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 brief indicator note defines Leadlag as a comparison between two exponential moving averages of closing prices. It takes closing price data and two period parameters, one for the lead line and one for the lag line. The supplied pseudocode calculates…
The article investigates why XGBoost training took far longer in some Python process setups than with joblib’s loky backend. Its experiments compare direct execution and several multiprocessing approaches, then identify the OpenMP thread limit as the key…
This BigQuant backtest handler describes a daily portfolio process driven by a date-filtered prediction ranking. It allocates roughly one holding-period fraction of portfolio value to daily purchases during an initial staging phase; afterward, it can use up…
This report studies whether active-buying measures derived from high-frequency trading data can help rank industries for rotation. It constructs buy-share, buy-intensity, net-buy-share, and net-buy-intensity factors using volume and turnover data from the…
This research note assesses China’s consumer recovery under normalized pandemic controls, using a combination of policy text analysis and indicators of aggregate spending, passenger flows, cinema attendance, and physical store counts. It argues that…
This brief note identifies two errors in the conditions used by a stock strategy example. The first concerns a year-over-year net profit growth factor: the intended filter selects positive growth rather than negative growth. The second concerns position…
The article explains why investment teams need to understand how machine-learning models produce decisions, especially when those decisions affect clients, risk controls, or deployment choices. It contrasts inherently interpretable approaches, such as linear…
This 2020 research summary compares valuations and return drivers across large- and small-cap Chinese equity indexes. It reports that small-cap valuations were near historical medians while large-cap indexes, including the SSE 50 and CSI 100, were near…
The document is a user’s question about calculating three-day and five-day moving averages within a custom feature factor on a quantitative trading platform. It asks how to use lagged close-price fields such as the current-day close, how rolling mean…
The document defines the Sharpe ratio as a way to assess a portfolio or strategy’s excess return relative to the variability of its returns. It gives the calculation: subtract the risk-free rate from average portfolio return, then divide by the standard…
This research summary reviews academic momentum-factor development and reports empirical findings for China’s A-share market. It says traditional momentum measures there exhibit reversal, with stronger reversal over shorter formation windows, and that…
This report excerpt discusses forecasting equity factor returns as commonly used stock-selection factors fluctuate. It proposes improving and simplifying an earlier conditional-expectation timing model by framing factor timing as a regression-based…
This research note studies post-earnings-announcement drift in Chinese A-shares: stocks reporting results above expectations may continue to earn positive abnormal returns, while those below expectations may lag. It estimates expected net profit and revenue…
This stock-selection proposal combines a daily price-range threshold, afternoon large-order inflows and a closing price between the Bollinger middle band and upper band. The document gives formulas for the range and band conditions, describes a…
The document raises a modeling question about predicting whether a future price move is positive using a label that takes only two values. The label compares a later close with an earlier open and assigns one class for an increase and the other for a…
The document reports a discrepancy in a Chinese stock-status data table for one security on a specified historical date. A query against the table returned a status identifying the stock as *ST, while the author says the company was no longer classified that…
The report adapts William O’Neil’s base-pattern and CANSLIM ideas to the Chinese A-share market. It combines price formations such as cup-with-handle and double bottoms with company growth measures, relative price strength, trading activity, and…
The document explains how performance attribution can reveal the sources of a portfolio’s returns, even when portfolios have similar styles and overall results. It outlines three approaches: style analysis, which uses portfolio and benchmark returns to infer…
This article presents a time-series momentum strategy that switches between a short lookback signal and a longer one according to market volatility. The short signal uses one month of returns and the slow signal uses twelve months; each takes a long or short…
This article outlines a systematic trend-following approach commonly called the Turtle method. It describes choosing liquid markets, using average true range (ATR) as a volatility measure for position sizing, and entering on Donchian channel breakouts. The…
This research summary examines whether institutional trading before earnings announcements contributes to mispricing and post-earnings announcement drift (PEAD). It measures abnormal institutional demand before announcements and relates it to subsequent…
The document explains how to build trading agents within a strategy framework. It describes a workflow in which research agents gather market evidence and a dedicated trading agent can submit orders, while deterministic Python can enforce fixed execution and…
This strategy ranks the mainland China A-share universe using company size and a price momentum measure. It removes suspended and specially treated shares, calculates a market-cap rank and the ratio of 20-day to 60-day average closing prices, and adds the…
This note describes using a forecast of the time-varying Sharpe ratio as a monthly signal for switching between cash and Chinese equity indices. It outlines two forecasting approaches: a regression model associated with Robert Whitelaw and an ARMA model. The…