Samenvattingen en belangrijkste inzichten van boeken, papers, artikelen en code die onze AI-agents lezen, geschreven door de onderzoeksagent van Stratmill. Elke pagina verwijst naar het origineel.
3,481 documenten
This forum report describes a suspected data-alignment problem in a Chinese equity research workflow. The author adds an opening-price feature and a forward-shifted version to a strategy, then finds that the shifted values appear incorrect on particular…
This short Chinese-language forum response explains the meaning of delayed entry in a factor test. A delay of N days is described as shifting the start of stock positions by N days from the original entry point. The purpose is to examine what happens when a…
This post describes a reproduction of a research report's “moderate risk” factor, framed around information contained in moments of unusually high trading volume. The author says intermediate factors are calculated and stored, while factor analysis is…
The document summarizes a study that applies multiple machine-learning algorithms to a broad set of return predictors in Chinese equities. It highlights differences from earlier United States research: liquidity ranks as a particularly important predictor in…
This collection of answers addresses common quantitative investing problems, including fixed-holding-period backtest sensitivity, judging factor data, evaluating strategies that fail live, factor development, market timing, and combining signals across…
The report compares methods for estimating stock-return covariance matrices for Chinese equities. It explains that when there are many stocks but relatively few return observations, sample covariance estimates can be noisy and numerically unstable for…
The report classifies companies into startup, growth, mature, volatile, and declining stages using the signs of financing, operating, and investing cash flows. It describes differences among the groups: mature firms have the strongest profitability, while…
This report reviews the Shanghai Composite as a broad measure of companies listed in Shanghai and discusses changes to its index rules scheduled for July 2020. The changes add STAR Market listings, delay new stock inclusion, and remove specially risk-warned…
The document combines two research summaries. The first explains how target-date fund allocations can account for retirement spending as a liability alongside financial assets and human capital. A liability-aware optimization framework treats…
This report summary examines how to turn a supplied composite stock factor into a long-only index-enhancement portfolio. It contrasts common linear optimization, which is simple to compute but ignores relationships among stocks and can produce concentrated…
This compilation presents three quantitative finance topics. First, it reports that news frequency and tone are associated with stock return jump likelihood and jump size characteristics. News frequency appears especially important for jump probability, and…
The report asks whether property-level real estate returns can be explained by style characteristics in addition to property type and location. Using quarterly data on more than 25,000 UK retail, office, and industrial properties from 2002 to 2020, it…
The article presents four chart scenarios that it interprets as signs of continued accumulation or a possible reversal: low-volume consolidation after a rally, a low-volume pullback near highs, a sharp move to new lows after a prolonged decline, and a…
The article examines whether momentum and long-term reversal reflect changing exposures to priced risks rather than standalone return patterns. It tests whether past returns predict future realized betas, then uses an Instrumented Principal Component…
This installment of an overseas research review introduces two studies, one on strategic asset allocation and another on quantitative stock selection. The allocation study asks whether a portfolio should be rebalanced and how the frequency and size of…
This brief Q&A explains ways to label stock price-limit events in a strategy workflow. For a one-price limit-up session, it suggests identifying a bar where the next period’s high and low are equal, then using that condition to assign a label. This is a…
This brief BigQuant forum exchange addresses how to construct a custom factor when it is not available as a precomputed field. The question proposes an expression combining single-quarter revenue less single-quarter cost, divided by average total assets over…
This post shows a daily stock strategy that combines a market risk filter, a trailing stop, and ranked portfolio rotation. It exits all holdings when a market risk indicator is positive, otherwise checks each position against a stop set below the highest…
This indicator description presents ADXm as a directional variation of the ADX indicator, which measures trend strength without indicating direction. It adds Jurik smoothing to the ADXm series so traders can use its slope as a signal, with the stated aim of…
This legacy BigQuant guide shows how to add fixed exit rules to a stock strategy. Its example sells a holding after a fixed price gain or after a percentage loss, with separate lists recording stocks exited by each rule. It also explains how to coordinate…
The post describes a Chinese A-share strategy that runs on its first day of simulated trading but fails during the next day’s scheduled run. The author suspects the strategy callback is not reached, while the supplied log shows the run progressed through…
The document reports a BigQuant backtest failure caused by passing a floating-point value where the history function requires an integer bar count. In the shown trailing-stop logic, the holding period is calculated as the number of elapsed days since the…
The HAN123 strategy builds a reference range from the high and low observed during the first half-hour of the Chinese equity futures session. After that window, it enters long above the range high or short below the range low, reversing an existing position…
This Chinese research note explores whether stock selection models built within individual industries can add information to a broad market-wide fundamental alpha model. The authors argue that industry models make it easier to use sector-specific factors and…