Historical Surprise Persistence
Track companies that consistently beat expectations.
Overview
This pillar identifies companies with a persistent history of positive earnings surprises. It's built on the principle that past overperformance and conservative guidance can be strong predictors of future results.
What It Does
It analyzes the serial correlation of a company's earnings per share (EPS) surprises over a rolling 4 to 8 quarter window. The pillar calculates the statistical significance of this trend, flagging companies that repeatedly and predictably outperform analyst consensus estimates. It also assesses management guidance to see if they consistently set low bars they can easily clear.
Why It Matters
The market often underreacts to a single earnings beat but is slow to price in a persistent pattern of them. This creates a recurring predictive opportunity around earnings season, offering an edge in forecasting both the earnings outcome and the subsequent stock price reaction.
How It Works
First, the pillar gathers historical analyst consensus EPS estimates and actual reported EPS for a specific company. It then calculates the percentage surprise for each of the last 8 quarters. Using this time series data, it computes an autocorrelation coefficient to measure the strength of the beat-to-beat trend. Finally, it generates a persistence score based on this correlation and the consistency of management's guidance.
Methodology
The primary metric is the first-order autocorrelation of the earnings surprise series, defined as (Actual EPS - Consensus EPS) / |Consensus EPS|, over a rolling 8-quarter window. A statistically significant positive coefficient (p < 0.10) indicates persistence. A secondary metric, the Guidance Conservatism Ratio, is calculated as (Management's EPS Guidance Midpoint / Analyst Consensus EPS); a ratio consistently below 1 suggests a pattern of conservative guidance.
Edge & Advantage
This pillar provides an edge by systematically identifying companies whose performance is predictably underestimated by the market, moving beyond single-event analysis to find durable trends.
Key Indicators
-
Serial Beat Correlation
highMeasures the statistical likelihood that a positive earnings surprise in one quarter is followed by another in the next.
-
Management Guidance Conservatism Ratio
mediumCompares management's forward guidance to analyst consensus; a low ratio suggests easily achievable targets.
-
Surprise Magnitude Trend
lowAnalyzes whether the size of the earnings beat is increasing or decreasing over time.
Data Sources
-
Provides institutional-grade historical earnings estimates and actuals.
-
A primary source for financial data, including historical consensus estimates and company guidance.
-
Official source for corporate 10-Q and 10-K filings containing actual reported earnings.
Example Questions This Pillar Answers
- → Will Apple (AAPL) beat its consensus EPS estimate for the next reported quarter?
- → Will NVIDIA's stock price increase by more than 5% in the 24 hours following its next earnings report?
- → Will the aggregate earnings surprise for the S&P 500 Information Technology sector be positive next quarter?
Tags
Use Historical Surprise Persistence on a real market
Run this analytical framework on any Polymarket or Kalshi event contract.
Try PillarLab