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 summary of a Haitong research presentation discusses practical problems that arise when applying multi-factor stock-selection models. It observes that assumptions used in basic portfolio construction may fail in real settings, so models may need…
The document explains quantile regression as an alternative to ordinary least squares for modeling different points in a return distribution. Because it does not rely on the same normality and constant-variance assumptions, it may be useful with skewed,…
This report studies the return-forecasting stage of a multi-factor investment process. Starting with daily factor returns estimated by regression, it compares three approaches: a rolling mean of factor returns, a recency-weighted rolling mean, and a…
The article explains how spot crypto ETFs provide exposure through regulated brokerage products and, in the case of Bitcoin funds, hold the underlying asset. It argues that ETF purchases can shift coins out of open-market circulation, while familiar account…
This summary of a Chinese A-share research report explains small-cap and value premiums through stocks’ differing sensitivities to cash-flow expectations and discount rates. It frames equity returns with a two-beta model that separates those influences,…
This article describes a strategy that changes exposure to equity factors according to a four-state business-cycle classification: recovery, expansion, slowdown, and contraction. It combines the US Leading Economic Indicator with a Global Risk Appetite Cycle…
This document summarizes an article on applying quantitative methods to environmental, social, and governance investing. It argues that quantitative portfolio tools can translate investor preferences into portfolio construction and extend sustainability…
This code documents a factor evaluation workflow. It computes Spearman rank information coefficients between factor values and forward returns, with options to demean returns by group and summarize results over time or across groups. It also translates…
This 2019 outlook reviews China’s financial product and asset management markets amid regulatory change. It argues that rapid market expansion had already slowed before the new asset management rules, with aggregate assets contracting in 2018. The report…
The article describes a constant-mix strategy inspired by Claude Shannon: hold equal portions of an asset and cash, then rebalance after large price moves. Rebalancing sells some of the asset after gains and buys after losses, harvesting price fluctuations…
This BigQuant workflow illustrates a daily Chinese equity ranking process. It prepares stock features from price and prefactor data, including lagged returns and volume ranking, then creates a supervised label from a clipped forward return. Training data is…
This Chinese-language summary introduces a research paper on factor timing, which combines ideas from market timing and factor investing. It frames the theoretical argument around the stochastic discount factor (SDF), assuming a conditional factor model can…
This research summary organizes equity factor allocation around two macro conditions: economic growth or recession, and tighter or looser liquidity. Their combinations define four regimes in a factor investment clock. The proposed framework treats corporate…
The document frames a multi-strategy allocation problem: how to adjust the capital assigned to several simulated strategies, including custom and subscribed strategies, as their performance changes. It proposes considering changes in the slope of each…
The document presents a SQL example for constructing a turnover-related stock factor and neutralizing it for industry membership and company size. It calculates the logarithm of total market capitalization, a 20-day average of turnover, and a neutralized…
This report summary describes a Chinese equity strategy that rotates among industry and theme ETFs, then uses the portfolio in absolute-return constructions. It selects 14 industry indices with matching ETFs from a broader set of industries, aiming to…
The document explains alpha as a signal or mathematical expression that can be converted into portfolio weights, then outlines a research cycle from investment hypothesis and data choice through backtesting and review. It discusses potential signal sources,…
The document outlines an equities statistical arbitrage method that uses principal component analysis to estimate common return factors. Asset returns are standardized, PCA components provide factor weights, and regressions of returns on factor returns…
The document presents a trading system as an operating plan, not just a rule for choosing when to buy or sell. It covers defining the mandate and instruments, selecting a strategy, allocating capital, setting entry and exit rules, sizing positions, and…
The article describes a data-driven way to group stocks for portfolio diversification. Instead of relying only on fixed business-sector labels, it trains models to predict future pairwise return correlations, then clusters stocks using the predicted…
This report outlines a framework for classifying and evaluating bond funds, with emphasis on medium- and long-term pure bond funds. It groups funds by investment objective, such as seeking returns above the bond market or favoring particular bond types,…
This short forum post considers whether combining quantitative stock selection with discretionary choices improves results. The author argues that adding human stock picks to a quantitative selection process may perform worse than either approach alone, and…
This tutorial explains attention and Transformer components before outlining a time-series stock ranking model. It adds Time2Vec features to factor sequences so the model can represent both periodic and linear time patterns, then applies self-attention,…
This configuration defines a Qlib workflow that trains a CatBoost model on Alpha360 features for CSI 300 constituents. It uses Chinese market data from 2008 through mid-2020, with training through 2014, validation over 2015–2016, and testing from 2017…