Gate Research Institute: Multi-Agent LLM Trading Framework Significantly Outperforms Buy & Hold Strategy in BTC Backtesting

Odaily Planet Daily News According to a recent report titled "Research and Backtesting Analysis of a Multi-Agent LLM-Based BTC Trading Framework" published by Gate Research Institute, compared to a single LLM directly generating trading signals, the Multi-Agent LLM architecture more closely resembles the research and investment process of real financial institutions. It can enhance the transparency and risk control capabilities of trading decisions through collaboration and debate among analysts, researchers, traders, and risk management teams. The study, based on the TradingAgents framework, constructed an AI trading system applicable to the crypto market for BTC and introduced multiple agent roles including technical analysis, news analysis, sentiment analysis, and macro/on-chain analysis.

The research conducted historical backtesting of the TradingAgents-BTC strategy using BTC/USDT 1-hour data. The results showed that the strategy achieved a +20.25% total return during the test period, significantly outperforming the Buy & Hold strategy's -7.89% over the same period. At the same time, the maximum drawdown was controlled at -17.41%, lower than the Buy & Hold's -27.06%. The study suggests that during consolidation and downtrend phases, the multi-agent framework can reduce risk exposure through Sell/Underweight and Flat states, and re-enter long positions during market rebounds, thereby improving overall risk-adjusted returns.

The report indicates that the Multi-Agent LLM framework demonstrates certain application potential in cryptocurrency trading scenarios. However, the current backtesting period only covers approximately three months, and 1-hour level trading may still be affected by transaction fees, slippage, and signal latency. Future research still needs to further validate the strategy's stability and generalization capabilities over longer historical periods, under different market conditions, and across more asset classes.