Strategies operating within a single trading day, holding positions for minutes to hours.
ChatGPT can generate profitable day trading signals by analyzing real-time Twitter news and selecting stock tickers to buy and sell. The strategy earns significant intraday alpha, especially from short positions, even without firm-specific prompts.
Can ChatGPT predict stocks? Yes—GPT-4 applied to headlines yields a high-Sharpe intraday strategy with 650%+ returns. Predictive power emerges only in large models.
This paper finds that stocks underreact to firm-specific news—prices continue to drift in the news direction for days. A strategy based on high-frequency news returns earns over 3% per month and remains profitable after trading costs.
This paper uses a unique intraday dataset to track insider trades and shows that insiders don’t try to hide their activity. They prioritize trading on returns, not timing liquidity—making their actions visible and their impact immediate.
Most stock market gains happen during a four-hour window before European markets open. This paper finds that almost all returns come from this short time period, likely because investors are reacting to overnight news. A simple strategy that trades during this window beats buy-and-hold.