Large Language Models and Generative Factor Discovery in Crypto Markets
“The top long-short factor delivers 658.01% annualized excess return with a Sharpe ratio of 4.13, and we find that leading signals remain robust in a strictly forward out-of-sample window defined by the model's training cutoff, even during a sharp downturn.”
AssetCryptospot crypto
StrategyNews sentiment
TechniqueLarge language model
ClaimSharpe ratioalpha
Also stated: Long-short, Text and news, Order book, out-of-sample test
Missing from abstract: benchmark, costs, multiple testing, universe
Institutional-grade audit of this strategy.
£5 — Request the audit- Audit of the paper claims accounting for costs, luck, crises, originality, geography, etc.
- Reproduction code integratable into Bloomberg, IBKR, etc.
- Bias-free datasets for backtesting included
- Delivered in 2 working days
The paper's neighbourhoodNews sentiment and text signals by year
The family is counted across the whole universe. Asset, method and authors are counted inside the family.
| Similar papers | citation | same method | same asset | same authors | |
|---|---|---|---|---|---|
| More Strategies, Same Zero: Multiple Testing Against LLM-Scale Alpha Search by Mehndiratta | 2026 | ||||
| From Hypotheses to Portfolios: NeuraQuant for Auditable Crypto-Alpha Discovery with Large Language Model by Zheng and Wang | 2026 | ||||
| EVOQUANT: Self-Evolving Verifier-Guided Strategy Optimization for Robust Quantitative Trading by Mao et al. | 2026 | ||||
| MadEvolve: Evolutionary Optimization of Trading Systems with Large Language Models by Kvasiuk et al. | 2026 | ||||
| QuantHarness: Price-Driven Multi-Agent LLMs for High-Frequency Trading by Xiong et al. | 2026 |