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
This report summary examines how disclosure requirements and historical accuracy of earnings previews differ across Chinese equity-market boards. It compares preview disclosure rates, reports that the ChiNext board generally has the broadest coverage and the…
This forum post raises a portfolio-selection question about factor neutralization when the stock universe changes over time. The author describes selecting the five highest-scoring stocks from an initial group, then adding more stocks the next day and…
The document describes an indicator that displays Money Flow Index (MFI) calculations as a sequence of candlesticks. Instead of presenting MFI only in its conventional indicator form, it processes the relevant price time series with the MFI algorithm and…
This note describes a Disparity Index indicator that compares the closing price with an exponential moving average. It calculates the standard deviation of that difference over the same lookback, then plots upper and lower thresholds at a chosen multiple of…
This indicator builds five nested ZigZag levels: the first uses a configurable lookback, and each higher level is derived from turning points in the level below. It is designed to condense price action into progressively larger swings and can optionally…
This brief help note answers how to test Python code with breakpoints, including whether a workflow similar to PyCharm’s breakpoint debugging is available. It suggests using Python’s built-in pdb debugger to pause execution and inspect a program while it…
This note describes a weighted moving average whose price weights follow a parabolic curve. A power parameter controls the weighting pattern, extending the basic idea of a linear weighted moving average. The document explains two special settings: power zero…
This research overview outlines a framework for constructing high-frequency equity factors from Level 2 market data. It recommends reducing tick or intraday observations into daily measures, then transforming those measures over time with operations such as…
This stock selection example combines three filters: turnover between three and twelve percent, outside volume greater than 1.3 times inside volume, and a positive institutional direction measure. The article frames turnover as a liquidity or activity…
This indicator adds a trend-strength display to a smoothed Derivative indicator. It measures the indicator’s standard deviation and marks the moving average with colored dots that reflect the current trend direction. Two thresholds divide the display into…
This Chinese-language research summary examines factors associated with the assets of index funds, separating them into fixed product attributes, accumulated long-term characteristics, and changing short-term variables. Fixed attributes include fund…
This document describes a MetaTrader workflow that records historical real ticks once, then replays them in mathematical mode for strategy testing. Sharing a tick file among optimizer agents is intended to avoid repeated data writes and reduce the time spent…
The post raises a reproducibility question: two strategy templates reportedly use the same model, factors, parameters, and data, yet produce substantially different backtest results. The author asks which result is correct and whether one template contains…
This report evaluates 18 financial-quality, price-volume, and analyst-expectation factors for Chinese equities. It compares factor-ranked portfolios, including high-ranked and low-ranked groups, considers the stability of their performance differences, and…
This discussion examines how starting portfolio composition can distort a simple account-value profit calculation for a cryptocurrency strategy. It compares two accounts following the same price move: one begins with a bitcoin and no cash, while the other…
The Nadaraya–Watson estimator, attributed here to Nadaraya and Watson in 1964, estimates values with a locally weighted average. A kernel function assigns the weights, producing a smoothed series that can help a reader visually assess a possible trend. The…
Qlib is introduced as a modular platform for researching quantitative investment strategies with AI and machine learning. Its components are loosely coupled, so parts of the platform can be used independently. The architecture is organized into…
This Chinese-language document proposes a stock screen combining three conditions: amplitude above a stated threshold, a shrinking negative MACD histogram on a 15-minute interval, and the stock price above its five-day moving average. The rationale is to…
Polarized Fractal Efficiency (PFE) is presented as an oscillator for identifying trend direction. Its configurable inputs are fast and slow rate-of-change periods, a moving-average period, and a scale factor. The calculation derives short- and long-horizon…
This document explains insertion sort, a method that builds an ordered list one item at a time by inserting each new element into its proper position among those already processed. It highlights the algorithm’s stability, constant extra-memory use, and…
This research overview introduces six broad model families used in investment work: neural, graphical, clustering and encoding, linear, tree-based, and ensemble methods. It describes their general strengths, such as nonlinear function fitting, relationship…
This guide explains how to obtain and read monthly partitioned stock bar data at four intraday frequencies, then align the local compressed tables with corresponding cloud tables used for prediction. It describes the Feather file layout, recommends loading…
This document describes an exponential moving average that can incorporate a smoothed trend component. Its settings are the EMA coefficient, the trend coefficient, a switch for enabling trend adjustment, and the price series to apply. The input price is…
This document explains an MT5 indicator that tracks runs of consecutive rising and falling bars. Separate plots show upward and downward runs, and each count resets when a bar prints in the opposite direction. A trader can set a threshold around commonly…