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.
5,922 documents
This code describes a deep learning approach for turning sequential market features into position signals. Its example model uses an LSTM layer followed by dropout and a time-distributed output constrained through a hyperbolic tangent activation. Training…
This introductory multi-factor stock selection method combines one profitability measure, return on invested capital, with two valuation measures, price-to-sales and price-to-book. The stated goal is to rank stocks by factors associated with returns and…
This research summary examines newly launched and large mutual funds in China during the first half of 2020, focusing on their managers’ possible portfolio choices as funds completed or continued building positions. It argues that manager-level analysis can…
This example describes a daily portfolio rebalancer for five exchange-traded funds spanning US equities, long-term Treasury bonds, and gold. It starts by checking each holding’s weight against a target allocation. A rebalance is triggered on the first…
The report describes applying Thompson Sampling, an online learning method, to adjust a multi-factor investment model as market leadership changes. It uses China A-share constituents from the CSI 300 and CSI 500, with valuation, profitability, growth,…
The report compares sparse covariance estimation, which sets selected off-diagonal elements to zero, with rotation-invariant methods that adjust sample covariance eigenvalues. The latter include simulation-based, linear, and random-matrix-based nonlinear…
This weekly report reviews market conditions following a sharp equity selloff and describes trend signals across U.S. and Chinese equities, Hong Kong shares, gold, and bonds. Its trend-following system indicated rising trends in U.S. equities and bonds,…
This report studies trend-based allocation across CITIC Securities’ first-level industry indexes. It frames the process as choosing assets, measuring momentum, selecting industries, assigning weights, and managing portfolio tracking. The analysis describes…
This document introduces core combinatorics concepts through examples involving trading signals. It distinguishes combinations, where only the selected elements matter, from arrangements, where their order matters. It also describes versions that allow…
The document describes pure factor portfolios as a way to test whether a given equity factor has explanatory or investment value while controlling exposure to other factors. A pure portfolio is constructed to have unit exposure to the factor under…
This weekly report compares three Huatai fundamental stock-selection portfolios during a market pullback: a high-dividend-yield model, a low price-to-sales model, and a relative price-to-earnings model. It reports absolute returns, benchmark-relative…
The document compares simple systematic strategies, such as indicator rules applied to liquid developed markets, with advanced approaches that use specialist market knowledge, statistical testing, rigorous portfolio construction, and less liquid instruments.…
This monthly allocation report describes a strategic portfolio framework that maps broad asset holdings to underlying macroeconomic risks. It uses principal component analysis to define five factors—economic growth, interest rates, credit, term spread, and…
This Chinese-language strategy note combines two China A-share stock-selection approaches: a small-cap, low-turnover portfolio and a dividend-focused portfolio that excludes the smallest stocks. The first ranks eligible stocks using market capitalization and…
This document presents the China Securities Innovative Drug Industry Index as a way to track listed companies whose main businesses involve innovative drug research and development. It describes the index’s exposure to pharmaceutical subsectors, company…
This report introduces gcForest, or multi-grained cascade forest, as a tree-based deep learning model proposed as an alternative to deep neural networks. It argues that the method may suit financial data because it is described as more robust to…
This report evaluates risk parity, an asset allocation approach that aims to equalize the risk contributions of assets in a portfolio. It tests the approach using two historical data window designs: rolling windows and expanding recursive windows. For each…
This tutorial describes a Python workflow for comparing two subfunds with the China Securities 1000 Index using weekly net asset values. It calculates rolling correlations over a 20-day window and plots how those relationships change, then compares period…
This study examines how factor information decay can guide multi-factor stock selection. It defines a factor’s half-life from the time required for its information coefficient (IC), or IC information ratio (IC_IR), to fall to half its initial level. The…
This configuration describes a Qlib workflow for training a PyTorch feedforward neural network on China’s CSI 300 universe. It uses Alpha360 features, robust feature normalization, missing-value filling, and cross-sectional rank normalization of labels. The…
The article presents a robust version of equal risk contribution (risk parity) portfolio construction. It treats the estimated covariance matrix as uncertain and builds a bounded uncertainty set around the estimate. The optimization then accounts for…
This research note describes an enhancement strategy for China’s CSI 1000 small-cap equity index. It combines factors representing valuation, growth, trading behavior, capital flows, and network relationships within index constituents. After neutralizing for…
This research summary describes how to build and compare active and passive portfolios of Chinese healthcare-themed funds. For active funds, it argues that screening should use actual underlying holdings, since a fund's label may not reflect a concentrated…
This 2018 research summary forecasts changes to the FTSE China A50, an index of large Chinese mainland stocks selected with free-float and liquidity considerations. It describes quarterly reviews in March, June, September, and December, and identifies index…