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.
4,574 documents
The document describes an FX portfolio allocation sweep designed to isolate position sizing from signal selection. It starts from frozen equal-weight baseline candidates ranked by validation backtest performance, then advances configurations while preserving…
This notebook refits the FX configuration already selected by validation and registers its predictions for a later holdout backtest. It resolves the selection from solvent validation backtests, constrained to a frozen candidate set, then rebuilds the…
This notebook turns currency-pair prediction rankings into a traded baseline. At each decision time, it forms equal-sized long and short sleeves from the highest- and lowest-ranked pairs, then evaluates them with an FX backtest engine. Equal weighting is…
This study compares gradient-boosted trees with a penalized linear model for ranking currency pairs by future returns. Its motivation is that trees can represent conditional relationships, such as using momentum in one market regime and carry in another,…
This exploratory analysis profiles a four-hour OANDA dataset covering twenty currency pairs. It explains that the volume field counts quote updates at one venue, not consolidated traded size, and therefore cannot support a market-wide liquidity ranking. It…
This notebook describes how to prepare and register NLinear forecasts for FX labels. NLinear subtracts the last observed level from each fixed-length input history, focusing the model on changes within the lookback. A shared sequence eligibility process…
This analysis compares registered double machine learning treatment effects from nine trading case studies. It loads current results from each study’s registry, checks registry integrity and duplicate labels, and makes coverage explicit. Effects and…
This notebook builds fitted-model features from foreign-exchange price histories, complementing indicators calculated directly from past prices. It describes three approaches: a state-space filter that estimates a slowly changing price level and related…
This case study configures an LSTM model for forecasting FX-pair labels. It contrasts the recurrent model’s carried hidden state with fixed lookback transformations and convolutional receptive fields, while deferring comparisons with other models to a…
This notebook measures how a validation-selected FX strategy responds to changes in proportional transaction costs. For each prediction label, it selects a parent strategy from the signal, allocation, and risk-overlay results, then reruns that strategy…
This notebook reports parity comparisons between Backtrader, Zipline Reloaded, and an internal trading framework on selected real case-study strategies. It includes ETF and US equity-panel comparisons for both external engines, plus CME futures and…
This notebook builds a reusable diagnostic survey for financial datasets before preprocessing or feature engineering. It checks time-index dtype, null timestamps, ordering and uniqueness; finds exact and key-based duplicates; measures missing values by…
This notebook describes how to evaluate a previously selected FX strategy on holdout predictions. It reuses the selected configuration’s registered backtest specification, keeping the signal, allocation, risk controls, rebalance rule, costs, and account…
This notebook evaluates candidate features for 20 currency pairs one at a time against forward returns. It limits the analysis to walk-forward validation periods, stamps each session with its fold, and leaves a buffered gap before the holdout so labels…
This case study explains how to construct forward-return labels for a cross-sectional foreign-exchange strategy that ranks currency pairs and buys or sells according to their relative ordering. It first maps four-hour spot bars into trading sessions using a…
This notebook walks through a deployment cycle for a daily FX pairs model using Interactive Brokers as both the live data source and execution venue. It trains a Ridge regression model on a historical panel, computes a compact set of daily return, momentum,…
This notebook examines retained results from a real-strategy audit comparing the LEAN engine with matching ML4T Backtest profiles. It identifies the asset-class workloads supported by the frozen inputs and reports parity evidence across fills, valuations,…
This study examines whether traders can act on the Gotobi anomaly: a reported tendency for USD/JPY to rise toward 9:55 on Japanese dates divisible by five. It focuses on a practical question: whether a strategy based on that pattern remains valid once…
The document introduces a method for inferring lead-lag networks from the actions of individual agents, then applies it to trader-resolved foreign exchange data. It presents network persistence as an explanation for why one trader’s activity can help predict…
The study compares the complexity of log-return series for Bitcoin, GBP/USD, gold, and natural gas. It applies Multifractal Detrended Fluctuation Analysis and Refined Composite Multiscale Sample Entropy, two approaches that quantify different aspects of…
This study models financial prediction as a population of agents that follow trends over different prediction horizons. It examines how the population changes in stochastic, time-varying environments represented by real currency exchange-rate series. The…
This note develops a pedagogical approach to pricing foreign exchange options when the exchange rate is confined to a target zone with attainable, reflecting boundaries. It explains a no-arbitrage implication of this setting: the difference between foreign…
This work studies how aggregate trend-following or mean-reverting trading behavior can alter asset-price dynamics. It models the direction of trading explicitly and links those styles to serial correlation: trend-following behavior can create positive…
The study describes a method for classifying forex trends with a convolutional neural network. It first converts quantitative market data into images resembling trading charts, then trains a CNN to learn chart patterns and predict trend classes. Model…