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.
2,013 documents
This review explains how climate change can affect financial institutions through physical hazards such as floods and droughts, and transition pressures such as new climate policy, technology shifts, litigation, and changing customer demand. It maps these…
The article explains Theory-Implied Correlation (TIC), a method for estimating portfolio correlations by combining observed correlations with an externally specified hierarchy of assets. It describes three stages: fit a hierarchical tree to empirical…
This introduction compares four portfolio selection benchmarks using a collection of 23 ETFs with closing prices from 2008 to 2016. Buy and Hold starts with fixed allocations and lets weights drift with asset prices; Best Stock selects the strongest asset…
This essay discusses how asset owners, asset managers, and companies can support sustainable investing by incorporating environmental, social, and governance considerations alongside financial analysis. It presents long-term ownership and broad market…
This article describes LightGTS, a time-series forecasting framework designed to handle datasets with different sampling scales and recurring periods. Its central method is adaptive periodic patching: estimate or otherwise determine a series’ cycle length,…
The article describes a Time-MoE forecasting architecture that represents each time step as a token, processes temporal context with transformer blocks, and predicts across multiple horizons. Its focus is the sparse mixture-of-experts component: a router…
This article describes a workflow for developing, optimizing, and deploying a multi-currency Expert Advisor built from simple trading strategies. It separates reusable library code from project-specific strategy code, then organizes parameters and…
The article tests how often markets qualify as trending versus flat by comparing five classification methods: ADX above a threshold, a Bollinger based trend indicator, Percentage of Trend, an RSI filter, and a ZigZag trend detector. It describes an MQL5 tool…
This article describes an infrastructure layer that lets an Expert Advisor request logical instrument names while resolving broker-specific symbol variants at runtime. Its components include a persistent mapping store, a resolver, an in-memory cache, and a…
This article shows how to build a MetaTrader 5 indicator that displays several timeframe charts inside a chart subwindow. Buttons let users add chart objects for selected timeframes and toggle chart properties, including settings that are unavailable through…
The document introduces gating as a way to adapt how an ensemble uses specialist models according to contextual variables or model outputs. It distinguishes selecting one model for each case from combining several models with context-dependent weights.…
The document describes an MQL5 chart indicator for comparing price movements across the current instrument and five user-selected symbols. It displays raw prices as lines, bars, or candlesticks and lets users choose where the series should be aligned: at a…
The article explains TimeFound, a Transformer-based foundation model designed to forecast time series across domains, including financial data. Its central preprocessing idea is multi-scale patching: segments at several resolutions are projected into a…
The article describes a proposed integration of ST-Expert, a graph-based mixture-of-experts approach, with Extralonger, a modular Transformer framework for time-series forecasting. It frames market data as both sequences and networks: assets or time points…
The article explains cross-recurrence quantification analysis (CRQA), a nonlinear way to compare the evolving states of two time series. Each series is embedded separately using matching dimension and delay settings; pairwise distances between embedded…
This article recasts a binary recurrence matrix as the adjacency matrix of an undirected network: observations become nodes, and recurrent pairs become edges, with self-links removed. It explains how network measures such as degree, clustering, transitivity,…
This installment describes organizing MQL5 code for projects that use a shared Expert Advisor library. It recommends keeping each project in its own repository and placing each dependency in an Include folder as a separate repository clone. A shared library…
This installment in a multi-currency EA series extends virtual position handling to pending buy and sell limit and stop orders. It proposes representing pending orders and positions with one class: when market price reaches an order’s trigger, the object…
The article concludes a series on GinAR, a neural architecture for noisy financial time series. Its cells combine contextual attention, adaptive graph processing, recurrent state, and separate forget and reset controls. Multiple layers process a sequence,…
This article introduces K²VAE, a framework intended to forecast distributions of future time-series paths rather than a single price estimate. It combines a variational autoencoder with a learned Koopman representation, which approximates nonlinear dynamics…
This article describes a small data preparation workflow for turning MetaTrader 5 price history into inputs for a computer-generated character whose size changes with price. It outlines linear and exponential mappings from price changes to object scale, then…
This article presents an architecture for running multiple trading models across instruments and timeframes within one MetaTrader 5 expert advisor. It frames the configuration problem as a set of model, timeframe, and instrument combinations, each of which…
This article explains how the ALGLIB port available to MQL5 developers can support numerical work needed for dynamic hedging and multi-asset strategies. It contrasts single-window correlation indicators and hand-built calculations with tools for matrix…
The article argues that limited market history can leave a trading system exposed to only one realized path, then proposes generating synthetic price series to examine a rebuy algorithm across many independent instruments. Its price model uses discrete…