Presage – Real-time insider detection for prediction markets

Name: Presage – Real-time insider detection for prediction markets
Availability: InStock
Author: Jonsh92

by Jonsh92·Mar 5, 2026·3 points·0 comments

Visit Project View on HN

AI Analysis

●●SolidBig BrainDark Horse

Iran strike case study is compelling, but backtesting on three events is thin.

Strengths

•Iran strike case study shows 94/100 signal detection 24h before geopolitical event
•Five-signal stack (wallet age, concentration, low-probability bets, whale volume, clustering) creates genuine detection logic
•Real on-chain indexing against Polygon/Polymarket, not speculative

Weaknesses

•Only three backtested events; unclear if this generalizes to market microstructure noise
•Waitlist model means no live product access; claims unverifiable without trading data

Post Description

Presage scores every Polymarket trade against five on-chain signals: wallet age, position concentration, bet probability, trade volume, and wallet clustering. Backtested against three confirmed insider events with clean separation between suspicious and legitimate activity.