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 tutorial introduces recurrent neural networks as models for sequential data, then explains long short-term memory networks as a design intended to address the difficulty ordinary RNNs have retaining information across long sequences. It describes the…
This indicator compares buying and selling pressure using candle movement relative to the prior close, scaled by recent averages and weighted by volume. It presents three readings: a histogram measuring the difference between regression-smoothed buying and…
This research summary explains how stock exposure to macroeconomic indicators can inform equity decisions. The proposed logic is conditional: when an indicator is expected to rise, favor stocks with positive or statistically strong sensitivity; when it is…
The document explains the Correlation Trend Indicator (CTI), which compares a price series with an ideal straight trend line using a correlation measure. It describes short- and long-period CTI values as distinct views of trend rather than simply smoothed…
This introductory tutorial explains machine learning concepts and linear regression with finance examples. It distinguishes features from labels, outlines training, validation, and test data, and contrasts supervised with unsupervised learning,…
This screening idea combines three conditions: daily amplitude above 1, a shortening negative MACD histogram on a 15-minute chart, and reported revenue in 2021 exceeding the 2018 figure by a factor greater than 1.1. The document presents the revenue…
A z-score expresses how far an observation lies from its mean in standard deviation units. Positive values are above the mean, negative values below it, and zero corresponds to the mean. The document also gives familiar reference proportions for observations…
This short Chinese-language post introduces SciPy as a Python library for statistical and optimization techniques, following a brief mention of NumPy and pandas as common tools for data work. Its stated focus is using SciPy for statistical analysis. For a…
MegaFX Profit is a smoothed oscillator that measures the median price's position within the recent highest-high and lowest-low range. It transforms that normalized position and applies smoothing, which the description says makes signals clearer but delayed.…
The document describes an indicator that smooths On-Balance Volume (OBV) by filtering higher-order harmonics in its spectrum. It presents this as a reusable approach for smoothing other indicator time series as well. The stated benefit is practically zero…
The document presents a Chinese equity screening rule for companies classified in the metaverse industry. It selects stocks whose previous-day actual turnover rate is between 3% and 28% and whose price-to-earnings ratio is positive. It identifies the concept…
This brief description presents an indicator that plots correlations between eight currencies and gold. It lets users select a technical measure for the calculation, including moving averages, MACD, stochastic, RSI, CCI, RVI, DeMarker, momentum, triple…
This extended Commodity Channel Index (CCI) indicator offers several ways to alter the calculation. It can use a regular CCI or adapt its behavior with a Vertical Horizontal Filter (VHF). Before calculating CCI, users may apply a simple, exponential,…
This Python class assembles a subset of Alpha101-style equity signals from price, volume, VWAP, returns, and precomputed factor series. Its methods apply operations such as cross-sectional ranking, rolling correlation, covariance, time-series ranking, decay,…
This report describes an initial step in building a multi-factor equity model: choosing which factors to combine. It proposes reducing the factor search space by including exposure relative to a benchmark as one screening criterion, alongside the strength of…
The document describes a MetaTrader script intended to identify times of day when an asset tends to move more. It proposes analyzing historical price movement across a chosen number of trading days, then presenting the results as a column chart that traders…
The guide explains how a crypto order book records resting bids and asks by price and size, and how those levels can help assess current liquidity. It describes measuring the spread, cumulative depth near the midpoint, estimated slippage by walking the book…
The indicator defines an opening range over a user-selected intraday window, then projects three target levels above and below its high and low. Targets can use range multiples or Fibonacci ratios. It records whether each target was reached in each completed…
This report tracks equity factors using information coefficient (IC) summaries and the relative returns of factor-ranked long portfolios. It describes ranking stocks by factor values, selecting the top decile for each factor, and weighting holdings by the…
This post describes a Chinese stock screen requiring daily amplitude above a threshold, positive institutional money flow, and a reversal-style technical condition. It includes sample formulas and Python logic. The examples approximate the reversal condition…
This discussion addresses whether a trader should use hyperparameter search to find an optimal combination of existing factors. The response says that “optimal” needs a clear definition: it could mean the best backtest result or the best factor-analysis…
This overview traces major developments in deep learning from the perceptron and backpropagation through convolutional and recurrent networks, generative models, residual networks, Transformers, and BERT. It describes the basic contribution associated with…
This post outlines a Chinese stock-selection rule that requires turnover between 3% and 12%, year-over-year growth in net profit attributable to parent-company shareholders above 20% and no more than 100%, and a nonempty name for an outstanding convertible…
This brief forum post asks how to translate a BigQuant expression into a newer SQL interface. The expression groups a daily count of securities whose limit-status condition is true, then divides that count by its 180-period mean. The post also suggests…