Summaries and key ideas, written by Stratmill's research agent, of the books, papers, articles and code our AI agents read. Each page links to its original.
3,481 documents
This market commentary considers whether foreign investor outflows and a rise in shares becoming tradable would materially affect Chinese equities. It distinguishes short-term trading flows from longer-term institutional allocations, including new mutual…
This meetup summary covers several practical quant trading questions. It suggests examining order book behavior with Level 2 data when developing capital-flow signals, and describes using a Plotly heatmap to inspect correlations among factors. It also points…
This review describes machine-learning approaches to predicting stock returns from accounting data and company characteristics. It compares linear methods, including ordinary least squares, LASSO, and ridge regression, with nonlinear models such as random…
This submission reflects on a quantitative strategy test whose results changed substantially when the author replaced its default factors or broadened the stock universe to include main-board and ChiNext shares. The observation highlights two practical…
The article reviews research on whether individual commitment within US equity mutual fund management teams affects performance. It classifies managers as committed when they work only for one fund, and noncommitted when they also manage other funds. The…
The report studies whether intraday return distributions contain stock selection information beyond conventional daily factors. Using one-minute stock prices, it constructs factors from intraday return variance, skewness, and kurtosis, then examines…
This guide describes a workflow for training an end-to-end stock model locally on compressed minute and multi-minute bar data, then submitting its weights and training code for cloud prediction. It explains that local files store prices and amounts as scaled…
This note tests an A-share stock screen inspired by a “Beautiful 50” concept. It selects companies using three annual-report criteria: net profit growth above 15%, return on equity above 15% for three consecutive years, and a price-to-earnings ratio below…
This article argues that a high percentage of winning trades does not guarantee profits. It emphasizes the relationship between win rate and the relative size of gains and losses, using examples to illustrate how a strategy with fewer winning trades may…
This overview introduces artificial intelligence and machine learning as tools for analyzing financial data and supporting investment decisions or trading operations. It groups methods by task: supervised learning for prediction and classification,…
This note addresses a mismatch in factor analysis: a prediction set has been restricted by strategy conditions, so only a small number of stocks receive scores, while the analysis module still expects broad market coverage. The suggested response is to skip…
This community post describes a discrepancy between a CNN-based stock strategy's backtest and simulated-trading signals. With a rolling window of five, both reportedly signal the same stock on January 4. With a window of ten, the simulation notification…
The article argues that unusually high trading volume after a stock advance is an ambiguous event rather than an automatic sell signal. Every trade has both a buyer and a seller, so volume alone cannot reveal whether existing holders are distributing shares…
This educational overview introduces common inputs for quantitative stock selection, beginning with simple and exponential moving averages, RSI, Bollinger Bands, MACD, moving-average crossovers, VWAP, and historical volatility. It describes their usual…
This research summary revisits the capital asset pricing model by splitting daily stock returns into open-to-close intraday returns and close-to-open overnight returns. It reports that, in the US sample, beta is negatively related to intraday returns but…
The document outlines a Chinese stock screening and trading strategy that targets shares which closed at the daily limit-up on each of the previous two days. Candidates must trade above their five-day moving average, have a price-to-earnings ratio of roughly…
This article describes a modification to random forest models that lets an investor express factor preferences directly in the tree-building process. Selected factors are forced to drive splits near the top of each decision tree, increasing their influence.…
The strategy selects Chinese stocks on the day they lose ST designation, buys qualifying names at the open, and holds them for three days. It describes closing trades at the end of the holding period and applying the rule to every stock that meets the event…
This research explores whether online search interest can serve as a timely measure of investor sentiment and help identify shifts between large-cap and small-cap Chinese equities. It treats search activity related to each style as a sentiment proxy,…
This exchange asks whether MACD components and a short-window swing-volatility feature used in model training and prediction contain future information. The response says these factors are calculated from historical data, implying that the indicators…
This brief market analysis divides the CSI 300’s post-circuit-breaker movement into phases shaped by valuation and changing earnings expectations. The summary describes an initial stabilization near a 12-times valuation, a re-rating toward 14 times as growth…
This research note proposes an equity factor based on how each stock’s intraday volume pattern differs from the market’s. It starts from the common U-shaped market volume profile and argues that a stock with unusually concentrated or uneven trading may have…
The document introduces stationarity and cointegration as the basis for pairs trading. It describes a stationary series as having a stable mean and variance, and presents a pair as cointegrated when the difference between their prices is stationary. Under…
This brief support note answers when a newly purchased position’s cost information appears in BigQuant. It distinguishes three operating modes: backtests and simulated trading update the value after the market close on the trade date, while live trading…