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
The article develops a market replay simulator that turns one-minute OHLC bars into synthetic tick sequences. It seeds a pseudorandom number generator with a changing value so the generated paths vary between runs, while noting that a fixed seed can…
The article explains why a market replay and simulation system can use SQL through SQLite to manage orders and positions, rather than building every data-management routine in MQL5. It introduces SQLite’s five storage classes—NULL, INTEGER, REAL, TEXT, and…
The article presents a temporal nested cross-validation design intended to reduce optimistic evaluation in financial machine learning. It divides observations into an outer training zone, an inner validation checkpoint, and a final test segment that is…
The article explains how to build a MetaTrader 5 dashboard that compares relationships among selected instruments. It computes simple returns from recent closing prices, calculates pairwise Pearson correlation coefficients, then presents the values in a…
The article outlines a workflow for using an LSTM regression model to estimate future trend strength and filter trend-following entries. It motivates forecasting ADX rather than relying on its current value, which may rise after much of a move has already…
The article explains why a lifetime Profit Factor can hide changes in strategy performance and presents a framework for calculating it over rolling windows of completed trades. Each window contains a fixed number of consecutive trades; as it advances, the…
The article uses a simple moving average indicator as a template for building an MQL5 digital filter. It first explains how to inspect compiler errors, fix the earliest reported issue, and distinguish syntax errors from logical mistakes. It then introduces a…
The article describes an MQL5 execution utility that controls order submission frequency with a token bucket. Tokens refill continuously at a configured rate up to a capacity, allowing an initial burst while limiting sustained throughput. When no token is…
The article explains the Relative Strength Index, its smoothed gain and loss calculation, and common interpretations such as overbought and oversold levels, divergence, failure swings, and trend range shifts. It then presents a mean-reversion setup using an…
The article introduces Gann angles as chart lines intended to represent different rates of price movement relative to time. It describes the 1x1 line as the balance reference and discusses steeper and shallower ratios as possible indications of stronger or…
The article outlines a zone-aware signal system that combines previously identified probability-based support and resistance zones with programmable candlestick confirmations. Instead of treating contact with a zone as an entry by itself, the approach waits…
The article describes a trading signal workflow that encodes selected Ichimoku and Wilder’s ADX conditions as binary pattern flags, then trains a β-variational autoencoder to learn compact representations. Its motivation is that scaling continuous indicator…
The article explains Dynamics-Aware Discovery of Skills (DADS), an unsupervised reinforcement learning method intended to learn reusable behaviors that are both diverse and predictable. It contrasts DADS with DIAYN: DIAYN encourages distinguishable behavior,…
This article builds a price channel from independently fitted conditional quantile lines rather than a mean and standard deviation. It fits lower, median and upper quantiles with pinball loss, using iteratively reweighted least squares to solve for each…
The article implements a version of the DADA time-series anomaly detection framework, focusing on its adaptive bottleneck module. The design uses multiple autoencoder experts with different latent compression sizes. A gating mechanism selects the most…
This article proposes two code-organization changes to speed up a news-trading Expert Advisor during backtesting. First, it recommends loading the in-memory calendar database’s required information for the current day, reducing repeated database access…
This installment describes a basic method for creating synthetic tick data from one-minute OHLC and volume records. It spaces simulated ticks across the minute, assigns each a starting volume, distributes remaining volume randomly across ticks, and generates…
The article presents an architecture for managing trade stops in expert advisors that run across MetaTrader 4 and MetaTrader 5. It recommends keeping the main order manager focused on entering and exiting primary trades, while separate stop objects handle…
The document explains FreDF, a training approach for direct multi-step time-series forecasting. Direct forecast models predict several future values at once and commonly calculate their loss step by step in the time domain. The article argues that this can…
This article explains how to build an adaptive trend indicator in MQL5. It calculates True Range using the candle range and gaps relative to the previous close, then smooths volatility in two stages: Wilder’s moving average followed by a simple moving…
This article outlines an MQL5 Expert Advisor architecture that periodically rebuilds and trains a multilayer perceptron, then uses its forecasts while later optimization runs. It describes a workflow of creating the network, preparing and normalizing input…
The article defines an inside bar as a candle whose high and low remain within the preceding mother bar's range. It treats the pattern as a possible continuation or reversal setup and recommends assessing its context, with higher timeframes, nearby support…
The article describes FinMem, an LLM-based trading-agent design organized around profiling, memory, and decision-making. Its profile combines company and sector context with a selected risk style: risk-seeking, risk-averse, or self-adaptive. The memory…
The article explains why indicators that redraw past values as new prices arrive are difficult to assess from a static chart. It proposes using a non-trading Expert Advisor in the MetaTrader visual Strategy Tester, attaching the indicator during a paused…