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 forum post raises implementation questions about historical data warm-up in VeighNa portfolio strategies. The author considers a strategy whose longest signal period is 20 days and asks whether an ArrayManager size of 25 is sufficient, and whether that…
This trading system combines three Stochastic indicators operating on different timeframes. Two indicators establish directional bias by comparing each Stochastic reading with its signal line. A third, lower-timeframe indicator provides the entry trigger…
The document describes a Chinese stock screening strategy that combines technical and fundamental filters. It selects stocks with at least five moving averages on the daily chart, at least two limit-up sessions within the past 500 days, and a pre-open gain…
This indicator description adapts the pocket pivot concept associated with Chris Kacher and Gil Morales. It flags an up day when the close is above a longer-term moving average and the day's volume exceeds the largest volume seen on a down day in the recent…
The trading system described uses the ColorTrend_CF indicator to detect changes in trend direction. It generates a signal when a bar closes and the indicator’s cloud changes color. The document identifies a historical test on XAUUSD at the four-hour interval…
This research summary examines how sell-side analyst reports may inform stock selection. It argues that report counts and recommendation strength alone provide limited differentiation, while target-price upside and changes in analyst views may be more…
This note describes a stock screen based on three chart conditions: amplitude above a threshold, a rising price base, and a rounded-bottom pattern. It presents the pattern as a way to find stocks whose lows are gradually moving higher, while filtering out…
This article explains why a strong historical backtest may fail in live markets, particularly when a strategy has been tuned to a small or unrepresentative sample. It recommends splitting time-ordered data into a training period for parameter selection and a…
This proposed EUR/USD strategy combines Bollinger Bands, a conventional RSI, a Traders Dynamic Index (TDI), and a custom ATR-based stop line. Long setups begin when price and the momentum measures reach specified lower extremes; the system then waits for the…
This paper summary examines whether historical trading data can predict the next month’s cross-sectional returns of Chinese A-shares. It describes a dataset of 108 stock characteristics from 1997 to 2019 and compares traditional econometric methods with six…
The document presents daily portfolio rebalancing as a Markov decision process. An agent selects nonnegative weights for Dow 30 stocks, normalized to sum to one, using a state that combines a rolling covariance matrix with MACD, RSI, CCI, and ADX indicators.…
This forum post asks how to use pyramiding in a strategy that combines a higher-level long signal with lower-level entry and exit signals. The author wants to add long entries whenever the smaller-scale long condition occurs while the larger long condition…
A forum post asks how to retrieve popular industry indices through iWencai and pass them into a variable. The author describes a query combining industry indices with weekly KDJ conditions and recent main-fund-flow filters, then reports that a Python query…
This article introduces support vector machines for classification and regression, then applies them to A-share stock selection. It explains the maximum-margin principle for linear SVMs, slack variables for imperfectly separable observations, and kernel…
This tutorial explains how support vector machines classify data by finding a boundary with a wide margin, and how slack variables allow some classification errors in noisy data. It introduces kernel methods as a way to handle nonlinear boundaries by…
This expert advisor is built on the idea that price crossing a moving average and traveling a specified distance may continue in that direction. It places Buy Stop and Sell Stop pending orders, then updates their distance from a long-period moving average at…
This documentation explains how to run a TqSdk strategy over historical data without changing its core logic, and how to retrieve trade logs and account statistics when the simulation ends. It describes catching a backtest-finished event, accessing summary…
The document presents a four-hour forex system: a breakout approach for NZD/USD and a reversal variant for AUD/NZD, described as using the same code with long and short orders reversed. The sample rules combine the direction of recent daily closes, a…
The document introduces a 2017 Guotai Junan research report that constructed 191 short-cycle alpha factors from price and volume characteristics and combined them in a stock-selection system. It notes that the factor expressions resemble those in…
This stock screen combines a price-amplitude threshold, a low K-line reading described as an oversold condition, and MACD above its zero line. The stated rationale is to find active stocks that may be oversold while retaining an indicator associated with an…
The example outlines a limit-order market-making loop. It computes a midpoint from the best bid and ask, adjusts a reservation price using a forecast and an inventory-related risk term, then places bid and ask quotes around that price. It rounds quotes to…
This configuration describes a Qlib experiment using a graph attention model, GATs, with an LSTM base model to predict near-term returns for CSI 300 constituents. It sets Chinese market data, defines a close-to-close forward return label, normalizes features…
This forum question examines why changing the initialization length of a trading system’s ArrayManager can materially alter a backtest. The strategy uses RSI generated through TA-Lib, and the author suspects that the indicator’s path dependence makes its…
This post describes an A-share stock screen combining a daily price-amplitude threshold, a nonempty convertible-bond name field, and a company classification filter that can be configured for criteria such as industry or state ownership. It proposes ranking…