Overnight Reversal and the Asymmetric Reaction to News
Log in to collectAcademic paper
Overnight Reversal and the Asymmetric Reaction to News
Thomas Dangl; Stefan Salbrechter
- ATTU Wien
- ?Vienna University of Technology
Strategy in a nutshell
This intraday strategy uses a BERT-based financial news model trained on Thomson Reuters data to classify overnight news as positive, neutral, or negative for S&P 500 stocks. Stocks are traded at market open and closed at market close, going long or short based on sentiment and prior-day idiosyncratic returns, capturing inefficiencies caused by overnight news.
Economic rationale
Overnight news has a directional impact on the next day’s opening price, but intraday returns remain partially predictable. By combining previous-day returns with company-specific overnight sentiment, the strategy exploits predictable return reversals during the trading day.
Backtest performance
Annualised return47.52%
Volatility36.27%
Sharpe ratio1.31
Maximum drawdown-59.7%