Zrooth
Menu

Judged in Public

Our AI Forecasts and Scorecard

Every forecast is a specific, dated statement with the probability we honestly give it. When the date comes, we check it against public information and record the result here, right or wrong.

Open Forecasts

0

Judged

0

Calls Right

Soon

First results as dates arrive

Brier Score

Soon

0 is perfect; 0.25 is a coin toss

Open Forecasts

Our first forecasts are published with the first Situation Reports, every Monday.

How We Score Ourselves

  • A call is right when something we gave 50% or more happened, or something we gave under 50% did not.
  • The Brier score also rewards honest probabilities: saying 90% and being wrong costs far more than saying 60%. Lower is better.
  • Void means the question could not be judged fairly (for example, the rule it depended on was withdrawn). Void forecasts are not scored.
  • Forecasts are our judgement, not advice, and never a promise of any outcome. We do not forecast markets, share prices, elections or individuals.

Reports covering: India, The Gulf, Africa, Jobs and Skills, AI in Business, Policy and Regulation.

Free Every Friday

This Week in AI, Explained

The week's most important AI news in one short email, with what it means for your work and your business.

One email a week. Unsubscribe in one click. Privacy Policy.