TradeSmith, led by CEO Keith Kaplan, has developed an AI-powered trading tool capable of identifying over 200 predictive market signals across nearly 2,500 stocks. During a recent event, Kaplan noted that these signals have shown historical accuracy rates of up to 90% and, in a six-year backtest, produced a model portfolio with a 12x return, including a gain of 16.6% in 2022, a year when the S&P 500 fell nearly 20%.
The system, which employs a Quality Score from 0 to 100 to evaluate each signal based on its win rate and average returns, operates with a focused model portfolio of just three S&P 500 stocks at any given time. This strategy delivered a compounded annual return of 54% compared to about 15% for the S&P 500 from 2020 to 2026, with a maximum drawdown of 18.1% versus the S&P 500’s 25.4%.
Mike Carr, TradeSmith’s chief developer, implemented 847 calculations across one trillion database entries to refine the signal generation process. This significant investment in data analysis reflects a growing trend where AI is harnessed to analyze vast quantities of market data, providing insights that were previously obscured.
