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 Chinese-language strategy post describes a daily stock screen run after the market opens. It selects stocks with MACD above zero, ranks them by popularity, and requires the previous day's turnover rate to exceed 8%. The rationale is to combine a trend…
This summary of a Chinese financial-engineering report examines whether Northbound capital flows can inform A-share timing and portfolio selection. It describes fixed daily and weekly inflow or outflow thresholds, then a dynamically adjusted timing indicator…
The Expert Advisor described opens two opposing positions when a new bar forms. Its intended outcome is for one position to close at a stop loss while the other can remain profitable through a trailing stop. The example is framed around EUR/USD on a…
This Chinese-language post collects questions about factor research and strategy implementation, but does not provide answers. Topics include adding data factors, scheduling entries and exits, evaluating and combining factors, handling collinearity,…
The document explains how to connect to Binance Spot Demo Mode through the API. Users first create an API key from the demo account interface, then use the standard Spot API documentation while directing requests to the demo endpoints. It lists the…
The document describes a seasonal earnings anomaly: companies may have predictable high- and low-profit quarters, yet announcements of peak or trough earnings can still surprise investors and produce abnormal stock returns. It attributes the mispricing…
This document is a categorized directory of software libraries and learning resources used in quantitative finance. Its sections span numerical computing and time-series data, financial instrument pricing, technical indicators, trading and backtesting,…
This indicator description explains an option to calculate Trend Envelopes from smoothed or filtered prices rather than raw prices. The selected filter is a parabolic weighted moving average, presented as a fast method for filtering price data. The stated…
This expert advisor uses the magnitude and direction of price-time correlation to generate trading signals. It offers trend-following and reversal modes, and four ways to calculate correlation: Pearson, Spearman, Kendall, and Fechner. ADX, Bollinger Bands,…
This indicator seeks bullish and bearish divergences between price swings and a zero-lag MACD. It builds the MACD and its signal line from double-smoothed exponential averages, then tracks price highs or lows and corresponding MACD extremes around…
This report compares an AI-driven portfolio built from a named set of large-company stocks with SPY over a very short backtest window. It presents standard performance and risk measures, including returns, drawdown, Sharpe and Sortino ratios, benchmark…
A VeighNa community discussion addresses why a custom strategy does not appear in the CTP backtesting module’s strategy selector. The reply advises placing a strategies folder in the application’s startup directory, which is usually the user directory, and…
The document describes an Expert Advisor that trades from the ColorBullsGap indicator. It waits for a bar to close and generates a signal when the indicator’s histogram changes color, then uses that signal to open trades. This is a simple indicator-driven…
This short community thread raises possible future-data problems in an equity screening and backtesting tool. One commenter warns that sorting by current market capitalization may select companies based on their later size, making historical results appear…
This example demonstrates a daily-iteration stock strategy that submits limit buy and sell orders alongside two trailing stop sell orders. The orders target the same symbol, while the trailing exits use either a percentage retracement or a fixed price…
This configuration describes a Qlib workflow that trains a CatBoost model on Alpha360 features for China’s CSI 500 universe. Its label is a forward close-to-close return, normalized cross-sectionally after missing labels are dropped. The data is divided into…
This report presents a backtest of a long-only RSI strategy on hourly ETH-USDT candles from Binance, covering January 2020 through May 2021. It reports 108 closed trades, a net profit of 152.12% from a starting balance of 1,000, a maximum drawdown of 31.13%,…
These notes summarize a systematic approach to trading that begins with personal goals, constraints, and psychological tendencies. They review common judgment errors in system design, testing, and live execution, including overfitting, hindsight bias, seeing…
Exp_Fast2 describes a MetaTrader Expert Advisor that trades signals from the Fast2 indicator. A signal is generated when the indicator’s main and signal lines cross as a bar closes. Running the advisor requires the compiled Fast2 indicator file in the…
The script describes a daily gold-futures strategy using the Commodity Channel Index (CCI) with a 10-period lookback and thresholds at 100 and -100. It enters long when CCI moves back from below the lower threshold into the middle zone, and enters short when…
The document explains how Qlib serializes objects such as data handlers, datasets, processors, and models to disk using pickle-compatible formats. A Serializable object saves public attributes by default, with options to configure what is included or select…
This indicator is designed to help assess how spread affects the potential profitability of strategies that forecast the market one bar ahead. Its output can be treated as a unit-volume position in a custom tester, allowing a researcher to consider accuracy…
This practice report compares a DeepAlpha deep neural network with LightGBM in a daily A-share stock-selection strategy. Both models use the same 32 short-term price and volume factors and five-year training sample. The portfolio buys the ten highest-ranked…
This introductory example builds a daily A-share strategy around a five-day and twenty-day moving average. It selects one stock, retrieves recent closing prices, computes both averages, buys to a fully invested target when the shorter average is higher, and…