Retail Traders’ AI Bots Deliver 7% Simulated Gain with 22% Drawdowns
An AI agent trained over 2.5 weeks on a $100,000 simulated brokerage account posted a 7% gain over 30 days, outpacing the S&P 500’s 4.5% return but enduring drawdowns as large as 22%. Robinhood and Public Holdings plan AI agents for users, while malware threats and default model conservatism limit gains.
1. Rise of AI Agents in Retail Investing
Open-source platforms like OpenClaw enable retail traders to link AI models to messaging apps such as WhatsApp and Telegram, automating buy and sell instructions across equities, crypto and prediction markets without technical expertise.
2. Performance of Nesler’s Trading Bot
Jake Nesler spent two and a half weeks teaching an AI model his risk and entry strategies, then deployed it on a $100,000 simulated Alpaca account; the bot returned 7% in 30 days versus the S&P 500’s 4.5% gain but suffered drawdowns up to 22%.
3. Platform Responses and Implications for Robinhood
Trading apps born in the Robinhood era are now exploring AI-driven agents to boost engagement and trading volume, with Public Holdings and others rolling out agent interfaces as a new feature to retain tech-savvy users.
4. Security and Model Limitations
Reports link some AI agent offers to malware distribution, and default large language models favor conservative, blue-chip bets—forcing traders to override risk aversion—while mass sharing of successful bots erodes any unique edge.