Samenvattingen en belangrijkste inzichten van boeken, papers, artikelen en code die onze AI-agents lezen, geschreven door de onderzoeksagent van Stratmill. Elke pagina verwijst naar het origineel.
1,033 documenten
The document presents a model-free reinforcement learning method for sequential optimization when outcomes are evaluated using dynamic convex risk measures. It uses a time-consistent dynamic programming principle to value policies, then derives…
The document introduces TrendFolios, a portfolio construction framework that combines momentum and trend-following signals across several asset classes and asset-class risk factors. It describes an active investment strategy and says the framework is…
This paper studies whether intraday price overreactions in Apple shares can be forecast and traded as momentum signals. It combines volatility-normalized returns with emotion features derived from Twitter messages, defines overreactions relative to…
The paper investigates whether CryptoQuant signals, including on-chain, exchange, and miner data, and whale-alert posts can help forecast next-day Bitcoin volatility, especially extreme spikes. It proposes a Synthesizer Transformer forecasting model and…
This study examines how a large trader’s finite period of steady buying or selling affects market impact in the Minority Game. The meta-order changes the market environment and creates statistical arbitrage opportunities, prompting other traders to adapt…
The paper addresses a gap in volatility modeling: intraday high-frequency data are unavailable between market close and the next open, so models may omit important overnight dynamics. It proposes a GARCH-Itô framework that represents open-to-close and…
The document introduces FutureQuant, a Transformer model for futures markets that uses attention mechanisms to process complex inputs, including real-time limit order book data. Instead of forecasting a single future price, it aims to estimate a range of…
This research examines high-frequency observations from European cryptocurrency markets to characterize intraday activity associated with algorithmic trading. It analyzes returns, trading volumes, periodic patterns in volatility, and return correlations with…
This study tests two expectations about fixed-income relationships: that stock prices and yields move in opposite directions, and that central bank rate changes help predict later stock-market direction. It applies the thermal optimal path method to monthly…
This paper presents a model-free reinforcement learning framework for reallocating a portfolio among financial assets. Its components include an ensemble of independent evaluators, memory of prior portfolio weights, online stochastic batch learning, and an…
This document explains how open interest in cryptocurrency perpetual swaps relates to market activity, liquidity, and potential exchange leverage. Since outstanding contracts can indicate a minimum collateral requirement, aggregating open interest across an…
This study analyzes high-frequency realized volatility and its transmission among bitcoin instruments, comparing bitcoin quoted in US dollars with bitcoin traded against tether. It finds that Binance's tether-margined perpetual contract is the dominant…
This document studies how to manage positions across several mean-reverting assets, including whether an asset with no mean reversion can still belong in the portfolio. It formulates position management as an optimal control problem for an investor with…
This document models a broker serving both informed and uninformed clients as an infinite-horizon stochastic control problem. It derives a closed-form optimal dealing strategy and uses that result to propose an algorithm intended to work in live trading…
This paper proposes a deep recurrent model that uses exchange order flow to model directional price movements at high frequency. Order flow is described as the stream of arriving orders that contributes to price formation. The central question is whether a…
This thesis backtests strategies built around the conventional MACD settings of 12, 26, and 9 on stocks represented in the Dow Jones, Nasdaq, and S&P 500. It evaluates win rate, profitability, Sharpe ratio, trade count, and maximum drawdown over a period…
The study tracks price, trading volume, and bid-ask spreads following unusually large 15-minute price changes in stocks listed on the NYSE and NASDAQ. It examines how these market measures evolve after the event, with attention to whether sharp moves are…
This study examines possible wash trading in Ethereum-based NFT markets, where traders can operate multiple addresses without identity checks. It analyzes sales from the 52 largest NFT collections by volume, covering January 2018 through mid-November 2021.…
PPO-HRAP combines Proximal Policy Optimization with a regime-derived target exposure. Its policy observes market features and portfolio state, then blends the learned action with the regime target. The reward accounts for portfolio log returns, increases in…
The document describes a mean-reversion approach that begins with yield-curve trading strategies and extends the idea to a multi-pair strategy involving major currency pairs. To refine its trading signals, the strategy incorporates machine-learning forecasts…
This study examines high-resolution USD-JPY exchange-rate data spanning 13 years. Its focus is on statistical patterns in the distributions and correlations of exchange-rate changes, with particular attention to whether price movements depend on recent…
This study presents a method for identifying likely one-way arbitrage (OWA) sequences in spot trade records from centralized cryptocurrency exchanges, where public data lacks trader identities or wallet addresses. The approach infers linked trading sequences…
This theoretical study asks when several causal risk mandates, each built from overlapping information, can be implemented by one self-financing portfolio that is optimal given pooled information. In an incomplete continuous Brownian market, it projects…
The paper describes an equity trading system that combines trend and momentum signals from exponential moving averages and MACD with mean-reversion signals from RSI and Bollinger Bands. It adds financial sentiment estimates from FinBERT, uses XGBoost to…