Descriptions:
Part Time Larry documents both the motivation and technical implementation of a real-time order book watcher for Kalshi, the US-regulated prediction market exchange. The backstory: while on a flight to Mexico City during the Super Bowl, the presenter monitored Kalshi prop bet markets without watching the game and identified striking mispricings — including a no-bet on a running back rushing touchdowns available at 51 cents against an ask of 94 cents — purely by reading order book dynamics.
The tutorial covers the full Python implementation: connecting to the Kalshi websocket API using a KalshiO authentication class, subscribing to the orderbook_delta channel for specific market tickers, and incrementally applying delta updates to reconstruct a live order book. A terminal UI (TUI) version is also demonstrated, featuring keyboard shortcuts for rapid trading actions — buy at bid (B), sell the ask (S), jump order, and cancel — enabling faster execution than the Kalshi web interface allows.
The presenter also references a previously built edge analyzer that uses NFL historical data and Kalshi’s historical dataset to calculate baseline probabilities for prop bet markets, providing a framework for identifying underpriced contracts systematically. Source code and an accompanying blog post are linked in the video description. This is a practical coding tutorial for developers interested in building algorithmic or semi-automated tooling around prediction market data.
📺 Source: Part Time Larry · Published February 21, 2026
🏷️ Format: Hands On Build
