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
The report proposes benchmark indexes built from funds that seek to outperform the CSI 300 or CSI 500. It combines contractually designated index-enhanced funds with other funds identified as similar through their benchmarks, tracking error, and historical…
This brief student response defines quantitative investing as using data, statistics, research, systematic data collection, backtesting, and programmed trading to make investment decisions. It highlights several perceived advantages: a repeatable process,…
This research summary reviews trend-following indicators and how to build strategies for broad asset allocation and industry allocation. It groups 41 indicators by their input data, filtering, moving-average construction, and signal generation, arguing that…
This document presents a Chinese-stock ranking workflow built around an XGBoost RankNet model. It constructs price-return and trading-activity features, labels stocks using clipped forward returns divided into quantile bins, and separates training data from…
This report outlines a first step in building a multifactor model: choosing which factors to combine. It proposes screening combinations using three criteria: how strongly a factor differs from benchmark exposure, how correlated the factors are with each…
The document presents a framework for deciding whether a candidate multifactor strategy adds value beyond a small set of investable reference factors, called elementary smart betas. It models each candidate through its exposures to those factors and its…
This commentary reviews the Chinese quantitative-investing environment in 2022 and presents a private manager’s expectations for 2023. It attributes a difficult path to excess returns to weak trading activity and rapid shifts in market style, alongside a…
This weekly report compares artificial intelligence stock-selection portfolios across broad A-share universes, industry-neutral portfolios, and portfolios restricted to the CSI 300 or CSI 500. It names XGBoost, support vector machines, random forests,…
This code module outlines methods for constructing sparse portfolios intended to exhibit mean reversion. It includes Box–Tiao canonical decomposition, greedy support selection, semidefinite optimization under volatility constraints, and sparsity methods…
This brief literature note introduces a VaR-augmented Black-Litterman approach for constructing an absolute-return fund-of-funds portfolio with market-risk controls. The described formulation incorporates value-at-risk alongside practical trading…
This note discusses market style and public mutual fund positioning in Chinese equities. It compares the share of CSI 300 constituents within the broader CSI 800 universe, reporting that the share was 81% at the time of the note. It also cites two recent…
The document describes a software implementation of the Johansen cointegration method for forming mean-reverting portfolios from asset prices. It computes cointegration vectors, orders them by eigenvalue, and converts each vector into hedge ratios normalized…
This April 2020 market note reviews a sharp decline in Chinese equities amid weakening fundamentals and falling investor sentiment. It reports that large- and small-cap indexes had similar declines, while consumer and utility sectors held up better and…
This programming note explains how an object-oriented trading advisor can manage its configurable inputs as a structured set. It contrasts manually declaring fields and writing string conversion methods with a macro-based library that generates parameter…
This Chinese-language research summary discusses forecasting stock-selection factor returns after commonly used factors became more volatile. It focuses on screening timing variables with Lasso and Elastic Net regression, then extending factor timing with…
This short weekly report summarizes a divided Chinese equity market, with major indices moving in different directions. It compares the reported changes in the Shanghai Composite, CSI 500, ChiNext, and CSI 300, then notes that most of the covered long…
This sector-rotation study uses the absorption ratio to measure how concentrated the risks embedded in a group of asset returns are. The ratio reflects the share of total return variation explained by a small set of orthogonal components: higher readings are…
This 2018 report reviews managed futures, including how CTA strategies trade futures and options and how they differ by analysis method, trading style, holding period, and markets covered. It describes systematic and discretionary approaches alongside trend…
This article outlines an asset allocation method that combines risk parity with the Black-Litterman framework. Risk parity portfolio weights serve as prior weights, which are used to infer prior expected returns. Short-term momentum views are then…
This weekly review summarizes Chinese equity market conditions for the week ending April 3, 2020, then presents Haitong Securities’ market timing, factor, industry rotation, and strategy-performance views. Its evidence includes index returns, valuation and…
The document summarizes a portfolio construction method that translates target macroeconomic factor exposures into asset weights. It considers six factors: equity, real interest rates, credit, inflation, emerging markets, and commodities. A standard linear…
The article argues that no single programming language is best for every algorithmic trading system. It recommends starting with system requirements and strategy characteristics, then selecting tools for separate components such as historical research,…
This Chinese research summary describes an equity index-enhancement approach using listed companies’ reported revenue by business line. It uses annual and interim report data, then applies Chinese-language keyword extraction and similarity analysis to…
This discussion describes a model-stacking problem in a Chinese quantitative research platform. A user has trained and persisted twenty StockRanker models using the same factor set but differing in date ranges, data filters, or prediction horizons. The goal…