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.
79,386 documents
This configuration sets up a Qlib experiment that uses a gated recurrent unit model with Alpha360 features to rank CSI 300 constituents. Feature values are robustly normalized with outlier clipping and missing-value filling; labels use cross-sectional rank…
This note proposes screening Chinese mainboard stocks using an RSI below 65, a daily gain above 1%, and prior-day trading value above 60 million yuan. The stated intent is to combine a moderate RSI reading, recent upward price movement, and a liquidity…
This note describes a Chinese equity screen combining daily price range, free-float market capitalization, and the direction of the 30-day moving average. It selects stocks with an intraday range of at least 1%, free-float capitalization above 10 billion…
This brief description presents a utility for showing the time remaining before the current candlestick closes. It is aimed at chart intervals that may be less intuitive to track, including intraday and multi-hour periods. The practical purpose is to let a…
This ProOrder indicator draws a zigzag by identifying local highs and lows over a symmetric window and connecting alternating turning points. Its configurable depth parameter determines how many bars are considered around a candidate fractal: increasing the…
This short-term A-share screening idea focuses on stocks classified in the metaverse theme that appeared on the prior day’s exchange trading-activity list, while excluding stocks that reached the stated near-limit-up threshold. The article expresses these…
This A-share stock selection idea combines a turnover filter with a short-term price decline and a longer-term trend condition. It screens for stocks with turnover between 3% and 12%, seven consecutive down days, and weekly MACD above zero. The article…
This forum thread concerns a futures simulation account that connects successfully but receives no market data. A respondent suggests that the user may not have connected to the trading server, which is used to retrieve contract information. Another…
This note proposes a Chinese equity screen requiring price amplitude above 1%, a reversal or engulfing-style pattern, and positive net buying attributed to major participants during the opening auction. It explains the signals as indicators of short-term…
This note describes a simple Chinese equity screen combining daily price amplitude above 1%, a closing price of 18.5 yuan, and a positive return below 35% over ten days. It frames the filters as a way to find volatile stocks with moderate prices and recent…
This reading note distills three rules attributed to the book The Ghosts of Wall Street. First, treat a new position as unproven: reduce or exit if the market does not confirm the trade within a reasonable period, with survival and small losses taking…
The article explains event-based tick data through changes to a limit order book: orders arrive, are canceled, or trade against resting quotes. With this event stream, a researcher can reconstruct the visible book, subject to venue rules and the depth…
This Chinese equity screening idea combines three filters: a daily increase in positions above 5%, previous-day actual turnover between 3% and 28%, and a dividend ratio above 25% for 2019. The accompanying explanation treats position growth as a sign of…
This announcement describes proposed changes to how Hummingbot Foundation would allocate maintenance work across its open-source trading software. The Foundation reported that connector upkeep consumed substantial effort while community trading volume and…
This document describes an indicator that displays power-weighted moving averages calculated on three selectable timeframes over a price chart. A trader can set the calculation period and power, choose how the values are drawn, and select the timeframes for…
This document outlines a Chinese equity screening rule that combines daily price movement with a short moving average and a lower low. The stated final criteria are amplitude above 1, the close crossing above a five-day moving average, and today’s low…
This brief description introduces a MetaTrader indicator that displays trend information from the SimpleBars indicator at a fixed timeframe. It mirrors the source indicator’s bar colors, with marks on the signal line appearing when a bar in the selected…
This stock-screening example combines a technical condition with reported earnings growth and trading activity. It selects A-share stocks with a 14-period RSI below 65, year-over-year growth in parent-company net profit from above 20% through 100%, and a…
This article examines how training-window length affects an AI stock-selection model. It compares long windows, expanded year by year from 2005 to 2021, with shorter windows ranging from a month to a few years. The author recommends checking each window…
This screening concept selects stocks tagged with the metaverse theme, whose close is above the prior day's low, and whose weekly five-period moving average has crossed above the weekly ten-period average. The article interprets the industry label as a…
This stock-screening idea combines three conditions: MACD above zero, upward-moving averages, and price above the five-day moving average. The accompanying Python example calculates MACD from exponential moving averages and compares the five-day average with…
This implementation describes convergence trading for two cointegrated assets as a portfolio optimization problem. It estimates error-correction speeds and other model parameters from price data, then computes portfolio weights under both unconstrained and…
This stock-selection screen combines three conditions: daily range as a percentage of the prior close above a threshold, the current high equal to the highest high over the current and previous day, and the 20-day moving average above the 120-day moving…
This expert advisor uses the Chaikin Oscillator to classify conditions as flat or trending, then looks for directional entries when the market is not classified as flat. It examines a configurable run of oscillator readings and labels the market flat when a…