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.
3,481 documents
The document outlines a data mining workflow for investment research, moving through data collection, cleaning, feature extraction, structuring, storage, analysis, and evaluation of results. It frames this workflow as a way to turn growing volumes of market…
This summary describes a style-neutral, multi-factor stock selection system for Chinese A-shares built from short-horizon price and volume behavior. The report develops nearly 200 trading-oriented alpha factors using daily stock data, aiming to capture the…
This essay argues that traders can accelerate learning by studying experienced practitioners’ actual trade records and decision patterns instead of relying only on abstract explanations or repeated personal trial and error. It recommends choosing examples…
This overview introduces risk parity as an asset-allocation approach and places it alongside the classical mean-variance framework. Mean-variance optimization seeks portfolios that maximize expected return for a given risk level or minimize risk for a target…
The report examines how fundamental factors predict stock returns differently across groups defined by market capitalization, profitability, and volatility. It argues that factor weights should vary by group rather than applying one set of weights across the…
This article presents the volume clock as an event-based way to measure market activity, in contrast with fixed calendar time. Sampling equal-volume intervals can reduce intraday seasonal effects and may produce observations better suited to conventional…
This research review studies whether stocks with unusually high trading volume earn a premium that reflects information about the real economy. It constructs high-volume premiums from stocks ranked by recent abnormal volume, then tests whether the premiums…
The article explains a method for separating a mutual fund’s total alpha into selection alpha and weighting alpha. Selection alpha is estimated by replacing reported portfolio weights with market-cap weights, isolating the contribution of choosing stocks;…
This overview discusses two research approaches to using deep learning for equity factor discovery. The first builds stock embeddings from co-investment patterns in fund portfolios: random walks on a stock-fund graph produce sequences for skip-gram training.…
This brief forum exchange addresses whether the minus-one term in a short-term stock factor formula removes shares trading below one currency unit. The reply says it does not: the subtraction applies to the value produced by a maximum operation. The…
This research summary describes a study of price spillovers from homes in Hong Kong identified as haunted. It reports progressively smaller average price declines across nearby properties: the affected unit, other units on the same floor, homes in the same…
This report examines whether changes in economic growth, inflation, interest rates, and risk premiums can help guide allocation among value and growth stocks and among large-, mid-, and small-cap Chinese equities. It uses GDP growth, CPI, the ten-year…
This Chinese-language research summary examines how differences between reported macroeconomic data and consensus forecasts may relate to returns in Chinese equities, bonds, and commodities. It compares forecast data sources, noting that Bloomberg records…
This report reviews U.S. mutual fund and ETF expense ratios through 2017, comparing equity, balanced, bond, money market, target date, active, and index products. It describes broad fee declines over time and relates them to investor demand for low cost…
This short introductory note describes quantitative investing as the use of mathematical models to turn investment ideas into measurable criteria. Those criteria can be used to identify opportunities, place trades, and manage risk. It highlights potential…
This platform update describes additions to an expression engine for quantitative analysis. The new rolling functions calculate sums, means, standard deviations, variances, skewness, and kurtosis over a specified lookback while handling missing values. It…
This article explains beta as an asset or portfolio’s sensitivity to a market benchmark, estimated through linear regression. It situates beta within a factor-model view, where returns are described as a combination of exposures to other assets or factors…
The document explains Bayesian optimization as a way to search expensive black-box objectives, such as portfolio return or risk-adjusted performance. It describes using a Gaussian process as a probabilistic model: observations update the model’s posterior,…
This article collects practical advice for building and debugging deep learning models, using a generative adversarial network for manga coloring as its motivating example. It recommends beginning with a small working baseline, visualizing predictions and…
This document describes John Ehlers’ Adaptive Relative Strength Index, an oscillator whose lookback changes with an estimated market cycle. It first smooths the midpoint of each bar, then applies digital filters to derive in-phase and quadrature components.…
The study examines timing the basis between Chinese equity indices and their futures, motivated by persistent futures discounts and uncertainty for market-neutral funds. To reduce the influence of carry components such as rates, dividends, and time to…
The document introduces the correlation coefficient as a measure of the direction and strength of a linear relationship between two variables. It defines the coefficient using covariance and the variables’ standard deviations, and explains that its values…
This research note studies whether stocks featured on China’s daily price-change leaderboards attract investor attention and how that attention relates to later returns. Using Sogou search index data as an attention proxy, it reports a U-shaped relationship…
The document answers a brief question about counting stocks within each industry group in a quantitative data expression. Its example groups records by the first-level Shenwan industry field and sums a conditional value: one for rows whose amount is…