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 strategy ranks a fixed universe of ten large U.S. stocks by price momentum over roughly one trading year. On the first trading day of each month, it selects the three strongest performers, sells positions outside that group, and adjusts each selected…
This Freqtrade strategy template describes a short-term, long-and-short system using a five-minute chart. It enters when price is above or below a simple moving average calculated on a resampled, longer interval, and a short-period EMA crosses the…
This MQL5 Wizard signal design combines pullback conditions with candle-pattern direction. A long entry requires price above a moving average, RSI below a configurable level, and bearish-side candle action; a short entry reverses those conditions. A spread…
This small utility module provides basic operations for preparing timestamped market data. It estimates samples per day from the first observed interval, estimates elapsed days between the first and last timestamps, and partitions a dataframe into monthly,…
This code describes a deep learning approach for turning sequential market features into position signals. Its example model uses an LSTM layer followed by dropout and a time-distributed output constrained through a hyperbolic tangent activation. Training…
This introductory multi-factor stock selection method combines one profitability measure, return on invested capital, with two valuation measures, price-to-sales and price-to-book. The stated goal is to rank stocks by factors associated with returns and…
This reference summarizes exchange-specific behavior relevant to configuring an automated cryptocurrency trading system. It compares supported spot and futures markets, margin modes, and available on-exchange stop orders, then discusses API rate limits,…
This post proposes a Chinese equity screening idea that combines several moving averages converging, an opening price near the 10-day average, and a positive but bounded 10-day return. It interprets converging averages as a sign of stable trends and recent…
This MetaTrader expert advisor uses the nearest ZigZag vertex on a new bar to reverse positions. A vertex at a bar high triggers a buy and closes existing sell positions; a vertex at a bar low triggers a sell and closes existing buy positions. A…
This stock selection approach ranks securities by volume ratio and keeps the top 100, then looks for simultaneous bottom crossovers in MACD, KDJ, and RSI. It also requires at least two limit-up sessions within the prior 500 days. The stated rationale is to…
This note describes an equity screen using daily price amplitude above 1, a positive weekly MACD signal, and prior-day actual turnover between 3% and 28%. Its refined version also requires market capitalization above 5 billion yuan. The proposed rationale is…
This stock screen selects Chinese equities with a daily high-low range greater than 1%, excludes stocks that hit the price limit on the previous day, and restricts the historical selection period to 2021. A later version adds a price-to-earnings ratio below…
This meetup Q&A covers practical choices in quantitative strategy design. It says researchers can choose the forward-return horizon used as the target when calculating information coefficients, and points to model-based and scoring templates. For…
This tutorial explains interface features for a trading strategy environment. It shows how a status display can present changing text, variables, formatted tables, and encoded images without filling the log with repeated updates. It also describes chart…
This report summary describes revisions to AlphaNet, a neural network that learns stock-selection factors from price and volume data. AlphaNet-v2 adds six ratio features, replaces pooling and fully connected layers with an LSTM to capture temporal patterns,…
This Chinese stock-selection note describes a screen combining turnover, price movement, large-order net volume, trading volume, and a price near 18.5. Its central condition selects stocks with turnover between 3% and 12% and a positive product of price…
The document outlines an Expert Advisor that generates a trading signal when the ColorJFatl_Digit indicator changes color at a bar close. It requires the compiled indicator to be installed in the platform's indicator folder. A supporting library is mentioned…
This article describes a portfolio component for an event-driven backtester. A portfolio receives signal events, produces order events, and updates positions and holdings from fill events. The NaivePortfolio example tracks each symbol’s quantity, cash,…
The document explains how FinRL models automated stock trading as a Markov decision process. An agent observes market prices and features, acts in a simulated environment, receives rewards, and adjusts its policy to pursue higher cumulative reward. The…
This document explains a graphical workflow for researching CTA strategies with historical data. It covers downloading market data, configuring a backtest with instrument details and trading costs, and reviewing equity, drawdown, daily profit and loss, and…
This short forum exchange explains that backtest prices and simulated-trading fills can differ because the two systems may use different execution models. A backtest may value an order at the price associated with a configured time point, while a simulated…
Millenium is described as a system that opens at most one trade per day, using a fast and slow moving-average crossover alongside a price-position filter. For a buy setup, the fast average crosses above the slow average and the current price is above both;…
This document describes a rules-based intraday strategy that takes short positions in the DAX on 15-minute bars. Entry requires the MACD to cross below zero while its signal line is negative, price to be below a long-period exponential moving average, and…
This short forum exchange addresses how to model a monthly rollover strategy that trades futures and options together. The question is how to configure the backtest engine when the two instrument types have different transaction fees, especially when closing…