bobaoxu2001/prediction-market-value-lab

This repository takes a research-first approach to prediction market arbitrage, distinguishing itself from the many 'copy this bot and get rich' repositories. Written in Python, it operates as a read-only research platform that scans both Kalshi and Polymarket for mispricings and arbitrage opportunities. The methodology is notably rigorous: decimal-exact pricing avoids floating-point imprecision that can affect small edges, independent probability models are used for each market (rather than assuming price equivalence), and a walk-forward backtesting framework validates strategies out-of-sample. Conservative expected value (EV) gating means only opportunities with substantial, statistically significant edges are flagged, reducing false positives from noise. Created on July 27, 2026 (today), this is the newest repo in the batch. With 0 stars and 1715KB, it's a fresh but potentially valuable addition to the Polymarket analytics ecosystem. The focus on research rather than execution makes it ideal for quantitative analysts, data scientists, and traders who want to validate strategies before deploying capital. The low malware risk is supported by its read-only design and transparent academic methodology. Worth cloning for anyone conducting systematic research into prediction market efficiency or building a data-driven approach to cross-platform arbitrage. The code quality signals suggest a disciplined developer with quantitative finance background.