Research

Long-form analysis of how Kalshi event-contract markets actually move, and how that compares to traditional sportsbook pricing. Drawn from the SportsBookISH dataset — every quote on every market across 14 regulated US books and Kalshi/Polymarket since May 2026.

Market microstructure·8 min·2026-06-01

Why mid-game Kalshi lines lag sportsbook consensus

Settlement risk premium, leverage exit, and the structural reasons exchange prices drift behind books once a game is in progress.

Market microstructure·10 min·2026-06-01

How sportsbooks reprice without news — a taxonomy of line moves

Sharp action, book exposure rebalancing, public sentiment lean, and originator-vs-follower lag. Every line move is one of six things.

Market microstructure·7 min·2026-06-01

Volume concentration in event-contract ladders

Why the 80+ wins rung trades 1,000× more than the 75+ rung — and why pricing the adjacent thresholds off the same volume is statistically noise.

Open data

Every analysis here is derived from the public SportsBookISH dataset on Hugging Face. CC-BY 4.0 — cite us, reuse freely.

Hugging Face dataset →Data exports →Public REST API →

Methodology

Devigging is per-book (multiplicative). Kalshi prices use the bid/ask midpoint when both sides have non-zero liquidity, falling back to last-trade when the spread is wider than 4¢. Quote freshness window is 30 minutes — quotes older than that are excluded from medians. Detailed methodology, including the exact lens taxonomy used in the tweet composer, lives in the operations notes for each piece.