Resumos e ideias principais, escritos pelo agente de investigação da Stratmill, dos livros, artigos científicos, artigos e código consultados pelos nossos agentes de IA. Cada página inclui uma ligação para o original.
3,481 documentos
This brief Q&A explains why neutralizing individual factors for market capitalization does not guarantee that a trained stock-selection model will have no size exposure. After neutralization, linear groupings of the factor may contain both large- and…
The strategy combines model-based stock ranking with trend and timing filters for a selected universe of China A-shares. The author describes ranking stocks by predicted prospects, then screening for recent relative strength, liquidity, large-order money…
This Chinese research note proposes using capital flow ratios within style-factor portfolios to guide style rotation. It argues that raw inflow and outflow totals mostly move alongside prices, while flows measured against trading value or by…
This brief BigQuant forum entry asks whether an already running simulated trading instance automatically follows edits made to its original strategy definition, or remains on the earlier version. The response indicates that the running simulation can be…
This BigQuant strategy note describes a stock selection approach based on a fast and slow moving average crossover, identified as MACD-style golden and death crosses. It buys after an upward cross and treats a downward cross as a sell signal. The universe…
This short troubleshooting note explains why a pandas table may appear out of order after a sort. In the reported case, rows were first sorted by date and a field named double_low, then sorted again by index, which disrupted the intended ordering. The…
This report studies whether the most unusual intraday price bar contains useful stock-selection information. It defines extremeness as the absolute distance between a bar’s return and the median intraday return, then examines the bar with the greatest…
This tutorial walks through a basic Skip-gram word2vec training workflow. A center word is represented by an input vector, and the model predicts each word in its context window using a softmax output layer. The article describes calculating the prediction…
This study compares six cross-validation approaches for tuning machine-learning models on financial time series. It argues that ordinary shuffled K-fold validation can leak temporal structure and select overly complex models. Expanding-window time-series…
The document gives a conversion procedure for obtaining forward-adjusted stock prices when a platform supplies backward-adjusted prices and their adjustment factors. The first step is to recover the unadjusted market price by dividing the backward-adjusted…
This short market commentary presents a July 2020 view of Chinese indices and sector positioning. It describes the Shanghai Composite and SSE 50 as correcting while potentially forming daily-chart continuation setups, and says the ChiNext remained in a…
This December 2024 market note tracks sentiment, fund crowding, and two stock-selection approaches in Chinese equities. It describes breadth and moving-average sentiment measures as elevated, with momentum indicators pointing upward. It also reports that…
This forum post asks how to implement minute-level backtesting for a stock strategy that receives signals at the daily level and refines exits using intraday prices. The proposed rules would sell a holding after it rises by a specified amount and then…
This research summary examines additional ways to monitor crowding in equity investment factors. Factor crowding refers to a situation in which heavy investment in a factor may weaken its returns or make them less stable. The analysis considers asset…
This summary of a 2020 study examines equity factor strategies in China’s A-share market, where market structure and regulation can affect how familiar investment signals behave. It reports that carefully designed traditional factor strategies delivered…
This fund analysis describes a Chinese equity strategy that combines quantitative signals with qualitative judgment. Its inputs include mean reversion in returns and valuations, sector rotation, fundamental reversals, event effects, investor behavior and…
This overview introduces ten supervised learning approaches: linear and logistic regression, linear discriminant analysis, decision trees, naive Bayes, k-nearest neighbors, learning vector quantization, support vector machines, bagging and random forests,…
This platform support discussion reports sparse cross-sectional coverage for two company growth fields: year-over-year growth in free cash flow and year-over-year growth in adjusted net profit. The author says the free cash flow field has only about 800…
The document summarizes research on a listwise learning-to-rank loss designed for constructing long-short stock portfolios. The proposed loss builds on ListMLE but gives extra emphasis to the top and bottom of the ranking, where long and short selections are…
This document describes an equity factor based on intraday volume moving through a low-to-high-to-low cycle. It smooths volume with a nine-minute neighborhood, identifies a peak and the lowest-volume points before and after it, then measures price-change…
The article presents a workflow for translating a written trading specification into a QMT strategy. Its example applies ADX and directional indicators to one Chinese-listed equity: buy when ADX exceeds a threshold and positive directional movement is…
This short strategy summary describes a quantitative stock-selection approach built with TabNet, a machine-learning model. It says the approach extracts 98 price and volume factors, trains the model on data from 2010 through 2018, and applies predictions to…
This document explains why cash dividends matter when pricing Chinese equity index futures and estimating hedge costs. Because the contracts track price indices, expected constituent dividends affect fair value and the apparent futures basis. The proposed…
This study groups 50 Wind order-flow indicators into eight categories, then tests whether they help rank A-share stocks. Its framework combines industry-neutral portfolio sorts, cross-sectional regressions, and information coefficient analysis. The tests use…