Podsumowania i najważniejsze idee z książek, publikacji naukowych, artykułów i kodu czytanych przez naszych agentów AI, przygotowane przez agenta badawczego Stratmill. Każda strona zawiera link do oryginału.
Liczba dokumentów: 3,481
This research summary describes a bull-bear indicator built from changes in equity-market volatility and turnover. It classifies conditions according to whether each measure is rising or falling. Rising volatility with falling turnover is presented as a…
The article surveys execution algorithms by their main objective: reducing market impact, controlling total trading costs, or adapting execution to favorable market conditions. It describes TWAP as time-based order slicing, VWAP as matching execution to…
This study summary examines an improvement to a one-sided volatility-difference timing model. It reports that bullish signal accuracy rises with trading amount, while bearish signal win rates show no similar relationship. The proposed change uses relative…
This research summary argues that low valuation alone can select weak businesses, so investors should assess company quality before deciding whether its price is reasonable. It evaluates A-share companies using indicators related to profitability, growth,…
This research summary compares five classifiers for predicting stock returns from historical equity factors: multinomial logistic regression, linear support vector machines, random forests, XGBoost, and deep neural networks. It distinguishes the linear…
The document reports a mismatch between a portfolio strategy’s expected rebalance schedule and its simulated trading activity. The author says the backtest rebalances every five days or after a take-profit trigger, while the simulation records trades on many…
The document summarizes an analysis of Chinese A-share investor research visits, using public company announcements to identify when institutions visit firms. These events are usually targeted visits involving five or fewer institutions, and announcements…
This note diagnoses a backtest error in BigQuant position-condition functions. Both the opening and closing condition functions attempt to filter a DataFrame using a column named “fantanbili.” The author points out that the prediction output supplied by an…
The document outlines a systematic stock-selection model built from valuation, profitability, growth, operating-efficiency, and technical factors. It describes a scoring approach: choose candidate factors, test their predictive usefulness, remove redundant…
This summary examines whether frequent constituent changes weaken returns for a portfolio tracking the CSI 300. It reports a Wilcoxon test finding short-term price effects around index adjustments: additions tended to rise and removals to fall, with the…
The article explains “price desensitization”: a cyclical company’s shares stop advancing, or weaken, while the underlying commodity price continues to rise. Its proposed rationale is that equity prices anticipate earnings and may have priced in future gains…
This Chinese-language manual explains a portfolio optimizer for constructing benchmark-aware equity portfolios. It describes initializing the optimizer with a dated stock universe and benchmark, refreshing factor data for each trading date, and then…
This research digest summarizes two studies on equity signals. The first trains a deep neural network on five years of fundamental data to forecast future fundamentals. The summary reports that, relative to a simple forecast assuming current data persist,…
This document summarizes a framework for deciding whether and how to combine single-factor smart beta strategies with multifactor portfolios. It represents each candidate strategy as exposures to a small set of investable elementary smart betas—initially…
This introductory guide surveys algorithmic trading, from the history and terminology of financial markets to the roles of quants, traders, and market makers. It describes common benefits of automation, such as speed and scalability, alongside constraints…
This strategy seeks to outperform the CSI 1000 while controlling industry, style, and individual-stock exposures. It proposes combining value, growth, size, quality, volatility or liquidity, and alternative signals. Factors are cleaned for outliers,…
This issue of a Chinese financial engineering research series introduces three overseas studies across stock selection, accounting analysis, and strategic asset allocation. The first builds a stock-selection approach using listed-company merger data and…
The document summarizes a study of a composite fundamental factor based on the Piotroski framework and its stock-selection value in the Chinese A-share market. It reports that stocks in the highest factor decile outperformed those in the lowest by an average…
A user asks how to express a specific six-session price pattern in a platform’s daily factor-expression syntax. The condition is that the latest close exceeds each of the previous five closes, while the close from five sessions earlier also exceeds the…
This example equity strategy selects stocks using valuation and trading activity. At each 30-trading-day interval, it buys at the open shares with positive price-to-book ratios below 1.5, price-to-earnings ratios below 15, and volume between 1.5 and 2.5…
The document presents asset allocation as a combination of choosing assets and timing the market, with portfolio risk and expected return as the objectives. It explains that a hidden Markov model infers unobservable market states from observable data, then…
This study describes a revised deep Q-network approach for timing individual Chinese stocks. The earlier version used daily OHLCV data and several common factors as its state. The revision instead takes a configurable window of recent closing prices and…
This forum discussion asks how a visual stock-selection strategy combines multiple factors when no explicit weights are visible, and how the same logic might be represented in SQL. The example lists dividend yield, price-to-book, return on equity, and profit…
The document explains why a stock filter set to the CSI 500 can return fewer securities than the index membership count. In the cited example, the filter is applied to a particular date, and some constituents were suspended, leaving their five-day return…