I took the certain gain instead of the maximum one

I've written about why I buy a butterfly early — I buy the butterfly before price arrives — and about how the payoff curve sharpens — Watching a butterfly curve sharpen.

Both land in the same place. A butterfly gets richer as expiration approaches. So I had written myself a rule:

I don't close a 0DTE butterfly until the last 30 minutes.

On August 26 I closed two hours early.

What I was actually thinking that day was not theta, and not the rule. It was: lock in what I already had.

📌 Dated observation. SPX 0DTE, August 26, 2026 — one session, one butterfly, center 7,665, 20-point wings. Sample size: one.

---

1. What the screen looked like

SPX had sold off in the morning, put in a low near 7,657, and was bouncing.

These were the three strikes I was holding that day:

7,685 put   buy  1     ← upper wing
7,665 put   sell 2     ← center
7,645 put   buy  1     ← lower wing

Wing width 20 pts · debit 6.10

Per contract, that is $610 committed. In a butterfly the debit is the maximum loss. If it goes wrong, the whole $610 is gone.

As the bounce ran, the position traded above 10.0 and came back down — because price was moving up past the 7,665 center.

---

2. What "moving up" means in this structure

The expiration payoff is one line. One index point is $100 per contract.

Value at expiry = 20 − |close − 7,665|     (inside 7,645 – 7,685)
                                            outside that, zero

Max gain = 20 − 6.10 = 13.90 pts
Max loss = 6.10 pts (the entire debit)

(For a butterfly this is also its intrinsic value — at expiration there is no time value left.)

The breakevens sit one debit inside each wing — not one debit from the center.

Lower breakeven = 7,645 + 6.10 = 7,651.10
Upper breakeven = 7,685 − 6.10 = 7,678.90

When price climbs above the center, three things happen at once:

①	Value drops one point per point
②	The distance to the upper breakeven (7,678.90) shrinks — cross it and the gain is gone
③	The distance to the upper wing (7,685) shrinks — cross it and the entire debit is gone

① is on the screen. ② and ③ are not, and they are the ones that decide the money.

---

3. What I was looking at — distance and speed

09:44 PT   low   7,657.41
10:25 PT         7,672.86        ← +15.45 pts in 41 minutes

At that moment, above the 7,665 center      +7.86 pts
       to the upper breakeven 7,678.90      about 6 pts left
       to the upper wing      7,685         about 12 pts left

Time to the close                           2 hours 35 minutes

In a market that had just covered 15 points in 41 minutes, 6 and 12 are not long distances.

So what I saw was not "the value slipped a little." It was "six points to where the gain disappears, twelve to where all of it does — and those distances are closing fast."

---

4. So I locked it in

Around 10:30 I sold at 8.55. I had paid 6.10, so +2.45 points.

That was not the maximum. The theoretical maximum was 13.90, and the position had traded above 10.0 earlier — I sold below both.

The decision that day was one sentence:

Even when it isn't the maximum, I take the gain while there is a gain and no loss.

The point isn't how much I made. It's not letting a position that is already ahead turn into a loss. Giving it back means +2.45 becomes −6.10 — a round trip of 8.55 points.

---

5. What actually happened next

Time (PT)	SPX	Value at expiry
Exit (~10:30)	~7,673	sold at 8.55
12:20	7,689.47	🔴 zero — past the 7,685 wing
12:30	7,686.45	🔴 zero
Close, 13:00	7,675.70	9.30

This column is what the fly would have settled at, not the quote at that moment. The 8.55 on the exit row is a live fill — before expiration the fly trades below its expiry value because time value remains. That gap is the subject of Watching a butterfly curve sharpen.

At 12:20 price was outside the wing. It did reach the level I had judged it could reach — price covered those remaining twelve points before 12:20, in under two hours.

Now look at the rule again. It said hold until 12:30. Had the position settled at that moment, it would have expired worthless.

---

6. The rule wasn't wrong — it left a condition unstated

Here is the original reasoning: "Theta accelerates into expiration, so hold until the last thirty minutes."

The first clause leaves a condition unstated. Theta works for me while price is inside the range that pays.

Where price is	What the remaining time does
Inside the breakevens	The two short puts at the body decay faster than the wings, and the fly gains
Outside the breakevens	Time to come back is running out — the position decays toward its expiry value, and past the wing that value is zero

Once price is outside, the remaining time isn't "time to accrue value." It's "chances to come back." And the closer expiration gets, the faster those chances disappear.

So I rewrote the rule:

I don't close a 0DTE butterfly until the last 30 minutes — but only while price stays inside the range that pays. If price moves past the center and keeps going toward a breakeven, I look at closing there.

I didn't discard the rule. I added a condition to it.

Both breakevens are fixed the moment the debit is paid — two lines of arithmetic: lower wing + debit and upper wing − debit. I hadn't written them down that day; I only measured the distances after watching the value slip. Writing them at entry would have collapsed my intraday decision to one question: is price still between them?

---

7. What I still don't know — and this day's counterexample

The close was 7,675.70, and the value at expiry was 9.30.

That is 0.75 points above the 8.55 I sold at. Holding to the close would have paid a little more.

And 7,675.70 sits inside the breakevens. So on this day price went outside the wing and then came back into the paying range to finish.

I'm writing that down because this post must not read as "sell early." On results alone, holding was better that day.

But capturing that 0.75 meant sitting through a stretch where settling at that moment would have wiped out the entire debit. I couldn't find a reason to sit through it. The position was already ahead, and what was at risk was the whole debit.

Sample size: one. How often price finishes back inside after leaving the paying range is something I have to count. When there are enough of those, I'll write it up.

---

8. Summary

• The maximum gain and the certain gain are different columns. The market sets the first; I set the second
• The further a butterfly moves from its center, the more value drops and the shorter the distance to zero gets. Both worsen together
• Breakevens are lower wing + debit and upper wing − debit — not center ± debit — and they're fixed at entry
• Time is on my side only inside the range that pays
• Rules like "hold to the close" can carry unstated preconditions
• Sample size: one. I'll keep counting

---

※ This is a record of my own trading and the reasoning behind it, not a recommendation to trade any security, strike, or structure. The sample is a single session and there is no guarantee the same setup produces the same outcome. Same-day options can lose the entire amount committed within one session. All investment decisions and their outcomes are your own.