Research paper

Unlocking Liquidity on Prediction Markets

How to price the risk of lending against prediction market positions when prices can jump before collateral can be sold.

By William Flanders and Stephen Flanders
Download whitepaper (PDF)

Overview

The paper proposes a risk engine calibrated on historical Polymarket data. It estimates lender losses over a fixed loan period and uses a Wang distortion to convert that distribution into an upfront premium.

What the paper covers

  • Why sudden price jumps and thin order books make liquidation unreliable on its own.
  • How market conditions at origination affect the premium charged to the borrower.
  • How the model compares with flat-rate lending in a simulation on held-out data. Maximum drawdown was 3.6% for the model and 85% for the baseline.