Prediction markets are markets in which participants trade contracts based on the outcomes of future events, and they have become increasingly popular in recent years. Given their rising popularity, an important question is how to develop quantitative betting strategies.
Reference [1] proposes an interesting approach based on an adaptation of the classic pairs-trading strategy. Specifically, the paper adapts financial-market pairs trading and mean reversion to U.S. presidential betting markets. But instead of modeling the price spread between two stocks, it models the sum of the implied probabilities of the presidential nominees.
The combined probability is modelled as a latent Ornstein-Uhlenbeck (OU) process with a time-varying mean plus additive observation noise. The OU component represents short-term mean reversion, while the moving equilibrium represents gradual changes in the election market.
The authors utilize political betting odds obtained through BetData. The 2020 election (Trump/Biden) is used for model construction and training, while the 2024 election (Trump/Harris) is used for out-of-sample testing. They pointed out,
This study demonstrates the feasibility of adapting a pairs-trading framework to political prediction markets. We model the combined implied probability of the two major-party nominees using a latent time-varying-mean Ornstein-Uhlenbeck process with additive observation noise, use bootstrap-based prediction bounds to identify trading signals, and apply a Bradley-Terry pairwise comparison model to select the candidate-specific odds. In the 2024 out-of-sample evaluation, the strategy generated 130 completed trades, with a mean realized odds-price return of 1.86% and an unannualized per-trade Sharpe-type ratio of 1.12. These results illustrate the potential of the integrated signal-generation and candidate-selection framework in electoral markets dominated by two leading candidates.
In short, the authors present the framework as a proof of concept showing that a pairs-trading-style mean-reversion model can be adapted to two-candidate political betting markets. They point to the 2024 130-trade out-of-sample result, +1.86% mean synthetic odds-price return and 1.12 per-trade Sharpe-type ratio as evidence for the integrated signal-generation/candidate-selection framework.
This is an interesting paper, as it provides a foundation for developing quantitative strategies in prediction markets.
Let us know what you think in the comments below or in the discussion forum.
References
[1] Haoyu Liu, Len Thomas, Benjamin Baer, and Carl Donovan, Adapting Pairs Trading to Gambling Markets: A Case Study of the U.S. Presidential Election, arXiv 2609.22639v1, 2026
Post Source Here: Statistical Arbitrage in Prediction Markets
source https://harbourfronts.com/statistical-arbitrage-prediction-markets/
No comments:
Post a Comment