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 research note summarizes tests of profitability and earnings-quality factors. It reports examining 14 profitability factors, with two additional measures of return on assets and return on invested capital calculated using EBIT. Factor effectiveness is…
This Chinese-language project note describes an attempt to improve a stock strategy using technical factors. The author combines volatility, price, momentum, and liquidity measures, and filters stocks with unusually high or low characteristics in an effort…
The note compares two strategy versions and argues that both feature quality and model choice matter. It uses a cooking analogy to convey that useful factors cannot fully compensate for a weak model, and a capable model still depends on informative inputs.…
This study examines whether Chinese mutual funds’ heavily held A-share stocks rise near quarter-end and then reverse early in the next quarter. Instead of testing fund returns alone, it measures cumulative abnormal returns and excess trading volume for…
The supplied abstract reviews the development and calculation of the CBOE volatility index, then describes applying an index calculation to 50ETF options in China. It reports that the resulting measure has small calculation errors relative to the Chinese…
This opinion piece argues that China’s A-share T+1 settlement rule can leave retail investors unable to sell shares bought that day, while institutions may approximate same-day trading through short selling or by trading around an existing position. It…
The document introduces two regression methods used in financial analysis. Logistic regression models the probability of a binary outcome with values constrained between zero and one, then assigns a class by comparing that probability with a threshold. It…
This document summarizes research on estimating KOSDAQ initial public offering stock values with artificial neural networks. The stated motivation is that multiple-based IPO valuation reflects sentiment in the market environment and may not reliably predict…
This platform support discussion explains a LightGBM training failure in a stock-selection workflow. The intended factor marks a day when the previous six closes were all below the five-day moving average and the current close moves above it. Inspection of…
This tutorial develops a grid market-making example for a large-tick asset, where queue position can strongly affect execution. Its first version estimates fair value from best bid and ask prices weighted by displayed quantity, then shifts a reservation…
This tutorial adds two risk controls to a Chinese small-cap equity strategy. The underlying template excludes stocks listed for less than a year and special-treatment shares, ranks by market capitalization, holds 30 stocks at equal weights, and rebalances on…
The document explains how an ascending percentile rank converts a factor’s cross-sectional ordering into a normalized value. For each stock, sort the factor from lowest to highest, assign its rank, and divide that rank by the total number of stocks in the…
The article explains inner and outer market volume as a basic measure of aggressive selling and buying: trades executed at the bid are counted as inner volume, while trades at the ask count as outer volume. It proposes reading these figures alongside price…
The document compares methods for using high-frequency factors with strong short effects to improve a CSI 500 index-enhancement strategy. It distinguishes screening out stocks before building the strategy from removing them afterward, and describes combining…
The document reviews a firm life cycle classification based on the signs of operating, investing, and financing cash flows. Their combinations map companies to startup, growth, maturity, shakeout, or decline stages. This approach is compared with…
This forum post shows a workflow for preparing Chinese stock data to neutralize valuation measures by industry and market capitalization. It joins daily bars, price-to-book values, market capitalization, and industry classifications, then pivots valuation…
This beginner overview surveys several approaches to quantitative stock selection: multi-factor scoring or regression, style rotation, money-flow signals, momentum and reversal, analyst consensus, ownership concentration, and trend following. For style…
The note translates a familiar trend-following approach into rules: buy when a short moving average crosses above a longer one, then use a volatility-based trailing stop derived from a rolling standard deviation. The stop can move upward as its calculated…
The article discusses why people who lose money gambling may keep playing in an effort to recover their losses. It relates this behavior to prospect theory: people may avoid risk when ahead but seek risk when behind. It also describes loss aversion, where…
This brief data-platform exchange explains why a Chinese stock’s prices may differ from those shown by another trading application. The user reports that extracted prices appear backward-adjusted and questions whether this discrepancy makes backtests…
This article explains how to assess whether a cryptocurrency pair spread is suitable for mean-reversion trading. It presents the Augmented Dickey-Fuller (ADF) test as a check for a unit root, with a low p-value used to reject non-stationarity, and the Hurst…
The report summary presents a two-layer framework for improving the stock universe used in multi-factor strategies. Rigid filters aim to ensure that securities are tradable, while flexible filters seek to improve universe quality. The flexible layer…
This document summarizes a research note on building a systematic framework for allocating among asset classes, with a focus on domestic equities and bonds. It says the research team had worked on asset allocation for several years and began consolidating…
This documentation describes a basic Transformer model for financial time-series prediction and how to configure and train it in a quantitative research framework. Inputs are represented as batches, time steps, and features. An embedding layer transforms…