For 110 of the 119 minutes this position was open, the screen showed red or barely green. The worst tick was -47%. At 3:49 PM — ten minutes before the exit — the trade was still sitting at -$27 with the short call sticking out the wrong side. Then SPX collapsed back into the short strike like it had been pulled there, the structure gave up its remaining premium, and the bot booked +$910 (+82.7%) in the last nine minutes of the day.
A discretionary trader does not get that money. A bot with no profit target and no stop loss does. That’s the whole post.
| Date | Bot | Entry SPX | Exit SPX | Credit | Debit | Hold | P&L |
|---|---|---|---|---|---|---|---|
| Wed May 13 | 2 PM Iron Fly | $7,442.92 | $7,445.12 | $11.00 | $1.90 | 119 min | +$910 (+82.7%) |
⚙️ The Setup
The 2:00 PM Iron Fly is the same bot we paid out a $2,050 claim on the day before . The configuration hasn’t moved: ATM straddle, $40 wings, no profit target, no stop loss, time-limit exit at 3:59 PM. The only decision the bot makes after entry is what minute to print P&L.
At 2:00 PM SPX was at 7,442.92. The bot sold the 7445 straddle and bought the wings $40 out.
- Underlying: SPX @ 7,442.92
- Short Strikes: 7445 C / 7445 P
- Long Wings: 7485 C / 7405 P
- Entry Credit: $11.00 ($1,100 per contract)
- Max Loss: $2,900 (wing width minus credit)
The setup is deliberately fragile to direction and aggressively in favor of theta. A flat tape from 2 PM to close means the short straddle bleeds and we keep almost the whole credit. Any sustained move puts one short leg deep in the money and the long wing too far away to help. That’s the trade we sold.
📉 The First 21 Minutes
The position opened green by a few dollars. By 2:11 PM SPX had drifted up four points to 7,450 and we were still +$20. Then the tape went from drift to push.
- 2:14 PM: SPX 7,452 — P/L -$73
- 2:15 PM: SPX 7,454 — P/L -$215
- 2:16 PM: SPX 7,458 — P/L -$505
- 2:21 PM: SPX 7,459 — P/L -$520 (-47.3%)
That is what short-strike gamma looks like when the underlying runs through you. A 16-point SPX move in 21 minutes took the short call from $5 to about $18 of value while the long 7485 wing barely budged. The position value had ballooned from $1,090 at entry to $1,620, and the bot logged the tick and did nothing.
This is the moment a discretionary trader closes. A 47% drawdown on an ATM fly with most of the session still ahead is genuinely uncomfortable. The bot’s only rule is the clock, so it stayed in.
The Middle Hour
From 2:21 PM the run exhausted itself and SPX started drifting back toward the strike. By 2:51 PM the position was effectively flat (-$15). The next sixty minutes were chop. SPX oscillated between 7,450 and 7,456, gamma cut both ways, and the P&L bounced with it.
A few of the peaks and troughs from that stretch, just to show what “chop” actually looks like on a position this short-gamma:
- 3:01 PM: SPX 7,450 — +$55
- 3:11 PM: SPX 7,455 — -$88
- 3:26 PM: SPX 7,453 — +$135
- 3:31 PM: SPX 7,456 — -$88
- 3:36 PM: SPX 7,452 — +$220
- 3:41 PM: SPX 7,455 — -$33
- 3:48 PM: SPX 7,457 — -$122
- 3:49 PM: SPX 7,456 — -$27
If you’d glanced at the position at 3:36 PM and seen +$220 (+20%) you might have closed it. If you’d glanced ten minutes later and seen -$122 you might have closed it for a different reason. The bot did not glance. The bot also does not have a profit target — the configuration says explicitly that 20% is not enough premium captured to justify giving up the theta still sitting in the position. The full credit lives in the last fifteen minutes.
The trade looked indecisive because the trade was indecisive. SPX was hunting for a level into the close and the fly was just along for the ride.
🔔 The Last Nine Minutes
At 3:50 PM SPX dropped from 7,456 back down to 7,455 and the position flipped to +$28. Nothing unusual. One tick later, SPX dumped six points to 7,449. Here is the tape from 3:50 onward:
| Time | SPX | Position Value | P/L | P/L % |
|---|---|---|---|---|
| 3:49 PM | 7,455.77 | $1,127.50 | -$27.50 | -2.5% |
| 3:50 PM | 7,454.96 | $1,072.50 | +$27.50 | +2.5% |
| 3:51 PM | 7,448.98 | $605.00 | +$495.00 | +45.0% |
| 3:52 PM | 7,446.64 | $477.50 | +$622.50 | +56.6% |
| 3:53 PM | 7,446.94 | $490.00 | +$610.00 | +55.5% |
| 3:54 PM | 7,447.44 | $502.50 | +$597.50 | +54.3% |
| 3:55 PM | 7,445.55 | $405.00 | +$695.00 | +63.2% |
| 3:56 PM | 7,445.19 | $325.00 | +$775.00 | +70.5% |
| 3:57 PM | 7,445.78 | $290.00 | +$810.00 | +73.6% |
| 3:58 PM | 7,445.50 | $240.00 | +$860.00 | +78.2% |
| 3:59 PM | 7,445.12 | $190.00 | +$910.00 | +82.7% |
SPX printed within twelve cents of the 7,445 short strikes at the close. Both straddle legs went from worth nearly $11 combined down to about $2. The wings collapsed to nickels. The exit debit was $1.90 against an $11.00 credit. Nine minutes did everything.
This kind of close — index pinned to a round number, premium evaporating into the bell — is not rare for SPX. It is one of the structural reasons the no-stop-no-target configuration exists. The bulk of the bot’s expected value sits in the final twenty minutes of the session, when theta is screaming and gamma is going to do what it does. Cutting the trade at 2:21 PM, or at 3:36 PM, or at 3:48 PM, throws that expected value away.
What This Trade Cost (Almost) and What It Took to Hold
The expensive part of running this bot is not the losers. The expensive part is the temptation during trades like this one. Three separate moments today were genuine off-ramps:
- 2:21 PM, -47%: the worst tick. Hard stop discipline says cut. The bot’s configuration says no, because a -47% mid-trade reading on an ATM fly is statistically just noise — the underlying has not yet expired and the position can mean-revert in any direction.
- 3:36 PM, +20%: the brief peak. A profit-target bot closes here. This one doesn’t, because $220 captured against $1,100 of credit at risk is leaving 80% of the work on the table.
- 3:48 PM, -$122: the relapse. A trader who’d been watching for two hours and was just relieved to be near flat would close here. The bot logs the tick.
None of those off-ramps happened, because none of them exist in the code. The only exit is the one at 3:59 PM. The +$910 was not a forecast and the bot was not patient — patience is a human virtue and the bot has no virtues. It just doesn’t have a button to push.
That, in one chart, is the difference between a system and a feeling.
The edge isn’t in any one trade. It’s in not being able to ruin it.
Related Articles
- Paying Out an Insurance Claim (0DTE SPX Iron Fly) — same bot, the day before, the opposite outcome. The no-stop configuration cost $2,050 on a one-sided tape. Today it earned $910 by sitting through a -47% drawdown. Both are the same decision.
- 0DTE SPX Iron Fly Recap: The Discipline of Probability (May 4–8) — same bot, full week. Wednesday of that week had a 9-minute gamma spike against us that turned a manageable position into -133%. Today was the inverse: a 9-minute gamma collapse for us. The bot doesn’t know which one it’s going to get.
- +$388 in 53 Minutes: The Day the 9:32 AM Iron Condor Worked — same trading day, different bot. The morning IC closed cleanly at a 35% profit target in under an hour. The afternoon fly did the same dollar work in nine minutes after looking wrong for two hours. Two very different paths to a green day.
Disclaimer: This log is for educational purposes only. 0DTE options carry significant risk. All trades described here ran in paper-monitoring mode. Always trade within your risk tolerance.