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
The document outlines a Chinese stock screen using three filters: a daily price-change measure greater than one percent, a listing age exceeding one year, and a closing price below 12 yuan. It presents the filters as a way to find active, established,…
The document describes a tick-data compressor that stores changes in bid, ask, and time rather than repeating full tick records. Small price and time changes can fit into a compact representation, while larger differences use additional bytes. It also offers…
This discussion examines whether the length of a model’s training window changes an AI strategy’s results. It describes manually rolling training for a visual template strategy, comparing longer histories of five to ten years with shorter windows ranging…
This indicator combines a Laguerre-filtered RSI with a parameter that adapts to a fractal energy calculation. The energy measure adjusts the filter’s gamma sensitivity, while the resulting oscillator is displayed alongside it. The description identifies the…
This document presents a hand-coded version of a KDJ-style indicator intended to match the implementation described by TradingView, after the author observed that TradingView and FMZ produced different values. The calculation first finds the highest high and…
This stock screen selects shares whose daily high-low amplitude exceeds 1%, whose price is above its five-day moving average, and whose 2021 revenue is more than 1.1 times its 2018 revenue. The proposed rationale is to combine recent price movement,…
This short indicator note introduces William Blau’s double-smoothed stochastic, attributing it to a 1990 article. It identifies the calculation’s inputs as the current close, the lowest low and highest high over a lookback period, and exponential moving…
The document describes Temporal Routing Adaptor (TRA), a way to extend a stock prediction model so it can learn from different patterns in market data. It notes that momentum and reversal behavior may coexist, which challenges the assumption that…
The document answers how to allocate weights across strategies in a multi-strategy backtest. Its proposed workflow is to extract each strategy’s daily return series and use an optimization package to find portfolio weights. This frames the task as portfolio…
This Chinese-language post describes an A-share stock screen combining three conditions: a 14-period RSI below 65, an outer-to-inner trading volume ratio above 1.3, and circulating market value between 5 billion and 10 billion yuan. It outlines a screening…
The document describes a price oscillator modified by incorporating volume before the indicator's final smoothing step. Because this changes the oscillator's scale and smoothing behavior, the overbought and oversold thresholds must be recalculated for the…
The document describes a basic workflow for evaluating a trained quantitative model. After fitting the model on training data, apply it to a validation set, then compare its predictions with the observed values to assess performance. This separates model…
This research summary explains how to build a machine-learning stock-selection process using historical factor values to predict subsequent returns. In the training stage, a supervised model learns the relationship between inputs and returns; in the testing…
This MQL5 demonstration illustrates supervised classification with a support vector machine (SVM), using a fictional animal-recognition task to explain labeled examples and learned decision boundaries. It generates seven-feature observations with rule-based…
This helper prepares spread changes and their lagged values as inputs for a regression model. It can expand the lag features with pairwise products, split a chosen in-sample period into ordered training and test sets, and keep a separate out-of-sample…
This example turns a CAPM regression into a monthly stock-selection process. It takes a recent window of daily returns for eligible constituents, adjusts stock and benchmark returns by a stated daily risk-free rate, and regresses each stock’s returns against…
This brief coding question outlines a way to calculate fund performance statistics from a price series. It first derives periodic returns from price changes, then uses a performance-analysis library to compute cumulative return, annualized return, Sharpe…
The document summarizes CapTE, a model for predicting stock movements from social media text. A Transformer encoder extracts semantic features from posts, while a capsule network is used to represent structural relationships in the text. The approach is…
The document introduces Temporal Routing Adaptor (TRA), a model designed to learn multiple trading patterns from stock market data. It describes using TRA with Qlib datasets and workflows, and notes that the paper’s reproduction setup first trains a backbone…
This article introduces the autoregressive moving-average model as a combination of AR terms, which use past observations, and MA terms, which represent past shocks. It describes choosing the orders p and q with autocorrelation and partial autocorrelation…
The document summarizes a study that develops a probabilistic classifier to identify high-frequency trading activity from intraday order data. Using French BEDOFIH market records, the researchers engineered features describing orders, including their prices,…
This indicator example applies spectral filtering to the Average Directional Index (ADX). It transforms the ADX time series into frequency components and removes higher-order harmonics to produce a smoother series. The stated motivation is to reduce…
This Chinese stock-selection rule screens for companies associated with the metaverse theme, with the previous day’s actual turnover rate between 3% and 28%, while excluding stocks on the STAR Market. The stated rationale is that turnover may reflect…
The indicator applies linear regression to a price momentum series, then plots a lagged copy of the regression line as a signal line. The suggested entry cue is a crossover between these two lines. Its parameters control the momentum lookback, regression…