Abstract
The goal of this project was to understand the Polymarket microstructure and the effect of different types of wallets on the price. This research utilizes public 2024 US Election data from Polymarket to analyze market microstructure and the efficacy of algorithmic trading strategies within decentralized prediction markets. The study first develops a behavioral taxonomy by clustering wallet addresses into interpretable participant types such as informed, retail, or arbitrageur, and then quantifies information asymmetry by estimating type-specific Kyle’s λ to measure price impact. These empirical findings inform the design and back testing of an Avellaneda–Stoikov-style market-making framework specifically adapted to manage the unique inventory risks associated with political event contracts. Finally, the project evaluates the performance of mechanical edge strategies, assessing whether systematic mean-reversion or programmatic wallet-copying can generate consistent alpha in high-volatility, binary outcome environments.
Categories: Quant
0 Comments