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
This work uses nonnegative tensor factorization (NTF) to identify trading activity patterns across multiple time scales in online financial markets. The method is first demonstrated with synthetic data, then applied to interbank transactions from Italy’s…
This study models price formation in a standard call auction by treating buy and sell orders as random draws from valuation distributions. An equilibrium condition yields distributions for both the clearing price and traded volume. Bid and ask volumes remain…
This article studies optimal investment and consumption in an incomplete market when preferences are expressed through nonlinear g-expectations. Trading strategies are subject to general constraints, so the investor’s choices must satisfy restrictions as…
The document describes constructing strategies by classifying market states instead of directly forecasting prices or returns. Labels define states, and market features such as moving averages are selected for their ability to distinguish those labels. A…
The document develops a model for how firm characteristics can shape asset returns in equilibrium. It assumes investors' demands depend separately on firm characteristics and asset log-prices, while asset supply is exogenous and not driven by factors. With…
The document examines whether models that map market conditions directly to portfolio weights can outperform conventional cross-asset futures strategies. It describes training these policies on sixteen liquid CME futures with a differentiable Sharpe-ratio…
The document examines whether current market trends can help forecast future volatility and correlations. It proposes using trend strength as an input to risk forecasts, extending earlier work that relates trend strength to expected returns. The approach…
The study analyzes a strategy in the Black–Scholes framework that borrows at the risk-free rate and holds a long stock position until the price reaches a deterministic barrier. It derives analytical expressions for the expected value, variance, and…
The paper examines statistical properties of Bitcoin returns, compares their dynamics with those of conventional currencies, and considers behavior across time scales. Its empirical sample consists of transaction data from a single Bitcoin platform over the…
The paper studies alternating direction implicit time-stepping methods for numerically solving the three-dimensional Heston–Hull–White partial differential equation. It first discretizes space with finite differences on nonuniform grids, then assesses ADI…
The document outlines a proposed USDCHF trading strategy that uses gold market dynamics as an intermarket filter. It models both markets with a coupled hidden Markov model, linking their state behavior rather than analyzing the currency pair in isolation.…
The paper presents a general framework for deriving an optimal trading strategy from a historical indicator. It constructs one strategy intended to maximize expected return and a separate strategy intended to maximize the information ratio. The…
The paper evaluates whether social media sentiment and attention measures can help forecast daily log returns for ten cryptocurrencies. Its predictors include Twitter and Reddit sentiment, Google Trends indexes, and trading volume. The forecasting method is…
The document examines how encrypted commit-then-reveal transaction systems may leave perpetual futures funding signals exposed to self-authored state manipulation. An attacker who knows its own transaction and holds a claim affected by the changed state can…
The document describes a cryptocurrency trading system that combines trend information across multiple timeframes with neural networks that predict near-term price direction. It brings market prices, on-chain measures, and order-book activity into a shared…
The paper presents a way to model volatile time series using v-transforms, which connect quantiles of a stationary series’ marginal distribution with quantiles of a predictable volatility proxy. Expressing these transforms as copulas lets the model combine…
The study proposes a stock-pair selection distance measure that accounts for lead-lag relationships between the two stocks. It starts from the observation that correlated securities may not move simultaneously: one can tend to respond before the other.…
The paper introduces TenSeg, a method for detecting structural changes over time across multiple related transaction networks. It addresses a challenge in fraud and market-manipulation analysis: activity may span several cryptocurrency platforms, so a shift…
This study applies deep reinforcement learning to continuous futures trading, examining both discrete and continuous action spaces. It incorporates volatility scaling into reward functions so that position sizes respond to market volatility. The algorithms…
This paper introduces a price-to-utility ratio intended to compare Bitcoin’s market price with a blockchain-based measure of fundamental utility. The authors derive the measure using monetary theory and Bitcoin’s UTXO accounting structure, then compare it…
This theoretical study modifies the continuous-time Kyle insider trading model by allowing the market maker to pursue profit rather than act as a fiduciary. The market maker has market power and uses information in order flow to set prices advantageously.…
This study tests whether market impact follows the square-root law using anonymous order flow for Apple stock. It reconstructs metaorders from Nasdaq TotalView-ITCH events across 178 trading days, then estimates impact against daily volatility and traded…
The paper addresses a limitation of historical replay backtests: they do not model how the market might respond to a strategy’s actions. It describes multi-agent simulation as a way to represent interactions among traders with different strategies, allowing…
The paper evaluates two deep reinforcement learning methods, Double Deep Q-Network and Proximal Policy Optimization, for trading. Their performance is compared with a buy-and-hold benchmark using daily data from 2019 to 2023. The assets described include…