Guidance & estimates

Definition

The expectations layer: company GUIDANCE (management’s public forecasts, given under Regulation FD to everyone at once) and ANALYST ESTIMATES (the consensus that surprises are measured against). Prices trade the gap between results and this layer — not results in isolation — which makes the layer’s own mechanics (who sets it, how it’s managed, where it clusters) tradeable structure. The documented core: managements systematically WALK ESTIMATES DOWN to beatable levels.

How it works / structure

• The guidance game: Cotter-Tuna-Wysocki (2006) and a broad literature document expectations management — guidance nudges consensus down during the quarter so the print “beats”; the beat rate for large caps persistently runs near 70-80% — a beat is the EXPECTED outcome, and the tradeable object is the beat’s SIZE vs the walk-down’s depth.
• Consensus mechanics (engine-executable): estimate count, dispersion (wide dispersion = genuine uncertainty = bigger event variance), revision trajectory into the print (sent-analyst-revisions — the momentum of the layer), and the “whisper” gap (unofficial expectations above official consensus on crowded names).
• Guidance forms: point vs range vs withdrawn (withdrawal is itself an event — documented distress/uncertainty signal); annual vs quarterly cadence; the mid-quarter update (lens-event-catalyst unscheduled events).
• The reaction decode: beat-and-drop / miss-and-rally prints reveal the true expectation level — positioning information the consensus number hid (qualitative-analysis reaction reading).

When it applies

Every earnings-event thesis (the benchmark IS this layer — event-earnings); revision-momentum strategies (sent-analyst-revisions); estimate-dispersion as an event- vol input; guidance-credibility grading per management team (walk-down depth and delivery history are measurable).

Risk profile & failure modes

• Trading the headline beat: an engineered +1% beat against walked-down numbers is noise; the market grades vs the whisper and the guide — naive surprise metrics mislead.
• Consensus staleness: thin-coverage names carry stale estimates; the “surprise” is against a number nobody believes.
• Guidance-credibility drift: serial guiders-down get discounted; the same guidance from different managements is different information.
• Reg FD boundaries: private expectation-setting is illegal; channel checks and expert networks operate near the line — the platform uses public information, full stop.

Evidence & limits

Reg FD is SEC law; the expectations-management literature (Cotter et al and successors) documents the walk-down; beat- rate persistence is published statistics (quoted as approximate and period-dependent). Revision-momentum return evidence lives in sent-analyst-revisions. Whisper numbers are unaudited folklore-adjacent data — labeled where used.

Falsifiable-thesis examples

Illustrations only, not signals:

• “X will beat consensus EPS but close down (whisper-gap thesis: real expectations exceed the walked-down number)” — falsified by the print-reaction pair.
• “Names with upward revision trajectories into the print will outperform post-print vs downward-trajectory names this season” — falsified by the cohort comparison.

Cross-references

• The event it frames: event-earnings
• The layer’s momentum: sent-analyst-revisions
• Reaction decoding: qualitative-analysis
• The filed baseline: fa-financial-statements; valuation link: fa-multiples-comparables (forward bases)