Library About the Grade
The Surprise That Actually Matters
Wall Street’s oldest anomaly has been declared dead. It isn’t — but the way most investors measure it is.
The Earnings Whisper Grade measures the likelihood that a stock will outperform or underperform from the first market open after an earnings release through its next report — a window of roughly 91 days.
It does not measure whether a company beat or missed the consensus estimate.
It measures whether the report changed the market’s true expectations for the company — and whether investors have fully adjusted to that new information yet. That distinction is the entire reason the Grade works.
The original earnings anomaly
In 1968, Ray Ball and Philip Brown published An Empirical Evaluation of Accounting Income Numbers, one of the foundational studies of empirical finance. Their breakthrough wasn’t showing that earnings mattered — it was defining an information event around unexpected earnings and tracking prices before, during, and after the announcement.
Stocks reporting good news outperformed stocks reporting bad news. More importantly, part of that outperformance arrived after the announcement. Prices did not adjust immediately and completely.
That delayed reaction became the Post-Earnings Announcement Drift — and one of the most durable challenges to the Efficient Market Hypothesis ever documented. If public information were priced instantly, an earnings report couldn’t predict returns weeks later. For decades, it did. Ball and Brown’s own replication, covering the United States from 1971 through 2017 and sixteen other countries, kept finding it.
More on the Post-Earnings Announcement Drift
Then the drift “disappeared”
Human behavior changes slowly. The machinery of the market does not.
Estimates are now broadly available. Results are distributed instantly. Algorithms read press releases and trade earnings, revenue, guidance, and margins within seconds. Analysts revise faster than they did a generation ago. Research has found that faster revisions and low-latency trading pull more of the reaction into the announcement window itself — and one modern study concludes the traditional PEAD signal has largely vanished for large-cap stocks since 2006.
Our own statistics point the same direction — with one critical qualification.
PEAD disappeared when the surprise is measured against the consensus estimate. It did not disappear when results are measured against the market’s real expectations.
The traditional model assumes the published consensus is what investors actually expect. Increasingly, it isn’t.
Beating consensus is not a surprise
A company can beat the consensus estimate and watch its stock fall. It can miss and trade higher. That isn’t irrational — it means the published number was never the bar investors had actually set.
Consensus estimates can be stale, anchored to company guidance, or slow to reflect what professional investors already know about demand, pricing, margins, and customer activity. An earnings surprise cannot be measured without first measuring what the market truly expected.
The earliest systematic evidence came in 1999, when Mark Bagnoli, Messod Beneish, and Susan Watts found that whisper forecasts were, on average, more accurate than published estimates — and better proxies for how prices actually reacted. Trading on whisper forecast errors had a stronger relationship with announcement returns than trading on consensus errors.
That study had real limitations: 127 companies from 1995 through 1997, over 90% of them technology firms, with many forecasts scraped from message boards and news reports rather than researched from professionals. But its conclusion has only grown more important since:
The published consensus and the expectation embedded in a stock’s price are not the same thing.
The Earnings Whisper® number was created to measure that difference properly — by polling professional analysts and examining their research, not collecting rumors.
More on the Earnings Whisper® number
Expectations are more than a number
The market’s expectation often can’t be reduced to one EPS figure at all. Investors may already be positioned for a beat-and-raise, accelerating customer growth, better margins, stronger bookings, or a milestone on a product the whole thesis depends on.
A company can clear the consensus number and still fall short of that. The headline says “beat”; the report is a disappointment. The reverse happens too — a technical miss that shows the business reaching an inflection point, forcing investors to raise their view of every future quarter.
The Grade evaluates the entire expectations reset, not the gap between one reported number and one published one.
Sentiment decides how far it travels
An earnings report has no fixed meaning on its own. The same result produces very different reactions depending on what investors believed walking in.
When sentiment is euphoric, even a strong report can fail — strength was expected, positioned for, and already in the price. When sentiment is bearish, a genuine positive surprise forces a real reassessment: estimates rise, positions get rebuilt, and the stock keeps moving long after the first reaction.
The research backs this up. Analysts are demonstrably slower to upgrade stocks they previously disliked, even after the evidence turns. Friday announcements — when attention is lowest — produce smaller immediate reactions and substantially larger delayed ones. And the 60-day return spread between positive and negative surprises widens sharply on days crowded with other earnings reports.
Investors anchor to prior beliefs, notice some information and overlook the rest, and adjust gradually. That behavior is the foundation the Grade is built on — and the reason it keeps working.
Turning the reset into a grade
The Grade combines two components:
The actual surprise — how the company performed against the expectations investors had genuinely priced in, not merely the published consensus.
The change in perception — whether the results confirmed the existing narrative or forced investors to reconsider the company’s future earnings power.
The result runs from A+ to F and measures the stock’s potential to drift higher or lower into its next report. Any single quarter contains noise — a weak company can beat a strong one over a stretch, and even an A+ stock can decline. Across thousands of announcements, though, the grades have produced a strikingly orderly hierarchy.
Quarterly returns by grade
The relationship holds across the whole scale: F and D grades underperform (D less badly than F), C grades track the market, and B grades and above outperform.
Stocks earning an A+ produced an average annualized return of 37.1% — more than 3 times the S&P 500’s annualized return over the same period.
The important result isn’t the A+ number. It’s the orderly progression across the entire scale: as the quality of the expectations reset improves, subsequent performance improves with it.
The A+ grade in real time
When the Grade launched in 2015, it was a back-tested idea — grounded in decades of research, but unproven live. Backtests are cheap. A decade of real-time results is not.
More important than any single statistic: the live record keeps validating the premise. The strongest post-earnings returns don’t come from companies that merely beat consensus. They come from companies whose results are meaningfully better than what the market actually expected — the reports that force investors to raise their assessment of future earnings.
Measuring the surprise that matters
The classic story was simple: beat the estimate, drift higher; miss it, drift lower. That story no longer works. Instant dissemination and algorithmic trading have made a routine consensus beat easy for the market to price on the spot.
But consensus was never the expectation. Investors still anchor. Analysts still adjust gradually. Sentiment still decides which information gets weighed and which gets ignored. And the full set of expectations around a report is still far richer than one published EPS number.
The drift didn’t disappear. The yardstick did.
The Earnings Whisper Grade measures the surprise that matters — the difference between what investors truly expected and what the company delivered. That difference is what changes expectations. And changing expectations is what drives the drift.
Grade returns average the move from the first market open after each earnings report through the first open after the company’s next report, across 121,683 graded reports from December 17, 2002 through June 4, 2026. The universe excludes stocks priced under $5 or averaging fewer than 150,000 shares of daily volume, and requires the next report to fall 45–135 days after the last. The S&P 500 comparison uses its open-to-open return over the same windows. Annualized figures compound the average quarterly return four times. Figures exclude transaction costs, dividends, and slippage, and are not a guarantee of future results.