{"paper_id":"2407.09546","title":"A Reflective LLM-based Agent to Guide Zero-shot Cryptocurrency Trading","authors":["Yuan Li","Bingqiao Luo","Qian Wang","Nuo Chen","Xu Liu","Bingsheng He"],"paper_date":"2024-06-27","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2407.09546v1","page_url":"https://thequant.space/flowcharts/a-reflective-llm-based-agent-to-guide-zero-shot-cryptocurren/","scores":{"empirical_rigor":6,"math_complexity":1.5,"quadrant":"Street Traders"},"proposes":{"status":"available","statement":"The paper proposes CryptoTrade, an LLM‑driven trading agent for cryptocurrencies that combines on‑chain and off‑chain data with a self‑reflective mechanism.","quote":"We introduce CryptoTrade, an innovative trading agent in the cryptocurrency domain, driven by LLMs.","summary":"The utilization of Large Language Models (LLMs) in financial trading has primarily been concentrated within the stock market, aiding in economic and financial decisions. Yet, the unique opportunities presented by the cryptocurrency market, noted for its on-chain data’s transparency and the critical…","strategy":"CryptoTrade (GPT-4o)"},"assumptions":{"status":"available","items":[{"assumption":"Trading decisions are executed daily (day‑to‑day frequency).","type":"execution","stated":"explicit","quote":"Another limitation is the frequency of trading actions, which is currently set to day-to-day."},{"assumption":"A proportional transaction fee is applied to each trade.","type":"execution","stated":"explicit","quote":"When a trading decision is made, a transaction fee is charged in proportion to the traded value."},{"assumption":"The model operates in a zero‑shot manner without fine‑tuning on validation data.","type":"model","stated":"explicit","quote":"Notably, CryptoTrade operates in a zero-shot manner without fine-tuning based on validation sets"},{"assumption":"Test period is chosen after September 2021 to avoid data leakage from the LLM’s knowledge cutoff.","type":"data","stated":"explicit","quote":"We carefully select the test period after September 2021, the GPT-3.5’s knowledge cutoff date, to prevent data leakage."},{"assumption":"Initial capital is $1 million split equally between cash and the three cryptocurrencies.","type":"execution","stated":"explicit","quote":"We initialize the trading agent with 1 million US dollars, split equally between cash and BTC/ETH/SOL"}]},"datasets":{"status":"available","items":[{"name":"On‑chain data","provider":"Dune.com","kind":"crypto-onchain","access":"public","period":null,"universe":"BTC, ETH, SOL","quote":"The foundation of our methodology relies on a…"},{"name":"Gnews API news articles","provider":"Google News (Gnews API)","kind":"text-news","access":"public","period":null,"universe":"cryptocurrency‑specific outlets (Bloomberg, Yahoo Finance, crypto.news)","quote":"on filtering news from reputable financial and cryptocurrency‑specific outlets such as Bloomberg, Yahoo Finance, and crypto.news"},{"name":"Market price data for BTC, ETH, SOL","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":"Bitcoin, Ethereum, Solana","quote":"CryptoTrade using GPT‑42, and GPT‑4o3, evaluating its proficiency in making daily trading decisions for Bitcoin (BTC), Ethereum (ETH), and Solana (SOL)."}],"known":[{"id":"bloomberg","name":"Bloomberg","provider":"Bloomberg L.P.","kind":"market-prices","access":"commercial","quote":"Ini-tially, we collect data where on-chain details such as transactions and broader market data are aggregated alongside off-chain data from established financial news outlets like Bloomberg and Yahoo Finance.","mentions":2,"used_in":1}],"sample":{"start":null,"end":null,"market":"BTC/ETH/SOL","asset_class":"crypto","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"unclear","headline_net_of_costs":false,"detail":null,"quote":null,"source":"conflict","rule":"R4b","candidates":{"front_matter":"no","evidence_card":"yes","evidence_card_quoted":true},"conflict_kind":"polarity","card_detail":"A transaction fee proportional to the traded value is charged for each trade.","card_quote":"When a trading decision is made, a transaction fee is charged in proportion to the traded value.","meaning":"the platform classifier and the paper-text extraction disagree, so no value is asserted"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"CryptoTrade outperforms time‑series baselines and achieves comparable performance to MACD across bull, sideways, and bear markets.","kind":"performance","quote":"and achieves comparable performance to trading signals like Moving Average Convergence Divergence (MACD)"},{"finding":"During the ETH bullish test period, CryptoTrade exceeded the Buy‑and‑Hold return of 22.59% by a remarkable 3%.","kind":"performance","quote":"For instance, during the ETH bullish test period, the Buy and Hold strategy secured a 22.59% return, while CryptoTrade exceeded this by a remarkable 3%."},{"finding":"CryptoTrade (GPT‑4o) improves total return and Sharpe ratio by about 3% in ETH’s bull market.","kind":"performance","quote":"CryptoTrade (GPT‑4o) excels in all metrics under ETH’s bull market by 3% in total return and sharpe ratio."},{"finding":"Even without fine‑tuning, CryptoTrade surpasses more than half of the trading signals across different market conditions.","kind":"performance","quote":"it consistently surpasses more than half of the trading signals across different market conditions, even without fine‑tuning."}],"reported_performance":{"strategy":"CryptoTrade (GPT-4o)","sharpe":0.23,"sharpe_benchmark":0.25,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"holdout","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":20,"citations":{"cited_by_in_corpus":4,"by_independent_authors":4,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":4,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T18:07:04Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}