Guides / Quant Modeling
Most prediction market models — regardless of whether the final step is a tree model, a neural net, or a simple rule — draw from the same small set of raw ingredients: bid/ask depth and its imbalance, spread, mid-price drift, recent trade size and direction, and time remaining to resolution. ProbSights' historical API exposes the underlying snapshots (orderbook depth and executed trades, Kalshi and Polymarket, BTC/ETH/SOL Up/Down at 5m–24h resolution) that these features are computed from — the feature engineering itself happens in your own pipeline.
Data from Kalshi and Polymarket. Search live API docs Pricing