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.
21,023 documents
This release note describes changes to Zipline 1.4.0, a quantitative research and backtesting platform. It removes implicit downloads of treasury and benchmark data, replacing benchmark retrieval with user-supplied files or instruments, or an option to run…
This indicator displays candlestick size as a histogram in a separate chart window. Its calculation can use either the distance between the open and close, which represents the candle body, or the distance between the high and low, which represents the full…
Using simple fruit-and-money stories, the essay offers an informal account of how land monetization, initial public offerings, asset repricing, and credit creation may relate to China’s property and equity markets. It distinguishes newly monetized land and…
This Chinese equity screen selects stocks with turnover between specified bounds, a listing year of 2021, and at least five overlapping moving averages. The article presents turnover as a liquidity consideration and moving-average overlap as a technical…
The document proposes a short-term stock screen that selects shares with a daily amplitude above a threshold, a high equal to the highest of the current and prior sessions, and a volume ratio within a specified band relative to a moving average. It presents…
The Long/Short Differential Gain indicator compares the best hypothetical long and short outcomes over a lookback period using opening prices only. Starting from the opening price at the beginning of the window, it finds the highest and lowest opens reached…
This proposed stock screen combines amplitude above 1, a shrinking negative MACD histogram on a 15-minute chart, and a daily percentage change between -5 and 2.6. The rationale is that elevated movement can identify active stocks, while a contracting…
Greatest Swing Value (GSV) compares recent upward and downward price movement measured from each session’s open. The indicator concept sums or averages moves from the open to the high and from the open to the low over recent periods, then flags a swing from…
This tutorial outlines a workflow for collecting daily Chinese stock limit-up and limit-down pools, counting the stocks in each pool, and storing the resulting records on the BigQuant platform. It reads the latest Chinese trading date from a trading-days…
The post introduces a financial time-series study using returns on the CSI 300 index. It says that EGARCH and related models are used for comparison, with an effort to fit and forecast both prices and volatility. The topic is volatility modeling through…
The article lays out a process for finding, screening, and preparing algorithmic trading ideas for backtesting. It begins with personal constraints: temperament, available time, capital, programming ability, and whether the trader needs regular income or can…
Power Measure is an oscillator intended to relate price changes to volatility over a configurable period. It calculates percentage change from the previous close to the current close, and a second percentage measure using the current high and the previous…
This indicator identifies directional trends from the ordering of short and long moving averages. It counts consecutive candles for which the short average remains above the long average, or the long average remains above the short average, and reports a…
This overview explains how unsupervised learning finds structure in unlabeled data, contrasting it with supervised prediction. It presents K-means clustering and principal component analysis (PCA) through equity examples. K-means groups stocks using scaled…
This module fits a bivariate mixture of Clayton, Student-t, and Gumbel copulas, motivated by a mixed-copula pairs trading approach. It first maps each input series to empirical cumulative probabilities, then estimates component parameters and mixture weights…
The document contains a short example that applies ordinary linear regression to the latest 30 closing prices from a crypto exchange data feed. It assigns each observation a sequential time index, calculates the regression slope and intercept, and evaluates…
This BigQuant example describes a stock-selection workflow that calculates each stock’s 15-day price change and 20-day average turnover from local CSV files. It ranks stocks by 15-day change, keeps the lower half, joins those names to a current stock list,…
This code describes a stateful method for comparing consecutive limit-order-book snapshots. It stores bid and ask levels from the current and previous snapshots, then flags each current level as unchanged, changed, or inserted based on price and quantity.…
This strategy turns changes in a spread series into long and short entry thresholds. It separates historical spread changes into positive and negative values, then calculates a chosen upper quantile of positive changes and a lower quantile of negative…
The document describes a third generation moving average designed to reduce the delay of a conventional moving average by increasing the moving average period according to a simple procedure. It attributes the method to Manfred G. Dürschner and says the…
This report outlines a method for analyzing open-ended bond fund returns and deriving a residual alpha measure for fund screening. It groups risk exposures into eight factors spanning interest-rate curve movements, credit risk, equity markets, and money…
This document defines an On Balance Volume oscillator as the difference between OBV and its moving average. The default averaging period is 20, and the code constrains the chosen period to a range of 1 to 999. The resulting difference series is intended to…
This indicator note explains the Range Expansion Index (REI), a momentum oscillator attributed to Thomas DeMark. It compares selected changes in highs and lows with the absolute size of those changes over a defined lookback. Conditions based on prior highs,…
This VeighNa community contribution describes additions to a backtesting statistics engine for evaluating strategies with regressed annual return (RAR), R-Cubed, and Robust Sharpe. RAR is calculated by regressing cumulative returns across time intervals and…