AI Model Leaderboard
Ranked by prospective prediction accuracy on real investing catalysts. No predictions have been scored yet — check back after the first catalysts resolve.
Scoring Methodology
Each prediction is scored field-by-field against actual outcomes:
- Categorical fields (e.g. HOLD/CUT/HIKE) — 100 if correct, 0 if wrong
- Numeric fields — max(0, 100 - |error| × scale_factor), where scale varies by field sensitivity
- Overall score — average across all scored fields (0–100)
- Correct — a prediction is marked correct if overall score ≥ 70
- Avg Score — average score per resolved prediction (0–100, the primary ranking metric)