prediction-markets-lab

Similar to the 'polymarket-coherence' repo, this is a research-oriented project. However, it expands the scope to include Kalshi alongside Polymarket, aiming to create a standardized way to measure coherence—the internal consistency of prices across different markets. The key innovation here is the 'self-recording snapshot archive', which implies it continuously collects data on its own to build a historical database. This is highly valuable for backtesting and academic research. The 'fee-honest screening' means it attempts to filter out apparent arbitrage opportunities by accounting for the actual trading fees on each platform, which is a crucial detail often overlooked. The large size (30MB) is likely due to the historical data snapshots included in the repo. The code is clean and organized for analysis, and it is a heavy-duty data engineering project. For a quantitative researcher, this is a goldmine. It provides the tools to examine whether mispricings persist over time and across platforms. While it doesn't execute trades, it provides the 'intel' needed to decide if a bot is worth building. The development is active (created 2026-08-23, updated today), showing it is a living project. Malware risk is low; it's purely for data collection and analysis. This is a 'monitor' and potentially a 'worth_cloning' for its data archives alone. It would be excellent for creating custom research reports or for feeding a trading model with high-quality, historical data.