Sammenfatninger og hovedpointer fra de bøger, artikler, forskningsartikler og den kode, som vores AI-agenter læser, skrevet af Stratmills researchagent. Hver side linker til originalen.
3,481 dokumenter
This overview explains a supervised learning workflow for ranking stocks by their expected future returns. It introduces features, a fitted function, model parameters, and a loss function, then describes labeling return targets, splitting observations into…
The article explains play-to-earn games as blockchain-based games that reward activity with tokens or NFTs that may be traded beyond the game. It outlines how a game built on a blockchain can assign digital identities to items, enabling players to hold,…
This forum entry shows a BigQuant initialization example for setting futures commissions through a per-contract commission object. The configuration maps an instrument to separate rates for opening, closing, and same-day closing trades, then passes the…
This BigQuant assignment submission describes a stock strategy with two adjustable parameters and a target of holding 50 stocks. The author says the implementation adds filters for non-ST stocks, non-suspended shares, and excludes Beijing Stock Exchange…
This strategy applies a 20-day channel to a fixed basket of five Chinese stocks. It defines the upper and lower limits as the rolling mean of closing prices plus or minus two standard deviations. A stock is bought when its close exceeds the upper limit and…
The article describes a framework for identifying crowded sectors and estimating whether crowding may be building or unwinding. It measures asset clustering from principal components of sector return covariances, using the concentration of explained variance…
This research note examines how stock trading volume changes across the trading day and how the timing of that volume relates to subsequent returns. It describes a W-shaped intraday pattern in China, with activity rising around the morning and afternoon…
This Chinese-language forum post asks how BigQuant’s StockRanker model works and whether its gradient-boosted decision tree component could be replaced with another ensemble method, such as XGBoost, while retaining a similar ranking strategy. It describes…
This report summary introduces over-the-counter options as customized, non-standardized contracts negotiated between counterparties or arranged through an intermediary. It focuses on binary, or digital, options, whose payoff is discontinuous: at expiry the…
The document describes a Chinese equity screening rule that combines a 14-period RSI below 65, the product of daily percentage change and super-large-order net inflow, and large-order net inflow above 0.05 over three consecutive days. Its stated aim is to…
This research examines how crude oil returns relate to the Chinese equity market and industry returns, then applies estimated oil sensitivity to stock timing, industry rotation, and portfolio risk control. It describes a changing economic relationship:…
This study examines whether company news sentiment can help select Chinese stocks. It describes a news score where higher values generally indicate more positive coverage, then measures changes in the average score over a rolling lookback period. The…
This short forum post recommends adding the Amihud factor to a high-frequency factor table associated with a multi-factor alpha research series. The author says the factor was omitted from the table and characterizes it as a strong long-side factor. The…
The document presents a process for identifying stocks linked to an investment theme using company business classifications and supplier relationships. It starts with leading firms tied to the theme, finds their shared business activities and common…
This introduction explains NumPy’s homogeneous multidimensional arrays and highlights broadcasting, vectorized operations, and mathematical functions as tools for working efficiently with numerical data. It applies these ideas to a short series of…
The report proposes measuring a company’s short-term debt capacity by comparing cash and cash equivalents plus net operating cash flow with interest-bearing debt due within a year. It then builds a related stability factor and discusses two uses: ranking…
The document describes how to combine custom buy and sell conditions with a stock-ranking strategy. Its example selects stocks when the current open is above the prior close and the five-day closing-price average is above the ten-day average. It then ranks…
This report uses a cycle model to assess global markets in 2018. It identifies recurring short, medium, and long cycles in year-over-year financial and economic indicators, and says regression analysis supports their role in asset-price movements. The…
This study reproduces high-frequency stock-selection factors from brokerage research using one-minute OHLCVA data and a sample of roughly 350 constituents of the CSI 300 over a historical period ending in 2022. It focuses on five daily factors formed from…
The author investigates whether the highest-ranked stocks from a convolutional neural network strategy deliver the best results. The experiments apply a model trained on a small number of stocks to broader selections, then backtest separate groups of ranked…
This strategy-sharing article describes an active China A-share approach that blends model-based stock ranking with technical timing. It starts from a manually selected China Securities 150-related universe, described as roughly 100–300 larger, liquid…
This Chinese research note presents a June 2022 outlook for equity style factors and a framework for combining short-term signals with longer-term cycle analysis. It favors valuation, volatility, and turnover factors overall. Its internal indicators are…
This Chinese-language post describes an optimization of a machine-learning strategy template for the BigQuant platform. The author reports reducing a two-year backtest from 65 minutes to 26 seconds by using parallel computation and delayed execution. The…
This learning note outlines a stock-price forecasting pipeline that uses an LSTM as a sequence generator and a CNN as a discriminator in a generative adversarial network. It describes a broad feature set: related assets and rates, technical indicators, news…