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 discussion addresses how to feed minute-frequency market data into an AI stock-ranking workflow. Its answer is to replace the stock ranker template’s basic feature extraction step with direct data-source extraction, because minute data does not come with…
This research summary reviews how unexpected U.S. monetary policy changes affected U.S.-registered stock and bond mutual funds from 2009 through early 2017. It measures policy surprises using narrow-window changes in Treasury yields around Federal Open…
The report estimates stock-specific volatility from residuals in CAPM, Fama-French three- and five-factor, and Carhart four-factor models, then studies its role in China’s A-share market. It describes a low-volatility portfolio that equally weights stocks in…
This market note evaluates several indicators for judging whether Chinese equities may be approaching a short-term bottom. It reviews northbound investor flows, fund equity exposure, and a turnover-to-volatility measure. The note reports that foreign…
The document summarizes a simplified market impact model for large investment portfolios. It builds on an Almgren-style impact function whose nonlinear power form can make portfolio optimization slow and can leave uncertainty about whether the numerical…
The document explains how cross-validation estimates whether a statistical or machine-learning model can predict data beyond its training sample. It describes ordinary k-fold validation, in which data is split into folds and each fold is used in turn for…
This research note examines Chinese property developers through their business mix, financial statements, operating efficiency, leverage, and return on equity. It groups developers by scale and argues that inventory turnover and debt repayment capacity…
This stock-selection method looks for shares with a daily high-low range above 1, turnover between 2% and 9%, and a rising five-day moving average over several consecutive observations. The combination uses price range and turnover as proxies for market…
The post describes a modified Chinese stock selection strategy that combines changes in executive shareholdings or share repurchase cancellations with momentum ranking. Its author reports that backtest return and drawdown metrics look acceptable, yet spot…
This brief troubleshooting note explains why sorted results may contain duplicate entries when working with Shenwan industry classifications. The platform provides two versions of the classification: an earlier version and one revised in 2021. The revision…
This study examines the shift from reversal toward momentum in Chinese A-shares, especially among stocks favored by institutions. It compares several ways to measure price strength: conventional lookback returns, returns excluding limit-up days, daily-rank…
This report explains the structure of US retirement savings, focusing on employer defined contribution plans, defined benefit plans, and individual retirement accounts. It describes the 401(k) as a participant-directed plan supported by tax advantages and…
This Chinese-language post outlines an equity screening rule: select stocks with RSI below 65, circulating market capitalization between 5 billion and 10 billion yuan, and a rising 30-day moving average. The stated rationale is to combine a momentum or…
This tutorial develops the Markowitz efficient frontier from mean-variance analysis, describing the minimum-variance portfolio, the maximum-Sharpe portfolio, and combinations that form the efficient set. It explains how forecasts of asset returns and a…
This report summary distinguishes active quantitative mutual funds from index enhancement and quantitative hedging products. Active quant funds generally face less stringent tracking error limits than index enhancement funds and typically use long-only…
The article argues that quantitative funds may exclude China’s specially treated A-share stocks because of trading constraints, limited strategy capacity, weak liquidity, and severe downside risks. It says restrictions on financing and securities lending…
The study examines whether news and social-media sentiment relate to intraday stock liquidity and returns. It combines stock-linked news sentiment with minute-level Twitter and StockTwits measures, then uses regressions and event studies to analyze liquidity…
The note explains how to interpret the rank_amount and rank_amount_20 features in a stock data set. It corrects the assumption that a 20-day rank is an integer from 1 to 20: the described rank_amount value is a cross-sectional percentile, comparing each…
This guide explains how to configure a genetic programming module to search for trading factors from supplied market data and input features. Users choose a market and data frequency, return field and horizon, time splits, fitness measure and thresholds,…
A user reported a training error despite having a dataset described as containing 115,478 rows. The discussion identifies a possible cause: at least one factor had the same value across all samples. According to the reply, the older version of the stock…
This document describes a market timing model for the CSI 300 based on the Guosen Investor Sentiment Index. It adapts a risk appetite approach by comparing 28 Shenwan first-level industries. The proposed index uses the Spearman rank correlation between…
This example stock-selection strategy looks for a bullishly ordered moving-average structure: the 5-day average must exceed the 10-day, then the 20-day, 40-day, and 120-day averages. Among those stocks, it identifies a pullback when the day's low falls below…
This report studies cross-sectional, market-neutral long-short strategies in commodity futures using inventory deviation and price momentum. The inventory signal measures how far a commodity’s stock level is from its trend; the strategy buys relatively…
The document summarizes a research paper on predicting stock returns by modeling momentum spillovers between related companies. It argues that conventional graph neural networks depend on predefined links, such as industry or ownership, and may pass…