Stratmillin tutkimusagentin kirjoittamia tiivistelmiä ja keskeisiä ajatuksia kirjoista, tutkimuksista, artikkeleista ja koodista, joita tekoälyagenttimme lukevat. Jokaisella sivulla on linkki alkuperäislähteeseen.
3,481 dokumenttia
This article introduces discrete-time stochastic control as a method for dynamic asset allocation. Unlike a traditional mean-variance allocation made without an explicit sequence of future decisions, the approach models portfolio choices across multiple…
This brief support exchange explains why a MACD histogram value can shift substantially when the start date of a stock data list changes. The explanation points to the initialization of the exponential moving average: its first value is not fully smoothed,…
This document describes a Chinese stock strategy that combines classification and regression models to forecast intraday returns. It uses XGBoost and an ensemble of machine learning models, drawing on 15 features related to overnight returns, opening auction…
This intraday system trades the CAC 40 on a 15-minute chart, taking both long and short positions during a defined daytime session. Long entries require the MACD line to cross above its signal line, RSI to remain below an upper threshold, and price to be…
This report summary explores whether macroeconomic indicators can help select individual Chinese stocks, rather than serving only for asset allocation or industry rotation. It proposes estimating each stock’s sensitivity, or MacroBeta, to economic indicators…
This brief BigQuant platform note explains how to use a factor written as a SQL function and field within a strategy workflow. It distinguishes platform expression factors, described as an operator paired with a field, from SQL factors, described as a…
This research review describes a study of whether year-over-year changes in company filings contain information about future stock returns and business performance. It compares filings using four text-similarity measures, sorts stocks by similarity, and…
The report describes a weekly market-timing model based on a linear-chain conditional random field (CRF). It frames the task as inferring market direction states from observed features, with CRFs able to represent dependencies between observations and…
The document compares two ways to build multifactor equity portfolios. Integration ranks securities using a combined score for all target factors and selects a subset. Mixing first builds separate single-factor portfolios, then allocates among them. The…
This overview presents a multi-factor model as a way to explain stock returns through shared risk exposures, factor returns, and stock-specific residual returns. It frames quantitative management as estimating factor behavior and managing portfolio…
This article describes modifying OpenFE’s automated feature engineering workflow for futures research. It adds rolling mean, standard deviation, and sum operators across instrument groups, with preset window lengths, and explains how these operators are…
This personal account explains how an equity prediction model can appear strong in-sample yet fail when applied to unfamiliar stocks or later periods. The described experiment uses fundamental and technical features, including valuation, profitability,…
This example describes a short-horizon stock rotation strategy. It selects shares whose prior-day return was between 2% and 8% and whose turnover was between 3% and 8%, then removes special-treatment shares, halted stocks, and listings younger than the…
This report summary describes a method for refining traditional liquidity factors by incorporating intraday trading structure specific to China’s A-share market. The idea is to adapt a broad factor to market-specific trading patterns, with the report…
The note turns a discretionary stock method into explicit rules: admit a stock to an initial universe when its daily low stays above 99% of the 34-day moving average for ten consecutive days, buy when its profitable-chip proportion exceeds 73%, and sell when…
This Chinese-language example sets up a test of whether stocks that have risen on each of the prior three days are more likely to rise the next day. It defines the condition as three consecutive daily price increases, each exceeding one percent, and labels…
This strategy combines model-based stock ranking with trend filters for a selected universe of China A-shares drawn from the CSI 150 and related large-cap core-asset stocks. It aims to rank candidates for potential five-day strength, then apply timing…
This research note examines an equity-selection factor built from differences in stock price behavior across the trading day. It reports that the original APM factor performed well in its in-sample period, weakened out of sample, and suffered a sustained…
This article is a catalog of beginner-oriented examples for learning TensorFlow and TFLearn. It lists introductory operations, nearest-neighbor and regression models, neural networks, autoencoders, model saving, visualization, and multi-GPU basics. The…
The document compares a traditional Chinese equity strategy that buys the smallest market-cap stocks with a StockRanker model that ranks stocks using market capitalization as its sole feature. Both are evaluated over 2015–2016, illustrating how a…
The article reviews the 2021 revision of China's Shenwan industry classification and its possible uses in quantitative investing. The revision expanded coverage to A-shares, Stock Connect shares, and selected Chinese firms listed overseas, added three…
The document describes a daily process for building a short-term stock watchlist. It starts by ranking stocks by traded value and keeping the top 50. It then removes stocks that gained less than 5% that day, those whose 30-day moving average is flat or…
The document surveys the development of enhanced ETFs in China alongside the expansion of actively managed ETFs in the United States. It presents enhanced ETFs as an effort to seek returns above a benchmark while retaining ETF features such as intraday…
This 2020 China equity-sector note combines macroeconomic regime signals, earnings relative to valuation, trend signals, historical regime matching, and public-fund holdings to discuss sector allocation ahead of interim reports. It argues that weaker growth…