Questions for the tape

Any name, any metric, any quarter — pinned to the company’s own reported numbers.

Binary

“Will NVIDIA beat consensus revenue next quarter?”

A probability of the beat, built from filings, guidance history, and supply-chain signals.

Numeric

“What EPS does Apple report for the December quarter?”

A full distribution — median, P10–P90 band, and the odds of clearing the whisper number.

Binary

“Will management raise full-year guidance on the call?”

Guidance questions read the company’s own language history, not just the model.

Binary

“Will the stock close up the day after it reports?”

The reaction, not just the result — positioning and expectations priced in.

Multiple choice

“Which mega-cap posts the largest revenue surprise this season?”

Cross-name questions get a probability for every candidate, surprises included.

Date

“When does the company’s AI segment first pass $10B a quarter?”

Milestone questions resolve on reported segments — median timing plus checkpoint odds.

How an earnings call gets a number

Which metric, which quarter, whose consensus — resolution criteria are fixed before anyone forecasts, so the answer is checkable.

10-Qs, transcripts, guidance history, peer prints, and the news of the last hour — live sources, cited in the report.

A red team attacks each analyst’s case. What can’t defend itself doesn’t make it into the number.

Panel views pool through deterministic, validated aggregation. The headline probability recomputes from its own tables — or it’s rejected.