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.
5,018 documents
This article describes an Expert Advisor that represents market conditions as vertices in a state-transition graph. It discretizes normalized trend, volatility, and momentum measures, counts transitions between states, smooths and decays those counts, then…
The article develops a machine learning mean reversion system for EURGBP, covering trade labeling, market regime identification, model training, and deployment in MetaTrader 5. It contrasts random horizon labels with labels based on how far prices deviate…
The article describes a binary genetic algorithm as an evolutionary method for optimizing parameters. Candidate solutions are encoded as binary chromosomes, with genes representing parameter values; the implementation uses Gray code to encode positive…
The article walks through adapting a supervised PatchTST time-series model for use from an MQL5 expert advisor. It explains the model’s use of reversible instance normalization to normalize inputs and restore forecast scale, decomposition into trend and…
The article explains how to build, train, and test deep neural networks with the darch package in R. It contrasts unsupervised layer-wise pretraining using restricted Boltzmann machines or autoencoders with training a multilayer network from initialized…
The article describes three complementary methods for reducing the chance that a strategy’s backtest reflects historical noise: Validation-within-Validation (V-in-V), Combinatorially Purged Cross-Validation (CPCV), and Combinatorially Symmetric…
The article adapts the deterministic Dendritic Cell Algorithm, originally developed for anomaly detection, to continuous optimization. It maps candidate solutions to antigens, fitness to danger and safe signals, and agents’ maturation cycles to decisions…
The article examines whether random number generator quality affects stochastic population optimizers. It distinguishes pseudorandom generators from true random sources, reviews common generator families, and describes a 64-bit Mersenne Twister…
The article presents the Electromagnetism-like algorithm, a population-based metaheuristic for unconstrained continuous optimization. It models candidate solutions as charged particles: objective-function scores determine attraction or repulsion, and the…
The article explains SARSA, an on-policy reinforcement-learning method, and describes using it as an independent signal model in an MQL5 Expert Advisor. Its Q-value update uses the action actually selected in the next state under the agent’s current policy,…
Chaos Game Optimization (CGO) is a population-based method for searching multidimensional spaces. Agents begin at random positions within defined bounds. Each agent samples a subgroup, calculates its average position, and generates a candidate using one of…
The article describes an ensemble for a moving average channel Expert Advisor. The channel uses moving averages of high and low prices to generate breakout signals, while a Ridge Regression model uses the same technical indicators. Trades are allowed only…
The article explains batch normalization as a way to normalize neuron activations during training, aiming to reduce changes in the distributions passed between network layers. It describes calculating a batch mean and variance, standardizing values, then…
The article describes a proposed MetaTrader 5 strategy that combines a deep learning model’s next-close predictions with sentiment scores derived from news headlines and descriptions. In the backtest, a position is taken only when the direction implied by…
This article explains polynomial regression as linear regression on features formed by raising inputs to different powers. That lets a model fit curved relationships while remaining linear in its coefficients. It walks through forming the normal-equation…
This article introduces conformal prediction and MAPIE as tools for adding uncertainty estimates to existing machine-learning forecasts. A model is trained on one dataset, then conformity scores from a separate calibration set are used to form prediction…
This article experimentally examines ways to improve bagged neural-network classifiers built from market data. It prepares indicator and OHLC-derived predictors, ranks features, and compares three treatments for noisy training examples: correcting their…
The article adapts the immune-inspired Dendritic Cell Algorithm (DCA), originally developed for anomaly detection, into a population-based optimization method. It maps candidate fitness to danger and safety signals, which dendritic cells accumulate over…
This article outlines an enhanced Chaos Optimization Algorithm intended to search complex, multimodal objective functions. Its three phases combine a broad search driven by chaotic maps, a weighted-gradient search that accounts for constraints, and a local…
The article describes a unidirectional gold trading workflow that combines K-Means clustering with CatBoost classifiers. Rolling standard deviations of closing prices form both the main model’s features and separate meta-features used to identify market…
The article explains how to use MQL5’s Alglib classes to build a basic perceptron for testing trading ideas in an Expert Advisor. It outlines the roles of network setup, training data, training routines and forward processing, and describes normalizing input…
The article describes the Artificial Searching Swarm Algorithm (ASSA), a population-based method for optimizing difficult, potentially discontinuous or mixed-type objective functions. Each agent can move toward a recently improved agent in response to a…
This article introduces Multitask-Stockformer, a stock-selection framework that combines wavelet decomposition with temporal, convolutional, and graph attention components. A discrete wavelet transform separates historical asset data into low-frequency…
This article describes an architecture for running multiple LLM trading agents in parallel in MetaTrader 5. Rather than pooling agent opinions into a consensus, each agent follows a distinct trading philosophy and has its own notional capital, performance…