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.
21,023 documents
This article describes a reproduction of residual network experiments on CIFAR-10, focusing on two changes from an earlier run: image augmentation and shortcut design. It pads training images by four pixels on each edge and randomly crops them back to the…
This article proposes using discrete-time signal-processing ideas to reduce the lag of moving-average trend indicators. It starts from an exponential moving average, frames the relationship between closing prices and the EMA as a transfer function, and…
The Arnaud Legoux Moving Average (ALMA) is described as a price-smoothing method that applies Gaussian weights across a rolling window. Its offset parameter shifts the weight distribution and adjusts the balance between smoothness and responsiveness, while…
The document outlines a broad framework for evaluating equity factors. It combines periodic cross-sectional robust linear regression to estimate factor returns, information coefficients (IC) to measure the relationship between factor exposures and subsequent…
This stock-selection strategy ranks shares using large-order net inflow and relative price performance. Its stated purchase screen looks at the prior 30 trading days, selecting stocks whose net inflow share from very large orders is among the top 5% of the…
This indicator aims to separate trading volume into estimated buying and selling components and display them as a two-color histogram. Its description associates green bars with buyer activity and red bars with seller activity. The calculation uses bar…
TRIX is described as an oscillator based on exponential averaging that can help identify potentially overbought or oversold conditions and can also serve as a momentum indicator. This variant replaces the regular EMA calculation with a double-smoothed Wilder…
This guide explains how to use historical backtests and parameter optimization as research checks before deploying a trading strategy. It outlines setup choices such as the instrument and exchange, bar interval, date range, fees, slippage, contract…
The article summarizes research on whether diversified business groups take longer than single industry firms to have earnings news reflected in share prices. It argues that interpreting segment results is harder when a company spans businesses with…
The document describes a rules based screen for Shenzhen main board stocks associated with robotics. It combines a daily turnover range of 3% to 12%, circulating market value below 10 billion yuan, price to earnings ratios from 0 to 29.01, and price to book…
The document defines Qlib data handlers for one-minute bars, aimed at high-frequency research and backtesting. The training handler creates normalized open, high, low, close, and approximate VWAP features, along with volume features. Price features are…
This brief question-and-answer explains how to assess whether returns across factor-sorted groups are ordered. A factor analysis sorts stocks in a chosen universe by factor value and divides them into groups; the lowest-ranked factor values form the bottom…
This Chinese equity screen selects stocks with turnover between 3% and 12%, year-over-year growth in net profit attributable to the parent company above 20% and no more than 100%, and prior-day actual turnover between 3% and 28%. The stated intent is to…
The indicator converts each candle into a weighted value based on its body, upper and lower shadows, direction, and gap from the preceding candle. Users set the relative weight of each component. A component’s contribution is scaled against its average…
Alphalens is a Python library for evaluating predictive stock factors. It turns a factor signal and pricing data into a structured dataset of forward returns, optionally assigning observations to quantiles and groups such as sectors. The resulting analysis…
This support exchange addresses errors encountered while following a BigQuant tutorial. It identifies two script issues: a data source table name cannot contain spaces, and the daily Chinese stock bar table exposes fields such as close, open, high, and low…
The document describes a chart indicator that plots the autocorrelation function for a signal. Autocorrelation compares a signal with a shifted copy of itself, showing how their relationship changes as the time lag varies. In market analysis, such a plot can…
The document introduces quantitative investing as using computer technology to analyze investment data, make trading decisions, and execute trades. It outlines two practical advantages: computers can process large amounts of information quickly, and…
This research summary compares three sources for studying foreign ownership of China A-shares: QFII/RQFII disclosures, Stock Connect holdings, and Refinitiv Ownership data. QFII reports arrive at lower frequency and can be substantially delayed. Stock…
The document describes a technical indicator that applies Gaussian smoothing to price data to form dynamic support and resistance levels. It presents the output as a price channel that traders can use to monitor changing levels across market timeframes. The…
The document shows a workflow for calculating daily industry-level average price-to-earnings ratios for Chinese stocks. It queries industry membership and valuation data, groups constituents by date and industry, filters valuation records to each group,…
The document summarizes a study of stock price movement prediction in China’s A-share market. It describes combining five data types: historical trading data, technical indicators, company announcements, news, and Baidu search activity. The researchers…
The document describes research that turns intraday stock volume data into factors for stock selection. It presents seven conventional volume distribution factors and six specialized ones. The methods include measuring minute volume through bucket entropy…
This guide explains how unit-root and cointegration tests can help identify mean-reverting combinations of asset prices. It presents the Augmented Dickey–Fuller test as a test of whether price changes depend on the current level, and relates the estimated…