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.
16,761 documents
This document outlines an automated strategy that trades breakouts from a price range during configurable trading hours. Its settings cover start dates, position volume, risk per order, stop-loss placement, take profit, trailing stops, partial exits, and…
This expert advisor trades signals from an Anchored Momentum Candle indicator. A signal is generated when a bar closes and the indicator’s candle color changes; the document says the advisor uses that event to initiate trades. It also notes that a…
The document surveys a platform’s patent database and describes how patent characteristics are turned into equity-selection factors. It groups patent information by patent type, application or grant status, and measures such as counts, claims, citations,…
The document describes a stock screen requiring turnover between 3% and 12%, a listing year of 2021, and a positive return. It presents these conditions as a basic selection rule and includes example implementations for screening stocks. The approach…
This configuration defines a Qlib workflow for ranking CSI 500 equities with a LightGBM model and Alpha360 features. It uses cross-sectional rank normalization for labels and trains on data from 2008 through 2014, validates on 2015–2016, and evaluates a…
This Qlib configuration defines a Chinese equity research workflow using the CSI 300 universe and Alpha158 features. It trains a double ensemble of gradient-boosted models, with both sample reweighting and feature selection enabled, then evaluates signals…
This report explains how to design a cryptocurrency arbitrage strategy across centralized exchanges and decentralized exchanges using automated market maker pools. It covers the differences between order-book prices and pool pricing, then identifies costs…
This BigQuant strategy tutorial outlines a framework for combining a broad-market risk signal with individual-stock timing and position management. The proposed workflow builds separate feature expressions for index-level risk and stock-level conditions,…
The article compares a StockRanker stock-selection strategy with a Deepalpha deep neural network using the same 17 short-term factors, training data, and backtest setup. The strategy selects the top-ranked A-share stock each trading day and sells it the…
This document describes an automated trading system that uses the 2XMA Ichimoku Oscillator. It generates a signal when a bar closes and the indicator changes direction, then uses that change to determine trading actions. The text also notes dependencies for…
This post describes an A-share screening rule combining daily turnover between 3% and 12%, an opening price within 5% of the 10-day moving average, and a moving-average relationship intended to indicate alignment. It includes a formula-style example and a…
This document explains how MetaTrader 5 strategy tester cache files for individual test runs can be read, saved, and inspected without rerunning the tester. Example scripts convert real account trading history into a tester-compatible file and retrieve the…
This stock screen combines price location among several moving averages, a Bollinger-band position, and a recent-gain constraint. The described filters use five averages spanning short to longer horizons, place the close between the band’s upper and middle…
This Expert Advisor generates a trading signal when a bar closes and the JBrainTrend1Stop NRTR indicator changes color. The description says the EA depends on compiled indicator files being installed in the platform’s indicator directory. It also mentions a…
This indicator defines the Asian session range, then looks for a London or New York wick through one boundary followed by a close back inside. A signal requires a later close beyond the opposite end of the sweep bar within four bars. It enters at the next…
This forum post describes an early automated trading system tested on EURUSD at the one-minute frequency. The author reports a very large gain over much of 2007 from an initial deposit of 3,000 USD, but says the system lost the deposit over 2005–2006 and in…
The article introduces logistic regression as a binary classifier that maps a weighted combination of features to a probability. It outlines fitting the model by maximizing the likelihood of observed labels, equivalently minimizing a loss with gradient…
This guide explains how to inspect a Moonshot strategy interactively after moving it from a development notebook into a Python file. In JupyterLab, the user opens the strategy file, attaches a console, loads the file’s contents, creates a strategy instance,…
This report proposes a cross-sectional stock-selection framework that combines choices of factor subspaces, model families, and ensemble rules. Its base learners include boosted trees, ExtraTrees, random forests, Ridge regression, and Lasso regression,…
This report explores whether stock selection factors built from Level 2 transaction records can predict returns over a monthly horizon. It derives measures from trade size categories and trade direction, including the shares of trading represented by…
The script demonstrates a calendar spread strategy for two nearby equity index futures contracts. It calculates the spread between their closing prices over a rolling window, estimates the mean and standard deviation, and sets upper and lower thresholds two…
The script describes a daily SPY allocation strategy driven by CNN’s Fear and Greed Index. A research agent retrieves the latest score from a prior day, while a separate trading agent maps score ranges to target allocations: higher equity exposure at low…
The document describes a local data-loading problem in VeighNa: a Jupyter Notebook backtesting example reports that historical data loading has completed but contains zero records. The proposed fix is to run Jupyter Notebook from the CTA backtesting example…
This document outlines a Chinese equity screen using three conditions: price amplitude above 1, a 9:25 price increase below 6%, and recent revenue at least 1.1 times the 2018 revenue figure. It frames the revenue comparison as a growth filter and combines it…