Exploring Arbitrage in Prediction Markets

The Quant team spent the Fall 2024 semester researching and developing a cross-arbitrage pairs trading strategy between prediction markets Kalshi and Polymarket. The focus was on event contracts related to Federal Reserve interest rate cuts, specifically the market for predicting four rate cuts in 2024. To execute this strategy, the team aggregated daily price data from both exchanges using APIs and formatted it into a CSV file for analysis. The team conducted a Pairwise Engle-Granger test to confirm mean reversion tendencies between the contracts. Based on the results, they designed an arbitrage strategy that analyzed the 10-day rolling average spread between the contracts and established Bollinger Bands at one standard deviation above and below the rolling average. Trades were triggered when the spread crossed these bands, leveraging mean-reversion tendencies for profit. The backtesting process recorded opening and closing trades, calculated profit and loss, and assessed the strategy’s performance through metrics like the Sharpe Ratio. In the upcoming semester, the team plans to extend their research to backtest strategies involving equities, derivatives, ETFs, and cryptocurrencies.