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.
16,761 documents
This guide explains how a backtest engine turns each OHLCV bar into four synthetic book updates, processes resting orders along that path, and dispatches the completed bar to strategies afterward. It covers timestamp conventions, venue and book requirements,…
This research summary examines whether inventory cycle phases can help time broad sectors and the Shanghai Composite. It describes four phases—active restocking, passive restocking, active destocking, and passive destocking—and compares qualitative phase…
The document raises a data-timing question about China’s Dragon-Tiger List, a report of stocks with notable trading activity. It names fields for returns one, two, and five days after a stock appears on the list, and says the five-day return is unavailable…
This Chinese Q&A discusses basic factor research and strategy refinement. It recommends using a correlation calculation to compare factors and gives rough absolute-correlation bands for weak and strong relationships. It also suggests tracking factor…
This research note develops an equity factor from intraday changes in volatility, which it treats as uncertainty about volatility. Using rolling five-minute windows of returns and volatility, it identifies periods when this uncertainty is elevated. It then…
This document outlines a stock screen using daily price movement, convertible-bond information, and a reported increase in holdings. Its stated conditions are an absolute price change of at least one percent, a nonempty outstanding convertible-bond name, and…
The document reports a failure during a VeighNa CTA backtest when the engine calculates performance statistics. The traceback points to annual downside-risk calculation, where the Empyrical statistics package accesses NumPy’s removed `NINF` attribute. NumPy…
This Chinese-language strategy note describes a screen for mainland Chinese equities in the metaverse sector. It selects stocks whose prior-day turnover exceeds 8% and whose MACD is above zero, with the DIF line also above the signal line. The stated…
This Chinese brokerage research summary examines Fuling Zhacai, a single Chinese listed company, and proposes adjusting exposure according to how much future earnings its share price appears to anticipate. The analysis describes relatively stable analyst…
This BigQuant workflow builds a Chinese stock portfolio from a broad exchange and sector universe. Its factor query computes lagged price measures, ranks stocks using longer- and shorter-window returns, filters out special-treatment stocks, and removes rows…
The document describes an Expert Advisor that trades signals from the BSI indicator. A signal is generated when a bar closes and the indicator histogram changes color. The system requires a compiled BSI indicator file, and it notes that a supporting trade…
The document outlines an Expert Advisor built around the intersection of two moving averages. It states that take-profit and stop-loss levels are virtual, and that the moving-average parameters are hard coded rather than adjustable. Thus, the described…
The document describes an expert advisor that trades from signals generated by the wlxBW5Zone indicator. A signal is recognized when a colored icon appears at a bar’s close. Running the advisor requires the compiled indicator file, and an accompanying trade…
This Expert Advisor trades signals from the Fisher Cyber Cycle oscillator. It generates a signal when the oscillator’s main line crosses its signal line on a completed bar, making the closed-bar crossover the strategy’s stated decision rule. The document…
This research summary explores technical factors built from turnover, illiquidity, and combined price-volume information in the CSI 500 universe. It considers turnover stability, turnover variation, abnormal turnover, several illiquidity measures, On-Balance…
This brief forum exchange answers whether data sources for VeighNa, an algorithmic trading platform, are free. One user reports authentication errors while trying credentials from an example configuration for a Tushare integration. A respondent explains that…
This script creates order labels from per-stock historical backtest data, then saves the resulting records in training, validation, test, and combined datasets. It rejects a stock’s data if it is empty, contains missing values, or has very small minimum…
The document explains tick volume as a count of deals during an intraday period. Because reported futures trade volume can arrive a day late, tick counts offer a more timely way to observe activity. They do not reveal the number of contracts in each deal, so…
This guide explains how to inspect an AI agent’s decisions during backtests and live or paper trading. It describes per-run Parquet records, per-call JSON traces, summary logs, and machine-readable artifacts. These records expose prompts, tool calls and…
This document presents an A-share stock screen combining a daily price-movement condition, a minimum circulating market capitalization, and more than two limit-up sessions within a recent ten-day window. It frames these filters as a way to capture…
The article explains why futures research often uses constructed continuous series: individual contracts expire, so a longer history requires rules for selecting and joining successive contracts. A direct splice can create an artificial price gap at a roll,…
The document explains a MetaTrader 5 backtesting workaround for cross-pair instruments. During a test, the platform may load and synchronize an additional symbol to convert the main instrument’s profit currency into the account currency. The article argues…
The document surveys a quantitative research platform’s components and workflow. It describes tools for strategy development, financial data access, factor research, model training, backtesting, simulated trading, and live execution. Its examples span stock,…
The document summarizes research on applying conditional generative adversarial networks to calibrate individual systematic trading strategies and combine them into ensembles. The proposed workflow covers training and selecting a cGAN on time-series data,…