Library The Power Rating

The Power Rating

The gap is the market’s first verdict on an earnings report. The Power Rating is about what happens over the five days after it.

An earnings report does not produce one reaction. It produces three, and they are worth keeping separate.

1 The gap

The immediate repricing when the stock reopens. Driven as much by positioning, liquidity, and options activity as by the report itself.

2 Initial drift

The first five trading days, measured from the post-announcement open. Analysts revise, institutions reposition, short sellers act. This is what the Power Rating measures.

3 Longer drift

The rest of the quarter, driven by earnings quality, guidance, and revisions. Measured by the Earnings Whisper Grade.

The Power Rating does not predict earnings, and it does not predict the gap. It is calculated after both are known, and it asks one question: given how the results landed against expectations, does the next week favor strength or weakness?

Why the first reaction is incomplete

An earnings release contains far more than an EPS number. Revenue, margins, guidance, cash flow, segment trends, customer behavior, expenses, and the outlook all have to be weighed — and then compared against what was already in the price.

The market reacts to the headline in milliseconds. Working out what the headline means takes longer. An algorithm can confirm that EPS cleared consensus instantly; determining whether the revenue was durable, whether guidance was conservative, or whether margins hold does not happen at that speed.

That lag is the drift. Ball and Brown documented it in 1968 and it has survived six decades of scrutiny. More on Post-Earnings Announcement Drift.

Attention appears to widen the gap between reaction and understanding. Friday announcements — when fewest people are watching — produce a roughly 15% smaller immediate response and a 70% larger delayed one. Reports landing on crowded days show the same pattern: weaker initial reaction, stronger subsequent drift.

Measured against the right expectation

Drift can only be measured if the surprise is measured against the number investors were actually trading against — which is often not the published consensus.

Estimates may be weeks old. Analysts frequently expect a company to clear their own published number without revising it. Institutions fold in channel checks, guidance, peer results, and economic data that never reach an estimate database.

Which is why a company can beat consensus and fall. It cleared the published number and missed the real one. The Earnings Whisper number exists to measure that second bar.

But the surprise alone does not explain the move. How the market was positioned for the news matters just as much.

Strongest setup Positive surprise, bearish sentiment

Nobody was positioned for it. Repositioning, short covering, and estimate revisions all push the same way.

Capped Positive surprise, bullish sentiment

Good results against a high bar. The news was expected, so much of it is already in the price.

Most dangerous Negative surprise, bullish sentiment

Estimates come down while crowded positions unwind. The follow-through can be severe.

Partly priced Negative surprise, bearish sentiment

Weak results, but investors were already cautious. The stock can still fall with less left to reprice.

How the rating works

The Power Rating combines the factors associated with initial drift into a single score from 0 to 100. Higher means greater potential for strength over the first five sessions; lower means greater potential for weakness.

Two questions sit at the center of it. How did the results compare with the Earnings Whisper number and the market’s broader expectations? And how was sentiment positioned going in?

The strongest signals come from the sharpest conflicts — a strong report into bearish positioning, or a weak one into bullish.

−17.6% −8.8% 0.0% +8.8% +17.6% 1–10 1 4–13 12 8–17 70 12–21 243 16–25 803 20–29 1,789 23–32 3,308 27–36 5,481 31–40 7,947 35–44 10,940 39–48 13,518 42–51 16,086 46–55 17,850 50–59 19,349 54–63 19,573 58–67 18,750 61–70 16,135 65–74 13,544 69–78 9,989 73–82 6,732 77–86 4,088 77–86 1 80–89 2,054 84–93 876 88–97 314 92–101 80 96–105 6 99–108 1
Average return from the post-announcement open through the fifth trading day, by Power Rating decile, with sample size beneath each bar. These buckets are round for readability. They are not trade thresholds — they exist to show the relationship holds across the whole scale.

What matters in that chart is not any single bar. It is that the bars go in one direction. A score that only worked at one cut point would be a curve fit; one that separates monotonically across all ten deciles is measuring something.

Where the edge actually is

The useful part is the tails. In the middle of the scale the expected advantage is small enough that it is not worth acting on.

The thresholds we treat as actionable are 67 and above for strength and below 35 for weakness.

+1.2% 67 and above · 24,141
+0.1% 35 to 66 · 134,805
+1.1% Below 35, shorted · 30,594

Average five-day return from the post-announcement open. Success rates: 56% long, 56% short, against 50% for all announcements in the universe.

The point is not to trade every report. It is to find the minority where the imbalance between results and expectations is wide enough to matter, and to leave the rest alone.

Both sides together

Taking the two tails as a single long/short portfolio, starting with $100,000 in December 2002:

$0 $1,253,662,500 $2,507,325,000 $3,760,987,500 $5,014,650,000 2002 2005 2008 2011 2014 2017 2020 2023 2026 Power Rating long/short
$100,000 starting capital. One fifth of the portfolio is deployed each trading day, split evenly across every qualifying signal that day and held five trading days — five overlapping positions at steady state. Days with no signals leave that fifth in cash. Open positions are marked at cost, so the line steps on settlement rather than moving continuously. Hypothetical, excludes commissions, borrowing costs, slippage, and the availability of shares to short.
$4,648,786,521 Ending equity
57% Annualized

December 2002 through August 2026.

Power Rating or Grade?

They start from the same earnings event and answer different questions.

The Power Rating is about the immediate setup: the surprise against real expectations, sentiment going in, positioning and the risk of forced buying or selling, and whether the opening reaction went far enough. Those things dominate the first several sessions.

The Grade is about the earnings story: sustainability, guidance, revisions, revenue and margin quality, cash flow, and whether the report changed the company’s trajectory. Those things dominate the rest of the quarter.

A stock can carry a strong Power Rating on a favorable short-term setup without the quality to support a quarter-long advance. Another can earn a strong Grade and make a poor immediate trade, because the market was already positioned for good news.

Using it

The Power Rating is published after a company reports and should be read as a short-term directional indicator. The most direct use is to focus on the tails and ignore the middle.

It is not sufficient on its own. Liquidity and spreads, the size of the opening gap, broader market and sector conditions, nearby support and resistance, company-specific news risk, borrow availability and cost, and position sizing all still apply.

A five-day average is an average. Individual stocks reverse sharply, gap intraday, or complete the entire move in the first session. The rating is an edge across a group of trades, not a forecast for any one of them.

What it does not do

  • It does not predict earnings. It is calculated after the results are known.
  • It does not predict the gap. The gap has already happened when the measurement window opens.
  • It does not assess long-term value. A favorable week says nothing about the next year.
  • It does not remove market risk. A large move in the market, the sector, rates, or the economy can bury a company-specific signal entirely.

The opening price after earnings is the market’s first attempt at valuing new information. First attempts are frequently incomplete. Investors still have to read the report, revise estimates, and move positions — and when the results differ materially from what was priced in, that process shows up as an accelerated move over the following days.

The gap tells you how the market reacted. The Power Rating helps tell you whether that reaction is likely to continue.

All figures measure the change from the opening price following the earnings announcement to the closing price of the fifth subsequent trading day. The earnings gap is excluded. The universe is limited to stocks priced above $5 with average daily volume of at least 150,000 shares over the quarter, and excludes announcements for which the five-day result is not yet available. Short results are stated as traded, so a positive figure reflects a profitable short. Results cover December 2002 through August 2026. Portfolio results are hypothetical, assume the stated position sizing, and exclude commissions, financing and borrowing costs, slippage, taxes, and the practical availability of shares to short. Past results are not a guarantee of future performance.