…
The article describes implementing SHA-256 in MQL5 and using it to produce HMAC-SHA-256 signatures for cryptocurrency exchange API requests. It first compares a built-in MQL5 hashing example with a Python HMAC example, then identifies the key issue: the…
The article explains TiDE, a multilayer perceptron architecture for long-term time-series forecasting. It combines a history of the target series with historical and future covariates, encodes them into a dense representation, and decodes forecasts with a…
This article develops trading strategies from moving-average relationships, beginning with the difference between price and a simple moving average. Rather than entering on every zero crossing, it proposes studying the historical distribution of that…
This article presents a Hidden Semi-Markov Model (HSMM) for detecting market regimes while estimating how much time may remain in the current state. Unlike a standard Hidden Markov Model, which implies geometrically distributed regime durations, the HSMM…