I Started Writing the Grading Sheet Before the Hypothesis

I write down a market call most mornings. A line or two before the open, so there's something to review against in the evening.

A while ago I noticed it wasn't doing much for me.

When I reread the note at night, I read the parts I got right first. If the direction was right I write "right." If the move fell short, I write "well, the direction was right." And when the direction was wrong, I go find the sentence where I hedged. That sentence is usually there, because I usually hedge.

The call was written down. What would count as right was not.

So on September 8th I added one more thing before the open. A grading sheet.

---

1. What I wrote 35 minutes before the open

The index was set to open just above 7,700, and that level happened to sit on the gamma flip — the point where dealer hedging changes sign.

📌 That flip is measured on the day's own expiry. Compute the same thing across all expiries and you get a different level. I've since read those two as if they were one number, so it's worth saying plainly: decide which one you're looking at first, or the same name refers to two different lines.

The call itself was simple.

Holds 7,700   →  chop, range-bound
Loses 7,700   →  be ready for 7,625

Nothing unusual so far. What was new is what I wrote underneath it.

#	What to look at	How it gets decided
1	Did 7,700 hold	Record both the close and the intraday low
2	Did it reach 7,625	Intraday low ≤ 7,625 counts as reached
3	Was it chop	Day's range ÷ the day's expected move. Below 1.0 is chop
4	Dow's 60-day line	Did it get there
5	Nasdaq's 5-month line	Did the rejection hold

And the verdict rules, also in advance.

• Hit: lost 7,700 and reached 7,625
• Partial: direction right, distance short
• Rejected: held 7,700 but the range expanded past the expected move

That last block is the point. At the top of the file I wrote "the standard does not get edited afterwards." It's a note to the version of me who reads this at night.

---

2. Filling in the boxes that evening

Here's how the day went.

Open    7,717.81
High    7,717.81      ← same as the open. It never traded above it
Low     7,666.99
Close   7,673.52

#	Item	Result
1	7,700 held	🔴 Lost. Gone seven minutes after the open, and never recovered
2	Reached 7,625	⚪ No. The low stopped 41.99 points short
3	Chop	🟢 Range 50.82 ÷ expected 58.4 = 0.87
4	Dow 60-day	🎯 The intraday low tagged it, and it closed back above
5	Nasdaq 5-month	🎯 Still pinned below

It landed in the "partial" box I had defined that morning. Direction right, distance short.

The point isn't that my call was half right. The point is that I had nothing left to decide that evening. I filled in the boxes and the verdict was already sitting there.

---

3. The interesting part — two boxes were true at once

Something only became visible once the table was filled in.

#1 says "lost the level." #3 says "chop." Usually it's one or the other. Days that break a level tend to run; days that chop tend to respect it.

That day the level went, and the range still stayed contained.

7,700 was lost           →  direction was down
but only 0.87 of expected →  the distance never came

I've started calling this a right-direction, short-distance day. If I had to describe that session in one line, that's the line.

And the same verdict showed up somewhere else

Options are priced on how far, not just which way. There are structures that need the distance, and direction alone does nothing for them.

That day, options priced around 7,625 didn't come to life — even though the index moved toward them. They were 41.99 points short.

The grading sheet said "partial." The option market said the same thing. Neither knew about the other.

---

4. The box that landed on the line — this is the real reason for the sheet

Look at #3 again.

The "expected move" comes from VIX1D, a one-day volatility index. In the morning I used the previous session's close — 12.03, from September 4th — which gave an expected move of ±58.4 points.

VIX1D closed that day — September 8th — at 10.43. The number I used that morning was four days old, and the day's own reading doesn't exist until the session ends.

Morning 12.03 (Sep 4 close)  →  expected ±58.4  →  50.82 ÷ 58.4 = 0.87  → chop 🟢
Closing 10.43 (Sep 8 close)  →  expected ±50.6  →  50.82 ÷ 50.6 = 1.00  → on the line ⚪

That flips the verdict.

When I found this in the evening, I did briefly think about re-measuring with the better number. The closing value does describe the day more faithfully.

Which is exactly what the sheet exists to prevent.

The only number available to me in the morning was 12.03. The 10.43 doesn't exist until the session is over. Changing the ruler after seeing the result isn't grading. It's interpretation.

So I graded it with the registered value, and wrote the near-flip down beside it. The verdict stands; the sensitivity goes in the margin.

📌 This one box is why I'm writing this at all. A grading sheet earns its keep on the ambiguous days, not the clean ones.

---

5. What I still don't know

I have a guess about why the distance never came.

There's a way to watch whether money is actually moving into puts. My working view has been that a real move down needs put buying behind it. Without it, price can bleed lower and still fail to break the level.

• August 19: put flow went the other way. The low that day was 7,700.07 — it missed the round number by 0.07 of a point
• September 8: put buying came in hard at the open, which is when 7,700 went. By the afternoon it had flipped to selling. And 7,625 never arrived

Two sessions pointing the same way. Two.

Two isn't a pattern. Getting it right twice by chance is not remotely unlikely. This sits in the "observing" column while I keep counting. I'll look again at thirty. I don't trade on it now.

---

6. What actually changed

One thing. I stapled a grading sheet under the call.

Before	After
Write the call	Write the call plus what would count as right
Interpret in the evening	Fill in boxes in the evening
Ambiguity resolves in my favor	Ambiguity gets the registered value, with the sensitivity noted beside it

Five table rows and three verdict lines. It goes down while you're writing the call anyway. (I haven't actually timed it, so I won't put a number on it.)

And I don't think this is really about trading. Anywhere you could set the standard after seeing the outcome, the same thing happens.

I've only been doing it this way for a few days. I'll write again once I've done it longer.

---

⚠️ This is a record of what I looked at and how I graded it. It is not investment advice and does not recommend any security or strategy. Options trading can result in the loss of your entire investment. Your decisions and their outcomes are your own.