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 research summary describes a method for studying whether public information-seeking behavior can reveal topics associated with subsequent stock market moves. The cited 2014 study analyzes historical data from 2004 through 2012, drawing on Google and…
This note describes three technical indicators for low-frequency analysis. The BBI is the average of simple moving averages over 3, 6, 12, and 24 days. A second measure, labeled DMI moving-average difference, takes the 10-day moving average minus the 50-day…
This workflow note explains how to prepare data when training models across rolling windows. As each window advances, the training sample changes, and learned processor state—such as means and standard deviations—must be recalculated for that window. Reusing…
The research summary defines a trend turning point as a period when short-window and long-window time-series momentum signals point in opposite directions. It reports that more frequent turning points are associated with weaker risk-adjusted returns for…
The document describes a linear regression line indicator designed to be applied to data from another indicator in MetaTrader 5, rather than only to ordinary price series. It explains the practical motivation: a regression channel drawn as chart objects must…
The document describes a Chinese equity screening rule combining price movement, trading activity, and liquidity. It selects stocks with daily amplitude above 1, a product of the previous day's turnover rate and the ratio of today's opening auction volume to…
This research summary contrasts how high- and low-frequency strategies should be evaluated. Frequent strategies make many short-horizon decisions, so win rate and operational capacity matter; low-frequency strategies hold positions longer, tolerate…
This research note presents a framework for explaining bond fund returns and identifying portfolio style. It decomposes returns into income, government yield-curve movements, credit spread effects, convexity, and bond roll-down. Income is further divided…
The document describes smoothing a DeMarker indicator series by filtering higher-order harmonics in its spectrum. It presents this as a general technique that can also be applied to other indicator outputs. The stated advantage is very low latency, making…
MaRsi-Trigger is an indicator that classifies an asset into three states: rising, falling, or flat. It encodes those states as plus one, minus one, and zero, respectively. Its signals combine two moving averages with two RSI oscillators that use different…
The document presents a workflow for evaluating a stock factor with Chinese A-share data. It builds daily stock returns and industry returns, represents industry membership with indicator variables, and subtracts each stock’s industry return to form an…
This research summary presents a method for assessing analyst earnings forecast errors and whether investors give analyst consensus too much weight. Instead of fitting historical forecast errors to company characteristics, the method predicts future earnings…
This overview describes quantitative investing as using mathematical models, statistical analysis, and computer algorithms to guide investment decisions systematically. It surveys time-series and regression analysis, factor models, automated execution, large…
This stock screen selects shares with a market capitalization of at least 200 million, a 14-period RSI below 65, and a 20-day moving average above the 120-day moving average. The moving-average relationship is used as a short-term-versus-longer-term trend…
This reference compiles expressions for the Alpha191 stock-factor set, attributed in the document to a 2017 research report on short-horizon price and volume characteristics. The expressions draw on open, high, low, close, volume, turnover value, VWAP,…
This report outlines a method for rotating among equity style factors as the macroeconomic environment changes. It proposes two signals: identify macro events historically associated with stronger performance in particular styles, and compare current macro…
This BigQuant question and answer explains the distinction between labels and input features in a supervised machine-learning workflow. A label supplies the target value that a model learns to predict, while the input feature list is used to produce the…
This Pine Script indicator plots a volume-weighted moving average as its center line and builds six upper and six lower bands from the standard deviation of the selected price source. The bands use Fibonacci-related fractions of a scaled deviation, from…
The article describes a Chinese-equity factor that looks for unusually large intraday increases in trading amount, then aggregates information around those moments into daily observations. Its rationale is that a volume surge accompanying modest, orderly…
This document explains how NautilusTrader’s shared network clients add trading-system behavior to HTTP, WebSocket, and raw TCP transports. It covers quota sharing, proxy selection, connection reuse, retries, response limits, streaming deadlines, and…
This document describes a method for turning a set of conditions associated with an ideal uptrend or downtrend into a trend-strength percentage. Each condition that holds earns a point; the total is divided by the maximum possible score and expressed as a…
This report tests thirteen momentum and modified-momentum signals in China’s A-share market. The signals include returns over several lookback periods, turnover-weighted returns, exponentially weighted turnover-adjusted returns, and a long-horizon regression…
This document describes an alert-enabled version of a Spearman rank correlation histogram indicator. It can notify a trader when the histogram crosses zero, breaks through an overbought or oversold threshold, or exits one of those extreme zones. The alert…
The document introduces multi-task learning (MTL), in which one model addresses several related tasks and uses shared patterns and task-specific differences. It suggests that this can improve learning efficiency and predictive accuracy compared with training…