AI-powered earnings call analysis for the top 1,000 US public companies — featuring structured equity research and a live management bullishness leaderboard.
Ranked by conservative bullishness score (μ − 3σ). Click any row for full details.
| # | Ticker | Date | Bullishness Score | μ Mean | σ Uncertainty | Tier |
|---|
EarningsIQ runs a competitive tournament — pitting anonymized excerpts against each other in thousands of head-to-head AI matchups to produce a calibrated bullishness ranking.
Company identifiers, executive names, and ticker symbols are removed before any comparison — the AI evaluates pure rhetorical quality, not brand recognition.
Bias EliminationEach comparison asks: which management team is more bullish? Tone, guidance strength, and confidence signals are evaluated holistically by a financial AI judge.
Pairwise EvaluationThe engine targets high-uncertainty matchups first — maximizing information per comparison and converging the ranking far faster than random approaches.
Adaptive SchedulingBorrowed from competitive gaming, TrueSkill maintains a full probability distribution over each company's true bullishness. Uncertainty σ shrinks with every comparison.
Probabilistic InferenceEach matchup runs in both orderings — only unanimous results count, eliminating the systematic LLM bias toward whichever excerpt appears first.
Statistical RobustnessFinal rankings use μ − 3σ — a lower-bound that rewards consistency. Narrow consistent winners rank above volatile big-win/big-loss performers.
Risk-Adjusted RankingStructured analyses — executive summary, forward guidance, tone assessments — are pulled from a live database covering hundreds of companies.
Each analysis is condensed to ~300 words and fully anonymized. The judge sees only the quality of management communication.
Thousands of concurrent pairwise comparisons run adaptively — targeting the most uncertain pairs until confidence intervals converge.
Once max σ drops below 2.0, the tournament ends. Final scores are published as a ranked leaderboard.