SKEW
ABOUT

Do you want to make money?

Enter, Skew.

The Skew score measures one thing: how big a move to expect around the next dated event, which is usually an earnings report. The size of that move is the part you can estimate with some confidence. The process is very straightforward: you enter a ticker symbol; it recognizes the date (which can also be entered manually) and gives back a score out of 100 that is compiled using Nasdaq, Yahoo, S&P DJI and Tradier data sources. There is a whole set of additional metrics for research purposes (pullback levels, insider activity, reward/risk ratio etc.), but they don't matter to me.

RUN SKEWGRADE A TICKER

What counts as a catalyst

Catalyst? Any kind of trigger for the stock that has a day that it occurred. Earnings, for the most part. FDA decisions, IPOs, additions to indexes, splits. But an acquisition rumor posted by someone at 2am is not a catalyst, since there is no day to highlight, so there is nothing to grade.

How it works

01
Find it

Begin with the Calendar. Earnings announcements, FDA events, IPOs, indexes, splits, dividend, and what's making the most noise today all in one place.

02
Grade it

Only one number. Mostly it is the size of movement in earnings that matters and then comes probability, the date, short interest, insider info, and the entry point. Skew selects Long or Short position for you. The missing information is just left empty. Skew does not make assumptions.

03
Track it

Record the transaction in your Ledger. After the smoke clears, write a post-mortem. Two scrappy lines is plenty, about what happened. Closed trades go on the Leaderboard. Post mortems contribute to your Process Score, which, in my view, reveals a lot more about you than any P&L ever will.

Why the grade holds up

Most research sites are built for people who plan to hold a stock until their grandkids inherit it. Skew cares about one event and the week or so around it. Unlike AI tools, it avoids hallucinated or outdated information. It won't promise a stock goes up either. Can't. What you get is how big the move should be and how far to trust that, plus a Long or Short lean that's usually about the width of a hair.

Make it yours

Same settings for all at the start. Do I screw up my weights? I'll admit it. Create a Variant, change its weight or floor price, and make sure it knows what to ignore.

GRADE A TICKER →

How grades get assigned

One grade between 0 and 100 based on where Skew wound up. All about magnitude and confidence. The direction is a coin toss. So I won't put lipstick on that pig either, and just call it what it is. Magnitude is what is being graded. A fantastic one ends up in the 80s. Hardly ever gets any higher than that.

Long or Short? Skew picks

Skew studies the market, the price action of the stock, the size of the company, and its short interest to figure out which way has more probability, then grades that lean. The stock has a Long when there is a 50 percent or higher chance of its going up, and a Short otherwise. Anything from 49% to 51% gets stamped “Slight lean”.

You'll never see one stock shown both ways. Green for Long, red for Short.

90–100
Very big, well-measured move
Unicorn territory. Options and past reports agree on a big earnings move, the date is confirmed and the entry is good.
75–89
Big, well-measured move
Large move, carefully measured, confirmed date. Definitely a place to look into.
60–74
Solid move
Decent move. Something is a little off about it.
40–59
Ordinary move
There's something on the calendar, but the move looks normal or it's hard to size.
20–39
Small or unclear move
Either the move's tiny or nobody has confirmed the date yet. A crummy entry price can sink a stock down here as well.
0–19
No clear move
Usually nothing dated. Or nothing really worth looking at.

Numbers for those interested. The earnings-only move is worth up to 45 points, meaning option prices with the regular daily jiggle scraped off, plus how the stock moves on its own from prior announcements. Ten for confidence in the size. The date is worth 15, although a guessed date gets half credit and limits the total grade to 75, no joke. Odds in your favor get 15, depending on market conditions, and they stay inside a 48%-56% chance of the stock going up. Days to cover, 5 points: roughly how many days of normal volume the shorts would need to buy back their shares. If you're Short and that number is high, watch out. Insiders, 2. Entry price, 8. All grades are for the side that Skew chooses. Beating estimates doesn't mean a stock rises, weirdly, so the direction of past reports earns zero. IPOs have their own grading. When a number's missing, we leave it missing.

Behind the grades

Our backtest first. Then the published research we checked it against.

Skew's backtest of 24,000 S&P 500 earnings reports, 2014-2026

Data: Yahoo Finance prices and earnings data, FINRA short interest. About 24,000 earnings reactions across 490 current S&P 500 companies, January 2014 to September 2026. A reaction is the close before the report to the close after it, with before-open and after-close reports lined up correctly. Short interest comes from FINRA's history from December 2017 on.

What we found:

  • Direction was a toss-up. Using past reactions and beat streaks we called it right 50-51% of the time, and just yelling “up” every single time did better, 52%. Embarrassing.
  • Market conditions make a small contribution. When the S&P is below its 200-day moving average, stocks have gained on earnings 54% of the time (and 55.5% according to the crude count, but I have rounded it down), and when above, around 51%.
  • The beaten-down stocks and small S&P stocks popped after earnings slightly more often.
  • Days to cover helped. Barely.
  • Thus most of the weight in the grade hinges on the size of the earnings-only move, pried out of option prices.

Skew's direction calls hit around 50-56% in testing, so the odds are capped at 48-56% and it never claims better.

Just one problem. The test only works with firms in the S&P 500 right now, and those that were kicked out are not there, so it might appear a little more positive than it should have.

Peer-reviewed studies

Working papers and other context (not peer-reviewed)

Vendor studies (not peer-reviewed, lower confidence)

They come from trading tool vendors, and therefore should be taken with a grain of salt.