← Research

Research tool

Prediction Index: trace geopolitical aggregates to their questions

Prediction Index combines an ecosystem directory and the Chaos Index. The research aggregate separates escalation and peace scenarios, Polymarket and Kalshi surfaces, and market pricing from a language-model ensemble.

Open Prediction Indexpredictionindex.xyz
Availability: public

Screenshots and interface

What is Prediction Index?

Constituent tables expose markets, clusters, forecast groups, weights and Yes odds, including entries with no weight. A war starting and a ceasefire failing are different questions; neither basket is an authoritative probability of every geopolitical outcome or a single tradable contract.

The operator says Grok, GPT and Claude use independent prompts hiding market prices. Machine history begins with the first experiment, without backfilling. The captured page showed a September 29 machine run and August 25 constituent revision; LIVE labels do not establish current inputs.

Who it is for

  • Researchers following geopolitical contract baskets.
  • Readers comparing market aggregates and model forecasts.
  • Builders mapping the prediction-market ecosystem.

Features and coverage

  • Separate baskets and venues

    Inspect escalation and peace separately on Polymarket and Kalshi views.

  • Visible constituents

    Read markets, clusters, forecast groups, weights and Yes odds, including zero-weight entries.

  • Market/model comparison

    Compare individual Grok, GPT and Claude assessments and their spread with the market, preserving different timestamps.

  • Ecosystem directory

    Filter projects by outcome type, chain and status. Classifications are research leads rather than independent confirmation of capabilities.

Workflows with Prediction Index

Explain an index move

Select venue and basket, inspect the largest weights and open underlying contracts to compare resolution conditions.

Investigate disagreement

Check model date and rules, then examine the largest gaps and model spread without treating older model values as current prices.

Pricing and total cost

The homepage, directory and Chaos Index are publicly readable without an account or displayed paid plan. A complete commercial API or subscription offer is not established.

Access: The research-subscription call to action is separate from reading the index.

Getting started

Before you start: Weighted event baskets do not cover every possible geopolitical outcome.

  1. Open Chaos Index and select the relevant venue.
  2. Read constituents, weights, revision date and machine-run date.
  3. Use gaps to select questions for investigation, then verify original rules and current prices.

Displayed model and constituent dates predate the source review; a LIVE label does not establish fresh data.

Useful links

What to check

  • Weighted event baskets do not cover every possible geopolitical outcome.
  • Escalation and peace are different event groups, not exact opposites.
  • Freshness and reconstruction of the aggregate are not independently established.

A closer look

Ratings & reviews

Editorial rating
3.0/ 5
0 user reviews

Our take on Prediction Index

Constituents and distinct market/model views make the research traceable. Older run dates, basket selection and unverified aggregation require care with headline numbers, which are neither comprehensive global-risk probabilities nor executable quotes.

Based on documented features, setup and access terms. How we rate

User reviews 0

The main score switches to user ratings after 5 published reviews.

Be the first to share your experience

A specific example helps the next person choose the right tool.

Related tools

Browse category

Frequently asked questions about Prediction Index

Can I buy Chaos Index as one contract?

No order-entry workflow is established. It is an aggregate of constituent questions.

Does LIVE mean the model ran today?

No. The capture showed a September 29 run and August 25 constituent revision; read explicit dates.

Is machine history backfilled?

The operator says it starts at the first experiment without backfilled historical forecasts.

Are war and peace complementary probabilities?

No. They group different questions, such as war starting and ceasefire failure.