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.
53 documents
The document presents a multi-asset model with two investor types: fundamentalists, who respond to perceived differences between stock prices and fundamentals, and chartists, who use price trends. Both groups make investment decisions by maximizing constant…
The document proposes using Bayesian optimization to select the ATR multiplier and ATR period in the Supertrend indicator. These settings determine how the indicator tracks price and generates trading signals, and the aim is to automate their selection…
The study uses historical prices and technical indicators to classify short-term moves in Bitcoin, Ethereum, and Litecoin. It compares classification methods on their ability to identify upward and downward movements over a one-hour horizon, reporting…
The document argues that trend-following performance can be described by a theoretical Sharpe relationship tied to the signal parameter. It reports an empirical fit to the formula derived by Grebenkov and Serror, and interprets this as evidence that a…
This survey reviews research on cryptocurrency trading, an area shaped by assets that share some features with traditional investments while retaining distinct market behavior. It covers trading systems and platforms, trading signals and strategies,…
The paper applies log-periodic power law (LPPL) patterns, interpreted through a behavioral herding framework, to U.S. Treasury yields. It reports evidence of an antibubble pattern in yields with maturities above one year beginning in October 2000. In this…
This study applies KM₂O-Langevin theory to classify stock price fluctuations as stationary, non-stationary, or intermediate. It uses the classification to choose between two types of signals: moving averages for stationary periods and psychological lines, an…
The document presents KBPT, a pairs trading policy that combines a Kalman filter, Bollinger Band signals, and a neural network based on the KalmanNet architecture. Traditional versions model the relationship between two assets as a linear state-space process…
The paper proposes a model for predicting Bitcoin’s short-horizon price direction by combining hourly price and volume features with sentiment from Reddit. A rolling volatility measure labels each observation as stable or volatile, and a learned gate changes…
This paper derives and proves an upper bound for cumulative returns in trading series, then evaluates conventional technical trading rules across developed and emerging stock markets. Its framing addresses an enduring question: whether these rules can…
The paper proposes a dynamic cryptocurrency portfolio method that combines technical signals with news sentiment. It uses the 14-day Relative Strength Index and Simple Moving Average to represent momentum, while VADER analyzes news sentiment and Google…
The study describes a trading approach that combines sentiment extracted from financial news with technical indicators and time-series forecasting. It uses GPT-2 and FinBERT for sentiment analysis, pairs the resulting signals with momentum and trend…
The document describes a long-only USD/TWD trading agent that combines Quantum Long Short-Term Memory (QLSTM) forecasts of short-term trends with Quantum Asynchronous Advantage Actor-Critic (QA3C). Its state uses QLSTM features and technical indicators,…
The study examines the statistical basis of Bollinger Bands by connecting them to a rolling regression time-series model. It applies the framework to pairs trading and derives a relationship between trade duration and returns for strategies based on the…
The document tests stock-selection rules based on maximum drawdown and the recovery that follows it. Across various equity markets, portfolios formed from these measures are reported to predict price direction and capture differences in returns across…
This paper proposes focusing day-trading activity on a single blue-chip stock rather than trading across many securities. The rationale is that specialization may help a trader build familiarity with one stock while applying established intraday approaches,…
This proposed system uses volatility relationships among large stocks to generate trend signals. It first explores price data with correlation and autocorrelation analyses, technical indicators, hypothesis tests, statistical models, and variable selection.…
The document presents QTMRL, a reinforcement learning agent designed to make equity trading decisions using technical indicators. Its input dataset combines daily price and volume data for representative S&P 500 stocks across sectors with features intended…
The paper proposes a strategy selection process called DNN-forwardtesting. Instead of choosing a trading strategy solely by its results on historical prices, it first forecasts a possible future price path with a deep feed-forward neural network and…
This paper compares a model-based optimal trading strategy with a technical rule when the asset’s trend is latent and follows an Ornstein–Uhlenbeck process. It focuses on how each strategy performs when the model parameters used to design it are incorrectly…
The document presents a general workflow for designing, implementing, and assessing algorithmic trading strategies across stock and crypto markets. It uses four examples: a moving average crossover, a volume-weighted execution approach, sentiment-based…
This study predicts short-term directional changes in a futures contract using features drawn from technical analysis, order flow, and order-book data. It trains a TabNet deep learning model on silver futures listed on the Shanghai Futures Exchange,…
This paper proposes a two-variable hidden Markovian time-series model for price inefficiencies, extending beyond geometric Brownian motion. One variable, the reasonable price, is unobserved. In the simplest form of the model, path-integral and…
The study tests whether trading and social media indicators improve cryptocurrency price-change classification beyond technical indicators alone. It evaluates hourly and daily Bitcoin and Ethereum data from January 2017 to January 2021, comparing a…