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 paper summary challenges common ways of measuring whether assets diversify one another during poor market conditions. It argues that estimating downside correlation only on observations when both assets perform poorly omits cases where one asset gains…
The Sine Weighted Moving Average is a price smoother that takes a finite window of observations and assigns weights based on sine values. The weights correspond to positions across the window, with the oldest and newest observations receiving different…
This post describes a Chinese equity screen using a capital-strength measure, moderate turnover, and historical revenue growth. Capital strength is illustrated as trading value divided by tradable market capitalization. The turnover range is above 2% and…
This article explains time inconsistency in investment decisions: a plan that appears optimal today may no longer suit the investor later as circumstances or preferences change. It contrasts rigid pre-commitment with a dynamic equilibrium strategy designed…
The report outlines a holdings-based framework for classifying actively managed Chinese equity funds by market capitalization, growth, and value characteristics. It uses semiannual or annual full-portfolio holdings, standardizes style measures, and combines…
This brief indicator description explains a channel display derived from a price histogram. It marks the histogram's upper boundary with a red line, highlights a central area where 60% of trades occur, and shows a wider area covering 90% of trades. Blue…
These reading notes discuss James Simons and Renaissance Technologies as an example of systematic investing built by combining mathematical research, statistical signals, computer models, and automated execution. The notes contrast discretionary investing,…
The document implements a threshold autoregressive model for testing whether a spread adjusts differently after positive and negative deviations. It first differences the input series to form changes, lags the spread by one period, and assigns each lagged…
The document describes a Chinese A-share stock screen that excludes Beijing-listed shares and selects from the Shenzhen main board. It filters for turnover between 3% and 12%, price-to-earnings ratios from zero to 29.01, and price-to-book ratios from zero to…
This indicator adapts the Commodity Channel Index by placing Bollinger Band boundaries around it to define overbought and oversold areas. Its inputs include the CCI lookback length, applied price, band deviation, optional smoothing, and horizontal shift. The…
This brief note introduces a Color Parabolic indicator based on the iSAR indicator and describes its application to the Relative Vigor Index oscillator. Its central idea is that iSAR uses current price time-series data and can therefore be applied to other…
The document briefly identifies an idea credited to Eugene and MQL5 code authored by Vladimir Karputov: analyze open, high, low, and close values on bars numbered 0 through 3. This implies comparing OHLC data for the current bar and the three preceding bars,…
This introductory overview surveys regression, neural networks, support vector machines, clustering, dimensionality reduction, and recommendation methods. It distinguishes numerical prediction from classification, explains how logistic regression turns a…
This research note outlines a cross-sectional regression framework for studying historical returns of technical, expectation, and financial factors. It uses sequential single-factor regressions: each stage models the preceding stage’s residual, so later…
The document explains how to translate two Tongdaxin indicator functions into BigQuant operations. A crossover, described as one series moving above another, can be expressed with a conditional filter and shifted values. The example applies this to a bullish…
This user report describes a discrepancy in Chinese stock prices around corporate actions. The author says that prices derived from a platform’s close and adjustment factor did not match prices shown by two brokerage or financial-data applications for Ping…
This indicator plots account balance and equity over trading history, with optional displays for used margin and free equity. It can also show period returns, maximum drawdown in absolute and percentage terms, profit factor, and recovery factor. Filters…
The document introduces excess kurtosis and skewness as statistical measures for classifying market conditions. It proposes treating negative excess kurtosis as a sign of a trending market, then using positive skewness to indicate an upward bias and negative…
The document summarizes a 2019 study that combines two representations of the same stock data to predict prices: numerical time series and images of stock charts. An LSTM extracts temporal patterns from the sequence data, while a CNN extracts visual patterns…
The document describes a daily factor based on the linear-regression slope of a stock’s minute-bar volume during the first 30 minutes of trading. A positive slope represents volume increasing over that window; the suggested interpretation is that early…
This document describes a slow stochastic oscillator displayed in a separate chart window. It first calculates a fast stochastic value from the close's position within the recent high-low range, then applies a simple moving average over the D period. A…
This brief description introduces an indicator formed by applying the Inverse Fisher Transform to a nonlag-smoothed RSX oscillator. It notes that RSX is presented as a smoother alternative to RSI and that nonlag smoothing can be turned off by setting its…
This short note describes an indicator that applies the Inverse Fisher Transform to a nonlag-smoothed version of RSX, a momentum oscillator related to RSI. It characterizes RSX as smoother than RSI and says the nonlag smoothing can be disabled by setting its…
This document describes a Chinese stock screening rule that selects shares with daily amplitude above 1%, a listing history longer than one year, and a volume ratio between 1.5 and 6. It explains the intended rationale: larger ranges may indicate active…