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 report proposes a cross-sectional stock-selection framework that combines choices of factor subspaces, model families, and ensemble rules. Its base learners include boosted trees, ExtraTrees, random forests, Ridge regression, and Lasso regression,…
This report explores whether stock selection factors built from Level 2 transaction records can predict returns over a monthly horizon. It derives measures from trade size categories and trade direction, including the shares of trading represented by…
This short question-and-answer thread addresses how to obtain a list of stocks underlying convertible bonds in the BigQuant data environment. A reply suggests reading the `stk_instrument` field from the `basic_info_CN_BOND` table. The questioner then asks…
This guide explains the workflow for a BigQuant contest focused on constructing 15-minute factors from three-second stock snapshot data. Participants are asked to write factor logic without using model-generated factor synthesis. The platform provides…
This research digest summarizes two finance topics. The first reviews competing claims about the size effect, the long-discussed idea that smaller companies have higher average returns than larger ones. It says the underlying study uses simple tests and…
This overview describes a practical Python workflow spanning data collection, database access, and machine-learning, deep-learning, and natural-language-processing tools. Its investing example uses traditional factors measured over rolling 12-month windows…
This report examines domestic Chinese fund products that provide exposure to Hong Kong equities and their role in broader asset allocation. It reviews access routes and fund types, including actively managed products and index funds, then compares the…
This overview explains the Kuramoto model, which represents interacting oscillators through phase dynamics. Each oscillator’s motion depends on its natural frequency, coupling strength, and phase differences with the others. Positive coupling draws phases…
This research summary treats the assignment of special treatment (ST) status as a proxy for financial distress among Chinese A-share companies. It describes using Fisher discriminant analysis and regression-based models to predict company financial…
This document outlines the development of US investment advisory services across four historical phases and describes the industry’s current structure. It proposes examining market scale, concentration, client types, service models, and fee arrangements, but…
This document connects prospect theory and the disposition effect to potential changes in stock supply. It describes a behavioral asymmetry: investors tend to realize gains quickly, while holding losing positions longer and sometimes accepting more risk in…
The document reviews thematic funds, which target long-term narratives such as artificial intelligence or alternative energy. It describes a surge in US thematic-fund assets and inflows through early 2021, then frames investment as a difficult selection…
The document outlines how sequence learning can fit into a conventional quantitative investment process: feature extraction, individual asset return prediction, portfolio construction, and trade execution. It describes using price and volume measures,…
The article offers three discretionary rules for stock traders: favor stocks that have previously produced gains for the trader, concentrate research on a small number of names, and use price trends to guide holding and exit decisions. Its most explicit…
This weekly factor report describes two forms of diagnostics for the ten highest-ranked anonymous ModelScore factors: average style exposures estimated through daily multivariate regressions, and average industry-level Rank ICs measuring the cross-sectional…
The study applies the CCK approach to detect herding among the constituents of Chinese equity indices. It treats herding as a period when constituent returns move more alike: their co-movement increases while the dispersion of returns falls. The authors also…
This research summary examines what drives convergence in the basis of stock index futures and how those drivers can inform cash-and-futures hedging. After accounting for dividend effects, it identifies the current contract’s annualized basis, the index’s…
The document outlines a daily stock rotation schedule: buy ten stocks at the open, hold each cohort for five days, and sell it at the close on its fifth day. It allocates capital in five equal portions, investing one portion in each day’s new cohort. From…
This document poses a model-design question: whether the number of input factors in a StockRank model relates to the number of leaves per tree or the total number of trees, and whether there are useful rules of thumb for choosing these quantities. It…
This note introduces the Average Directional Index (ADX), an indicator calculated from high, low, and close prices over a chosen lookback length. It outlines the intermediate calculations: true range, positive and negative directional movement, and the…
This summary reports tests of 19 factors derived from sell-side analyst consensus data, covering forecast financial measures and analyst attention. For financial factors, it compares raw values with industry-relative rankings and quarter-over-quarter…
The article discusses Chinese stocks that close at the daily limit and then open sharply lower. It offers two possible explanations: short selling followed by repurchasing at a lower price, and conventional large traders either distributing shares near a…
This article reviews Word2Vec as an efficient method for learning word vectors from context, and compares it with matrix-based and clustering approaches to distributed representations. It discusses CBOW and Skip-Gram, emphasizing their different update…
This document outlines a machine learning workflow built around feature engineering, model training, and model combination. Feature engineering includes creating, extracting, and selecting variables. It contrasts linear models, which capture linear…