prediction-arbitrage-scout

Cross-market price discrepancies occur when YES/NO prices for aligned events differ across platforms. This Skill identifies those gaps by fetching public endpoint data, applying fuzzy matching to event names, and computing theoretical costs after fees to rank profitable opportunities.

You can compare prediction market prices by fetching data from public endpoints using the requests library and processing it with pandas. The workflow normalizes prices, matches events across platforms, and generates audit-ready Markdown and CSV outputs for review.

No, you do not need API keys to fetch market data for analysis. The pipeline retrieves YES/NO prices from publicly available endpoints on both platforms, requiring only Python 3.x with standard libraries like requests and pandas to execute the discovery workflow.

The best way to normalize and match event names is using fuzzy matching algorithms to align similar market titles across platforms. This ensures accurate comparison of YES/NO prices and reliable calculation of arbitrage opportunities before generating final reports.

Limitations include dependency on public endpoint availability and potential mismatches from fuzzy matching imperfect event names. Theoretical profit calculations rely on current fee structures, meaning real-time execution may face slippage or liquidity constraints not captured in the generated CSV report.