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 introductory note defines quantitative investing as using data and a processing method to guide investment decisions, with the aim of predicting returns consistently. It lists potential advantages: model-based decisions, historical performance that can…
This summary describes version 3.0 of Xingye Securities’ Crystal Ball timing model, which uses activity in options on the SSE 50 ETF as a source of investor sentiment and potential leading information about the underlying. Its premise is that options…
A VeighNa community discussion asks how to display a five period average of bar lows on the strategy’s built in candlestick chart during a backtest. One reply shows how to calculate the current value from the strategy’s array manager, but the original poster…
This post describes a Chinese equity screening rule combining turnover, recent price movement, and the ten-day moving average. It selects stocks with turnover between 3% and 12%, three consecutive declining sessions, and an opening price near the ten-day…
This document describes a DAX strategy on hourly bars that uses separate exponential moving average pairs for long and short trades. A 60-period average crossing above a 200-period average opens a long position, while the reverse cross exits it. For shorts,…
This document describes a trading system built around reinforcement-learning agents implemented in an MQL library. An agent accepts a vector of predictor values and contains an ensemble of trees; the example supplies normalized closing prices as inputs and…
The document describes a simple long-only strategy for the S&P 500, based on the premise that the broad index trend is upward and that short trades would target corrections. Rather than buying a fresh high, it seeks a pullback followed by signs of renewed…
The document presents an outline for a VeighNa training course on machine-learning-based cross-sectional multi-factor strategies, also described as alpha strategies. It frames these methods as a form of statistical arbitrage applicable to equity selection,…
The document surveys ten cryptocurrency exchanges for automated trading and introduces basic bot setup. It compares venues through stated attributes such as available trading pairs, geographic availability, reputation, liquidity, fees, matching speed, and…
This intraday breakout strategy defines a range from 09:00 to 09:30 and applies it to the DAX on a 15-minute timeframe. It places buy-stop and sell-short-stop orders at the range high and low, respectively, during the subsequent session. The stop loss is set…
This article proposes screening stocks using price amplitude, large-order net-volume ranking, and a turnover range. It interprets these as rough proxies for volatility, trading activity, and liquidity, respectively, with the aim of finding candidates for…
This proposed Chinese-equity screen combines three criteria: membership in the metaverse theme, a curved price pattern, and a stock code beginning with 60. The article sketches a Python workflow that obtains stock data, computes a custom measure from daily…
This MetaTrader Expert Advisor uses signals from the Karpenko oscillator. It opens a signal when a bar closes and the indicator cloud changes color, making the cloud transition the described entry trigger. The EA is presented for GBPUSD on a four-hour chart,…
This Qlib configuration defines a daily stock-ranking experiment for the CSI 300 universe, using the CSI 300 index as its benchmark. Its features combine price and volume transformations, including residual and fit measures, intraday range, and rolling…
The described expert advisor uses no indicators. At the start of a new bar, it opens two positions in opposite directions and then waits until both positions close through their stop-loss or take-profit levels before opening another pair. The approach relies…
The document describes research that extracts information from high-frequency order-book data and turns it into two groups of factors. It says the researchers aggregate these measures into monthly signals and evaluate their stock-selection performance in…
This document outlines a long-only mean-reversion setup for the DAX on a 30-minute chart. It looks for a prior low below the lower Bollinger Band, followed by a close back above that band while remaining below the upper band. Entry also requires a close…
The report summary challenges the traditional claim that a doji candlestick predicts a market turning point. It says tests found that returns associated with ordinary doji patterns were primarily explained by reversal effects, rather than by the candle…
The post outlines a Chinese-equity screening heuristic based on price amplitude, ranking stocks by the day’s auction amount, and requiring a minimum listing age. It frames amplitude as a volatility measure, auction amount as a sign of activity, and listing…
This brief forum exchange addresses where to obtain historical futures data for strategy backtesting. Participants suggest using open-source financial data tools for daily bars, applying to a research service that offers a temporary trial, or downloading…
This tutorial introduces FMZ Quant’s M language through a moving-average trend strategy. The example opens long positions when price and a short moving average are above a longer average that is rising, and opens short positions under corresponding bearish…
This report describes three Chinese equity portfolio approaches designed to compete with active mutual funds rather than broad market indices. The fund-enhancement portfolio selects funds by performance layers, uses their holdings as a universe, and applies…
This troubleshooting note describes a simulation failure in a single-stock AI ranking strategy: the pipeline reports that no data remains after missing values are removed, even though the strategy completes a backtest. It suggests checking whether the data…
This community post asks how to start a live or simulated trading strategy with positions already held in the account, rather than having the strategy assume it begins flat. It also raises a related backtesting question: a user supplied an initial-position…