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.
24 documents
The document examines how crypto asset returns across five blockchain ecosystems move in relation to one another. It reports that sharp gains on one chain are often accompanied by losses on others, a pattern that differs from the positive co-movement often…
The document describes CryptoTrade, a cryptocurrency trading agent built around a large language model. Its approach combines on-chain information, which the authors characterize as transparent and immutable, with off-chain signals such as news, which can…
The study examines whether news and social media text can improve Bitcoin and Ethereum price forecasts when combined with financial and blockchain data. It compares pre-trained and fine-tuned language models with dictionary-based sentiment methods, and uses…
The document introduces Real-CATS, a dataset intended to support the development and evaluation of cryptocurrency address detection methods. It combines addresses identified as criminal in real-world reports with benign addresses associated with exchange…
This study uses internal Bitcoin network data and external features to predict daily price direction with multiple machine learning models. It evaluates those forecasts through a trading strategy on unseen observations gathered during the first quarter of…
This research addresses how Bitcoin mixing services obscure the paths of funds through transactions and addresses. The authors describe methods for tracking mixer activity and identifying mixing transactions, mixer addresses, and sender and receiver…
This study tests whether on-chain flows predict intraday returns and volatility for Bitcoin, Ethereum, and Tether. It examines one-to-six-hour intervals over 2017–2023 and distinguishes flows into exchanges from flows among investors’ wallets. The authors…
The document compares quantitative patterns in cryptocurrency markets with stylized facts associated with established financial markets. It studies return distributions, volatility clustering, temporal multifractal correlations, cross-market correlations,…
The document presents a framework for measuring price improvement from onchain order flow auctions (OFAs). It describes attributing execution gains to adjustable system inputs, including route quality, gas use, and priority fee settings, with the goal of…
The paper models Bitcoin block arrivals and price jumps using a multivariate Hawkes process, a point-process framework that represents events as capable of exciting later events. It uses publicly available blockchain data and estimates model parameters by…
This study examines whether information beyond price history can help forecast Bitcoin–dollar volatility. Its proposed CoMForE system combines historical trading data with sentiment from related tweets, search-volume measures of public interest, and…
The Lockchain Protocol uses Bitcoin transactions’ mandatory four-byte nLockTime field as a compact metadata header for a meta-protocol. By restricting encoded values to an unused range of past Unix timestamps, it assigns fields for a protocol signal, type,…
This study examines Bitcoin holding durations and age-dependent transaction activity using blockchain data. It reports a heavy-tailed holding-time distribution, with an average power-law exponent near 0.9 across durations from one day to at least 200 weeks.…
This chapter surveys how wealth and trading behavior evolve in cryptocurrency systems that issue new coins through Proof of Stake. It first considers wealth accumulation without trading, with attention to how the protocol may affect decentralization. It then…
The study examines whether cryptocurrency on-chain measures can help explain prices and distinguish risky assets. It analyzes historical data, measures correlations between price and other parameters, groups cryptocurrencies using clustering, and applies…
ManiScope is a visual analytics system designed to help users assess trade-based manipulation risks in cryptocurrency markets, including wash trading. Its coordinated views bring together token ownership distribution, relationships among holders, individual…
This study examines whether publicly displayed trading intentions reduce execution costs, using protocol-native TWAP orders on Hyperliquid as an example of “sunshine trading.” It compares visible TWAP executions with hidden metaorders reconstructed from…
This study asks whether two independently configured data collectors observe the same tokens on the Solana pump.fun launchpad. One pipeline detects cohorts associated with coordinated early-buyer wallets; the other records decisions from a trader-side…
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 paper asks whether cryptocurrency transactions can be linked across separate blockchain ledgers, extending the study of tracing activity within a single currency. It focuses on exchange services that convert one cryptocurrency into another, which 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 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 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…
The study examines fraudulent tokens on Uniswap V2 using transaction data to characterize trading activity and identify suspected scams. Its detection approach combines a guilt-by-association heuristic with machine-learning techniques, aiming to flag tokens,…