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
This study examines how factor exposure limits affect optimized portfolios built around the CSI 300 and CSI 500 indices. It reports that raising the limits can increase excess returns, drawdowns, and tracking error, so the change in return relative to the…
This brief question and answer discusses how to choose moving-average pairs and crossover rules for a dual moving-average strategy applied to an individual stock or fund. It lists comparisons among several simple moving-average windows and asks which…
This note describes two stock selection approaches, XGBoost and StockRank, using the same five inputs: ranked dividend yield, market capitalization, closing price, a 20-period price ratio, and average turnover. It also gives example hyperparameter…
The document outlines a simple stock market study of whether shares that rise more than 1% on each of three consecutive days are more likely to gain the next day. It defines a signal for that three-day run and a next-day return outcome, then calculates…
This research summary studies mainland Chinese stocks held through the Hong Kong–mainland Connect programs. It compares the holdings’ industry and Barra-style factor exposures with the CSI 300 and CSI 500, then tests the proportion of a stock held by…
The report describes a sector-rotation framework that pairs quantitative clustering with macroeconomic reasoning. It first groups industries using features such as returns across market conditions, return on equity, and revenue growth, then makes limited…
This weekly review compares equity style factors across broad Chinese stock pools and major index constituents. It reports that size and beta performed relatively well during the week, while profitability and financial quality remained strong. Results for…
This report studies factor performance within the SSE 50 rather than across the full market, arguing that analysis by investment universe can reveal patterns hidden by broad aggregate tests and can inform index enhancement or Smart Beta portfolios. It tests…
This study tests whether hedge funds can time returns in particular industries and whether that ability predicts later performance. It estimates industry timing using rolling regressions on market-adjusted industry returns, then evaluates the resulting…
This BigQuant forum thread discusses whether rolling parameter optimization can run in simulated trading and whether newly optimized parameters are automatically applied to a strategy. Replies say that rolling optimization can work in simulated trading,…
The author describes a stock reversal approach based on a prior limit-up move followed by a pullback, with a bullish candle engulfing a bearish one as a defining pattern. A proposed variation focuses on a leading stock's first bearish session and subsequent…
This document introduces a lecture on deep reinforcement learning by researcher Matteo Hessel. It says the session discusses practical considerations and algorithms, along with implementing them using automatic differentiation tools such as JAX. The subject…
The article explains the price-to-earnings ratio as a way to relate a stock’s price to its earnings per share. It illustrates that two shares with the same price can have very different P/E ratios when their earnings differ, and describes the ratio…
The document summarizes research on long-horizon rotation between large-cap, high-ROE sectors and smaller-cap sectors. It argues that their relative performance can shift over periods of years, and that size-based and ROE-based rotations may move together.…
This article outlines a method for anticipating business conditions in three Chinese consumer sectors: food and beverages, textiles and apparel, and retail. It treats financial measures such as revenue, net income, returns, margins, and operating efficiency…
This report outlines a CANSLIM-inspired stock-selection model adapted to China’s A-share market. It first identifies price bases, including cup-and-handle, double-bottom, and butterfly patterns, then combines these setups with fundamental and technical…
This Q&A compilation collects practitioner questions about developing, combining, and deploying quantitative strategies. Topics include variation between XGBoost rolling-training backtests, selecting and evaluating factors and models, overlapping holdings…
This report surveys U.S.-listed growth-factor ETFs, describing their number, assets, growth over time, and return characteristics. It groups the funds by market-cap segment and notes that the strongest recent absolute performers were mostly large-cap…
This single-stock strategy builds upper and lower price bands from the previous 20 days of closing prices, using the mean plus or minus two standard deviations. It enters when the close rises above the upper band and exits when the close falls below the…
This article recounts the 2022 nickel market crisis involving Tsingshan, Glencore, and the London Metal Exchange. It uses a simplified commodity-delivery analogy to explain how a seller with a large short futures position may face stress when prices surge…
The document answers a question about retrieving one company metric, such as management expenses, for multiple stocks. It points readers to a platform template for choosing the metric and explains that a code-list module can be used to enter the stock codes…
This Chinese equity strategy combines two types of candidates: stocks in the top 30% by return over the prior 120 trading days, and stocks in the bottom 10% by performance over the prior 30 days. A candidate must also have net inflows from very large orders…
This tutorial explains the skip-gram model as a neural network trained to predict a nearby word from a chosen input word. Training examples are word pairs drawn from a configurable context window. The input word is represented as a one-hot vector, and a…
The discussion concerns a visual trading workflow where an earlier filter can return no rows for a given day, causing a later rolling sequence window module to fail. One response explains that an empty input cannot be processed by that module as configured…