Hexcast

Problem

Probability and prediction markets (and sports odds) drown operators in competing forecasts. Raw consensus blends noise with signal; fixed weights go stale as markets and seasons move. Teams need a way to rank sources, aggregate probabilities, and improve that ranking over time — without pretending a single model already “solved” the market.

What this product is

Hexcast is Ikosagon’s R&D line for ranked probability aggregation: take many probability estimates (models, books, markets, experts), score how they have ranked against outcomes, then combine them into a clearer aggregate view. The stack sits on Ikosagon’s proprietary ranking system and is aimed at prediction/probability markets and sports.

In plain language: Hexcast ranks who has been useful lately, weighs forecasts accordingly, and keeps re-learning that ranking as new results land — a self-improving aggregation loop rather than a static average.

Principles

• Rank first, then aggregate: source quality is measured relative to outcomes, not assumed equal
• Proprietary ranking stays proprietary: public copy describes the idea, not the secret sauce
• Markets + sports focus: same aggregation ideas, different event calendars and data shapes
• Honest R&D: no invented win rates, ROI, backtests, clients, or prices on this site
• Explainable enough to discuss: operators should understand what is being ranked and aggregated, even when the ranking method itself is not open-sourced

Status

In progress. Hexcast is proprietary ranking + aggregation research and product shaping. This portfolio card is the public home for the name and thesis while the work continues. Rename and roadmap details can change — contact Shawn via /contact or shawn@ikosagon.com for serious inquiry, not for performance claims.