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
A user asks why backtest results change when the number of historical days fetched before a test is increased, even when the strategy calculates a moving average such as a ten-period average. The question assumes that fetching at least the indicator’s…
This post describes a simple stock-selection screen for Chinese equities. It seeks stocks with a daily trading range above a threshold, upward movement in short-term moving averages, and circulating market capitalization above a stated floor. The…
This article explains Reflex and Trendflex, two price indicators attributed to John Ehlers. Both first smooth prices with a SuperSmoother low-pass filter, then measure average distance between the smoothed price path and a reference line. The distance is…
This research report develops equity-selection factors from four types of intraday information: price behavior, price-volume relationships, pre-open auction data, and selected trading periods such as the close or high-volume intervals. It evaluates 46…
This article compares neural networks, decision trees, random forests, support vector machines, and a stock-ranking algorithm for equity selection. It describes broad trade-offs: neural networks can model nonlinear relationships and varied data structures,…
This report examines corporate tax burdens in China at the economy, industry, and listed-company levels. It distinguishes taxes from mandatory non-tax charges, describes differences across industries, and notes that the shift from business tax to value-added…
This overview explains how factor models can identify portfolio risk drivers and how factor strategies can target their associated return premia. It describes equity factors such as value, size, momentum, volatility, quality, yield, growth, and liquidity,…
This educational-session overview introduces machine-learning and deep-learning approaches to quantitative stock selection. It lists core workflow topics: defining prediction labels, engineering features, separating training, validation, and test sets, and…
The document answers a programming question about applying a per-stock price-pattern test to stocks first selected by daily market filters. The suggested workflow reads hourly bars for the candidate instruments in one batch, then passes the shared data frame…
The document explains the price-to-earnings ratio (P/E) as a way to relate a stock’s price to the company’s earnings per share. Comparing two stocks at the same price but with different earnings shows how the ratio distinguishes their valuations: the stock…
The report explains how option prices can be used to construct volatility and skew measures for the Chinese equity market. VIX represents expected volatility over a forward period, while SKEW captures the market’s pricing of asymmetric returns and tail risk.…
The article presents a discretionary momentum setup for Chinese equities that avoids chasing rapid advances and instead looks for a shift from prior weakness to renewed strength. It characterizes a previous session with heavy trading and a contested limit-up…
This short post introduces a template for monitoring a personally selected stock universe, aimed at individual investors who may trade manually rather than buy many stocks in batches. The stated use is to track dozens or hundreds of shares and view their…
This summary of a Haitong Securities study examines whether the cyclically adjusted price-to-earnings ratio can predict long-run US equity returns. Historically, high CAPE readings tended to accompany weaker subsequent ten-year returns, and low readings…
This study compares loss functions for neural models that forecast five-day cumulative returns for selected Chinese A-shares. It uses a DeepAlpha DNN as the main model and compares mean squared error with absolute error, Pseudo-Huber, negative information…
This research summary explains why predictive machine-learning models can mistake unstable associations for causal relationships, especially when selection bias or confounding drives observed patterns. It presents propensity-score analysis as a way to make…
This lesson explains binary classification through logistic regression, using the example of predicting whether a stock will rise. It describes encoding outcomes as zero and one, interpreting the model output as a probability, and applying a 0.5 cutoff to…
The document introduces a video about applying “做T” trading—buying and selling within a trading period—to a value stock, using China Yangtze Power as its example. Its stated aim is to explore whether this approach can generate returns beyond simply holding…
The article argues that algorithmic trading can exploit retail investors’ predictable reactions to price action and news in China’s A-share market. It illustrates this with two anecdotal patterns: a sharp late-session reversal in high-flying consumer stocks…
This research update describes two A-share stock-selection approaches. The product-price strategy tracks changes in listed companies’ main product prices and highlights cases where prices have risen but share prices may not yet reflect the development. The…
The report builds a composite trading-heat signal from stock-specificity and market-cap-adjusted turnover, using these trading-behavior measures as proxies for speculative activity. It argues that the two inputs capture most of the useful information in a…
The document explains how members access BigQuant market data through its Python SDK. It outlines installation and login, describes checking table permissions, and demonstrates querying selected daily stock fields over a date range. It also shows how to…
This overview explains how quantitative strategies use factors to express hypotheses about future returns. It describes candidate signal sources such as prices, fundamentals, macroeconomic data, and text, then outlines a research process: translate an…
A BigQuant support discussion addresses a user’s difficulty training a StockRanker model for individual-stock prediction. The replies recommend moving to the newer StockRanker template because the older module is no longer actively maintained. A second…