Forecast Credibility

Leadership Under Uncertainty uses cryptographic timestamping to prove our AI forecasts existed before their outcomes were known. Every forecast is hashed, sealed in daily Merkle trees, and anchored to public repositories.

Three-Layer Integrity System

🔐

Layer 1: Content Hashing

Each forecast is hashed using SHA-256. The hash includes the question, probability, model configuration, and evidence sources. Any modification would change the hash.

🌳

Layer 2: Daily Merkle Trees

All daily forecasts are combined into a Merkle tree. The root hash commits to every forecast made that day. Each day's tree chains to the previous day's root.

⚓

Layer 3: Public Anchoring

Daily root hashes are committed to a public GitHub repository. The commit timestamp provides independent proof that the forecasts existed at that time.

Verify a Forecast

Enter a forecast ID and date to verify its integrity and retrieve its Merkle proof.

Daily Merkle Seals

Each seal cryptographically commits to the records included as leaves in that day's Merkle tree — proving they existed at or before the seal time. It does not claim to cover every record created that day: anything written after the seal ran is covered by a later root.

Loading seals...

Technical Details

Hash Algorithm: SHA-256

Tree Type: Merkle Binary Tree with sorted-concatenation pairing

JSON Serialization: Canonical (alphabetically sorted keys)

Anchor Repository: github.com/cberzins1973/luu-attestation-anchors

API Endpoint: GET /api/public/forecast-integrity-proof?forecastId=&dateKey=

Leadership Under Uncertainty Forward Testing Integrity System - Cryptographic timestamping for AI forecasts