币界网报道：Token Metrics, a cryptocurrency research and analytics platform, has released findings indicating that its AI-driven trading signals significantly outperform traditional technical indicators. According to their internal analysis, the AI models achieved returns 35% higher than conventional methods like moving averages and RSI over a six-month backtest period. The platform's machine learning algorithms analyze over 100 on-chain and market data points, including exchange flows, social sentiment, and whale activity patterns. CEO Ian Balina emphasized that their system continuously adapts to changing market conditions, unlike static indicators. The company reports particular success with Bitcoin and Ethereum predictions, where AI signals correctly identified 78% of major price movements in 2023. Token Metrics currently offers these signals through subscription plans ranging from $50 to $500 monthly, with institutional clients accounting for 40% of their user base. The findings come as Wall Street firms increasingly explore AI applications for crypto trading, though some analysts caution about potential overfitting risks in machine learning models.