94% of my winning ITM credit spreads were closed before expiration — finding the disposition effect in my own trade log

In short — I re-checked 260 credit spreads. Among in-the-money (ITM) spreads, 51 of the 54 winners were closed before expiration, and only 3 of the 55 I held to expiration won. This isn't a story about expiration being dangerous. It's a story about me closing positions that were working and holding the ones that weren't. Behavioral finance gave that pattern a name forty years ago: the disposition effect.

1. I kept running into the same rule

There's a rule I've labeled an "iron rule":

No in-the-money credit spreads. No credit spreads sold past the opposite wall.

An ITM credit spread puts the short strike on the in-the-money side of the current price. It collects a large credit, and the max loss — width minus credit — is small. On paper, it looks attractive.

Here's the history of that rule, in order:

Date	What happened
Aug 14	ITM credit spreads accounted for 63% of the day's loss. I remembered it as one mistake on the call side; the log showed I'd done the same thing on the put side. "Not a one-off misjudgment — a habit." That's when I wrote it down as an iron rule
Aug 19	Carving out an exception, I added a note to myself: "on days when net put flow is falling, better not to use this ITM approach"
Sep 23	Two credit spreads sold past the opposite wall. Both expired at max loss
Oct 6	On a day when net put flow fell from open to close, two ITM put credit spreads with short strikes above the call wall. Then, writing up that day's journal, I left this rule out of the rule log

That last line is why this post exists. Breaking the rule mattered less to me than the fact that it didn't come to mind — even in the calm of writing up the day.

So I set the feelings aside and went back to the trade log.

2. I went back through 260 trades one by one

📌 Dated observation — SPX 0DTE credit spreads, Aug 11 – Oct 6, 2026. 260 trades, one trader. The sample is small.

From August 11 to October 6, I pulled all 260 SPX same-day-expiration (0DTE) credit spreads from my live accounts, one by one. I compared SPX at the minute of entry with the short strike; if the short was more than 2 points in the money, I classed it as ITM.

118 ITM credit spreads	Trades	Winners	Win rate
Closed before expiration	63	51	81%
Held to expiration	55	3	5%
Total	118	54	46%

51 of the 54 winners (94%) were closed before expiration. The three that won at expiration were all on October 5, the same 7,770 call credit — 2 to 9 points in the money at entry. SPX still settled slightly above the strike, but below each trade's breakeven, so each kept part of the credit.

And this wasn't only about ITM spreads.

Credit spreads held to expiration	Trades	Winners
ITM	55	3
ATM (±2 points)	5	1
OTM (out of the money)	15	1
Total	75	5 (7%)

Even out-of-the-money spreads won just 1 of 15 when I held them to expiration. What decided the outcome wasn't the structure — it was which positions I chose to hold.

The easy conclusion — and why it's wrong

"Close early and you win 80% of the time. So close early."

That's not what the table says. The causation runs the other way. These didn't win because I closed them early; I closed them early because they were winning. Most of the ones that were losing, I didn't close — I held them to expiration.

How I use ITM credit spreads makes this clearer. I usually place the short strike near my target level. When the index reaches the target, the short strike is roughly at the money; there's still a fair amount of credit left, but the trade is in profit, so I close it. That's the kind of trade I've made most often.

In other words, the winning side had an exit set in advance — the target. The problem was the other side. If the index didn't reach the target, I hadn't decided when to close. Where there was no exit, "what if" filled the gap.

So what my log shows is half of the disposition effect. Closing the winners was the plan; the part that reads as the disposition effect is holding the losers.

In my October 6 journal I wrote:

"Why didn't I close them? Only one reason was left — "what if it turns back up and breaks 7,850." That wasn't a view backed by evidence."

I suspect many of those 75 held-to-expiration trades rested on the same kind of "what if."

3. This has a name — the disposition effect

The disposition effect is the tendency to sell winners too early and hold losers too long.

1985 — Hersh Shefrin and Meir Statman named it. Their paper's title is the effect itself: "The Disposition to Sell Winners Too Early and Ride Losers Too Long" (Journal of Finance, 1985). Among the causes they pointed to: mental accounting (keeping a separate ledger for each position and avoiding "booking" a loss), regret aversion (closing a loser means admitting you were wrong), and difficulty with self-control.
1998 — Terrance Odean studied 10,000 individual accounts at a discount brokerage (1987–1993). Investors sold winning stocks noticeably more often than losing ones — and the winners they sold went on to do better than the losers they kept. In other words, the habit wasn't justified by what happened next (Journal of Finance, 1998).
2005 — Peter Locke and Steven Mann looked at full-time floor traders on a Chicago futures exchange. The professionals also held losing positions longer than winning ones. They found no clear evidence that holding losers longer cost these traders money — but the more disciplined traders did better afterward (Journal of Financial Economics, 2005).

The third study mattered most to me. It means this isn't a habit that experience simply wears away. Professionals have it too. The difference wasn't whether it was there, but how well it was managed.

4. For me, ITM credit spreads were where the disposition effect came easiest

Looking back, ITM credit spreads had one feature that fed this habit.

The max loss is small. Collect 4.35 on a 5-point-wide spread and the most you can lose is 0.65. A small number makes holding on feel "cheap." "If this is all I can lose anyway, I might as well wait and see."

	How closing feels	How holding feels
When it's winning	"Lock it in — feels good"	"Worried the gain will vanish"
When it's losing	"I have to admit I was wrong"	"The max loss is small, so what if…"

The diagonal of that table is the disposition effect. On the winning side, closing is the comfortable choice; on the losing side, holding is. ITM credit spreads make holding the loser especially comfortable — because the loss is capped and small.

But small losses add up. As I wrote on October 6, a capped loss is still a loss.

Still, as the second table in section 2 shows, the same thing happened with out-of-the-money spreads. ITM spreads made the habit easier — but the habit wasn't in the structure. It was in me.

And on a green day, a second effect joined in

On October 6, I'd already made a solid day by 08:29 PT. After that, I opened more upside positions.

In experiments, Richard Thaler and Eric Johnson showed that people take on more risk after a prior gain — the "house money effect" (Management Science, 1990). Money already won feels less like "mine" and more like the casino's.

That day I did both at once: I bet more with the day's gains (house money) and held on to the losing positions (disposition effect).

September 24 had the same shape. After one put made a big gain early in the session, I made five more bearish put trades — and when I cleaned up before the close, I left one contract open because its loss had grown too large. In that day's note I gave the reason in one word: "attachment."

5. Which is why the rule didn't come to mind

Back to the question from section 1: why didn't a rule I'd called "iron" come to mind in the moment?

Here's my answer. I set the rule while I was calm, and I broke it while I wasn't. The person writing the rule on the night of August 14 and the person at the screen at 08:51 PT on October 6 were the same person, but not in the same state. The disposition effect and the house money effect do their work in exactly those not-calm moments.

That's why relying on memory kept bringing me back to the same place. The four dates above are the record of that.

6. Handing it to systems instead of memory

Here's what I'm doing now. These aren't proven fixes — they're systems I'm testing.

I keep the account small. I don't rely on willpower; I rely on account size. Resolutions bend with the day's mood, but you can't bet money that isn't in the account.
Orders are meant to go through a pre-trade checklist. The rule is to enter only when five checks are green: time, risk, boundary, signal, playbook. The ITM credit ban is one of those checks.
I turn "what I sense" into "what shows up on the screen." On October 7 I added a check: when net call flow turns down from its day high, new upside entries get a warning, along with a prompt to close the upside positions I already hold. This one is aimed squarely at the disposition effect — when I'm inclined to hold a losing position on a "what if," the screen asks first.
I listen to my Super Trader Songs to keep the rules in mind. So I don't lose sight of my principles, I listen to the trader songs I wrote, over and over. One of them, "By the Rules 1," repeats "Wait. Read. Feel. Confirm." — listen on YouTube · all songs on this site

But systems can be routed around too. On October 6, only 1 of my 14 orders went through the checklist. There's the same gap between building a system and using it. So the conclusion of this post isn't "solved."

6-1. What I changed the night I wrote this

This post left me with two things: the winning side had an exit and the losing side didn't, and I don't know yet whether systems can reduce the disposition effect. So I turned those two things directly into systems. They go live on October 8, and I don't know yet whether they'll work.

① I write down the exit on the losing side first

I added one more field to the pre-trade checklist: "If this position is wrong, where do I cut it?"

A price (e.g., the index closing below a certain level)
A time (e.g., if it hasn't reached the target by a certain time)
A flow signal (e.g., net put flow starting to rise again)

Until one of the three is filled in, the proceed button stays locked. Where to cut is my call. The system only blocks a blank field. Just as the winning side has a target, the idea is to set one exit on the losing side before entering. If I wait until I'm losing to decide, it isn't an exit — it's a "what if."

② 40 minutes before the close, holding becomes a decision too

At 3:20 p.m. Eastern (40 minutes before the close), the app opens a window with a sound. It lists every open position that's currently losing, and next to each one it shows three things:

The current unrealized P&L
The loss-side exit I wrote down at entry
My own past record — for example, "0DTE credit spreads held to expiration: ○ wins out of ○"

Then there are two buttons: [Close it] and [Hold it — write the reason]. To hold, I have to write a reason.

The disposition effect doesn't happen through a decision. It happens through doing nothing — just leaving the position there. So I turned leaving it there into a decision I have to put in writing. This window looks only at positions that are open right now, no matter how the order was placed, so it shows up even on days I skipped the checklist. It's aimed at the gap I described in section 6, between building a system and using it.

③ On a green day, it asks one more question

Once the day's realized gain reaches my daily target, any new entry triggers a question on the checklist: "Is this entry a bet with money I just made?" It doesn't block me. I have to write a reason to proceed. This field is aimed at the house money effect in section 4 — the "betting more after a win" of October 6 and September 24.

④ A yardstick for whether it shrinks

If I build the systems and stop there, this post becomes another resolution. So every day after the close, I measure my log with the same yardsticks the disposition-effect research used. Below is the baseline: the 30 days before the systems went live (September 7 – October 7).

Baseline (30 days before the systems)	Winners	Losers
0DTE credit spreads: share held to expiration	6 of 73 winners	44 of 76 losers
0DTE credit spreads, holding time (median)	130 min	180 min
Single-option buys, holding time (median)	40 min	54 min

I held the losers longer, and more often all the way to the end — the same shape Locke and Mann saw in Chicago floor traders.

📌 This is a different cut from the tables in section 2 (August 11 – October 6, "how many of the 75 held to expiration won"). Different period, and it divides the other way: this table asks, "of the winners, how many were held to expiration — and of the losers, how many?" Because the period (it includes October 7) and the set of trades differ, the two tables' numbers don't line up exactly. Same log, seen from the other side. The next 30 days will be measured against this table.

One more. I adapted the method Odean (1998) used. Each time I close a position, I count it as a realized gain or loss, and count every other position open at that moment as a paper gain or loss (using unrealized P&L saved every 5 minutes).

Proportion of gains realized (PGR) — of positions in profit, the share I actually closed: 0.207
Proportion of losses realized (PLR) — of positions at a loss, the share I actually closed: 0.157

I closed winners about 1.3 times as often as losers. The disposition effect showed up in the numbers too. That's from 214 closing moments over 30 days.

⚠️ These two ratios count only positions I closed myself. Positions that ended at expiration settlement aren't counted as "closing moments," so losses carried to expiration are never counted as realized — which means this 1.3x may lean toward overstating the disposition effect. That's why the main evidence in this post is the "share held to expiration" row in the table above, with the two ratios as a secondary measure. From the next measurement on, I'll show both definitions side by side — with and without expiration settlement.

If the systems work, the "Losers" column and the gap between those two ratios should shrink. If they don't, I'll write that down just the same.

6-2. And I turned it into a song — "Two Doors"

A rule document gets read when you're calm. To have it come back in the moments when I'm not, I turned this post's conclusion into a Super Trader Song — a classic rock track that picks up "Wait. Read. Feel. Confirm." from "By the Rules 1" — in this one it's sung as "Wait, Read, Feel, Know."

(Chorus, translated from Korean)

Two doors! I already know The winning door, the guarding door — both are in my hands Two doors! The moment I close the door I rise even higher

The "guarding door" is the exit on the losing side. It's a song about setting both doors before I go in, and closing whichever one I reach with a light hand. I plan to listen to it before entering a trade, and before closing one. (The song is in Korean with English hooks.)

🎧 Listen to "Two Doors" on YouTube · All songs on this site

7. What I don't know yet

"5 wins out of 75 held to expiration" is one trader's record over two months. I left out 13 trades with no SPX value for the minute of entry. It doesn't mean the same ratio would show up for anyone else, or in any other market.
This table doesn't tell you when to close. It's mixed with a selection effect: trades look like winners because I closed them while they were winning.
I now write down a loss-side exit at every entry. Which kind of exit — price, time, or flow signal — actually cuts losses, I'll only know once the written exits and their outcomes pile up. Right now the sample is zero.
The spots where ITM credit spreads did win (under the call wall while net call flow was falling; in the direction of a strong trend) seem to share something, but the sample is small, so I haven't made it a rule.
I'll measure whether the systems reduce the disposition effect using the table and the two ratios in 6-1. The baseline is the 30 days before the systems went live. In 30 days I'll measure again with the same yardsticks and add the results to this post. This is one trader's record from my live accounts, and with unrealized P&L sampled every 5 minutes, the ratios are approximate.
I'm also collecting the reasons I give when I choose "Hold it." Because each one is written before the outcome is known, I'll be able to sort them honestly later — evidence, or "what if."

This is one trader's record. It describes what I did and what happened, not a recommendation to do anything. With options, you can lose the entire amount at risk in a short time. Your decisions, and their outcomes, are your own.