So Close to $15,000: My 9/16/2026 FOMC Three-Day Recap

This is the sequel to the 8/28 Jackson Hole post-mortem. Same trader, same event-driven playbook — down $2,887 last time, up $2,211.98 this time. But it isn’t a comeback story, because at the closing price I should have made $17,388.32.

---

Pre-Meeting: There Was No Way They’d Hold

September 16 was the Fed’s policy meeting. In the days before it, the loudest thing being passed around was JPMorgan’s five scenarios — every desk had a different script and nobody was willing to commit.

My own read was simple: there was no way they’d hold rates. How much and how were separate questions, but the direction wasn’t going to change. The real uncertainty wasn’t the decision itself, it was the press conference — hawkish wording and the market sells off, dovish and it rallies. So my plan was to wait for the presser before doing anything.

There was just one problem: I was in a meeting off-site that day.

---

The Decision Drops: I Read It as Dovish and Bought In

I caught a glimpse of the screen right as it was released — just the decision and the dot plot. I never got to hear the press conference.

What the dot plot told me: one more hike in 2026, none in 2027. I read that as dovish, because the market had been pricing 3 to 4 hikes and the dot plot came in softer than that.

The market was rising after the release, so I followed the momentum and bought QQQ calls (9/18 expiry, $700 strike, 3 contracts).

Then I went back to my meeting and left it alone.

---

The Twist: The Presser Was Hawkish and the Tape Rolled Over

When I got back to a screen, everything had flipped — the press conference was hawkish and the market was selling off.

Here’s the part that’s interesting in hindsight: because I was in a meeting and missed the presser, I never got shaken out of the position. If I’d been staring at the tape, the hawkish tone would have hit, the index would have started dumping, and I’d almost certainly have cut it right there.

I didn’t cut it. Two reasons:

First, $700 held. The calls I bought were the $700 strike for 9/18 expiry — and that strike was also the support level at the time. Break it, and I stop out and admit I’m wrong; hold it and bounce, and the support is real. Into the close it held, so I stayed in.

Second, I stuck to a judgment: this was “jawboning” — a fake-hawkish Fed. The wording at a press conference can be as hard as it likes, but that isn’t the same as actually hiking repeatedly. Talk is talk; the path depends on the data.

There’s one more thing worth spelling out: why did I pick the $700 strike? To keep the option in the money. An ITM option carries intrinsic value, so it doesn’t collapse toward zero the way a deep out-of-the-money contract does. That was the cushion I built in for myself.

So I didn’t close it that night — I carried the QQQ calls overnight.

The next day (9/17) the market rallied and I closed the position for +$454.92.

In hindsight I got that trade half right — my expectation that they wouldn’t actually keep hiking was correct, but I never anticipated the press conference striking a hawkish pose. What made it work wasn’t a correct forecast; it was holding a support level, on an option that wasn’t going to zero on me.

---

9/17: Memory Chips Were Strong, So I Bought SNDK and NVDA

The market gapped up on 9/17 and semiconductor memory was the strongest group on the board. I entered two positions at the open: SNDK and NVDA.

The NVDA trade: it chopped around all day with nothing much happening, so when I found a spot I took my profit and left — +$76.94. Nothing to say about it, and no reason to sit and wait.

SNDK was more complicated — I bought in at a high price.

It didn’t immediately punish me for it, so I did two things:

1. Looked at the tape into the close: a modest pickup in volume, and price closing at the high of the past three days, with the rate-hike overhang now cleared. I judged there was still upside.
2. Held it overnight — and left a limit order to take profit at $1,000 the next day.

That decision was the turning point of the three days.

---

9/18: The Limit Order Filled at the Open for $975

At the next open the limit order filled immediately. The SNDK 9/18 $1,600 call: +$975.57, a +23.1% return — the single most profitable trade of the three days.

The limit order did my discipline for me. I wasn’t at a screen; the order took the money off the table according to plan.

Then the story starts to sour.

---

Where I Went Wrong: Big Direction, No Precise Levels

Mid-session I noticed SNDK was exceptionally strong, so on a pullback I put on two more (9/25 expiry, the $1,720 and $1,680 strikes).

But those two swung violently, looked like they couldn’t hold up, I didn’t dare sit on them, and around midday I took profit: +$216.52 and +$103.51.

And then — SNDK closed up $177 on the day, roughly 11%.

The $1,600 call I sold off, its premium closed around $194 per contract. My cost was $42.24 per contract (the whole position cost $4,223.66), and I sold at $52.

Here’s the math:

	Premium	P&L on this call
Actual ($52 exit)	$52	+$975.57
Held to the close ($194)	$194	+$15,176.34
Difference		−$14,200

Run this call through a calculator and you can see where its real breakeven sits:

• Strike $1,600 + premium $42.24 = breakeven at $1,642.24

In other words, SNDK had to climb to $1,642.24 before this call was even back to even. When I sold at $52 it was still short of that — no wonder I couldn’t hold it.

That’s the real problem.

Working through it afterward, one thing became clear: what I got wrong wasn’t the act of “selling early” — it’s that from beginning to end I only ever had a directional view, and never a precise take-profit or stop-loss level.

• I had the view that “semiconductor memory is strong” → but I had no rule for “at X I’m out, at Y I fold”
• I did leave a $1,000 limit order (that part was right) → but that’s just “how much profit satisfies me,” not “how this position exits”
• On those two midday trades I decided everything by feel — any volatility and it “looked like it couldn’t hold,” so I bailed at midday

Feel is the most expensive input in the market. That day, feel left $14,200 on the table.

That said — this SNDK call couldn’t be scaled out of. SNDK itself is an expensive stock, and a single $1,600-strike contract cost $4,223.66. I held one contract, so “sell half” wasn’t on the menu. Scaling means buying two or three contracts from the start, and that doubles the capital one trade ties up.

That’s the constraint of position granularity: trading options on high-priced names with a small account, you’re often choosing between all-out and holding the whole thing.

(Which is also why I’m putting that options visualizer in here: breakeven and expiry are things you’re supposed to work out before you buy, not things you confront after the close with your hands on your head.)

---

The Rest of 9/18: One MU Trade in the Morning

MU was sluggish on Friday morning, so I did one round: the $980 call made +$377.25 (+10.1%) and the $1,000 call made +$7.27 (roughly flat). After that I didn’t touch it for the rest of the day.

That day totaled +$704.55.

---

The Three-Day Tally

Date	Contract	Qty	P&L	Return
9/16	QQQ 9/18 $700 Call	3	+$454.92	+11.6%
9/17	NVDA 9/18 $210 Call	3	+$76.94	+2.7%
9/17	SNDK 9/18 $1,600 Call	1	+$975.57	+23.1%
9/18	SNDK 9/25 $1,720 Call	1	+$216.52	+2.9%
9/18	SNDK 9/25 $1,680 Call	1	+$103.51	+1.3%
9/18	MU 9/21 $980 Call	2	+$377.25	+10.1%
9/18	MU 9/21 $1,000 Call	2	+$7.27	+0.2%
	Total		+$2,211.98	+6.6%

• All 7 trades were profitable, on $33,370.62 of capital deployed, for a three-day return of +6.6%
• By day: 9/16 +$454.92, 9/17 +$1,052.51, 9/18 +$704.55
• Had that SNDK $1,600 call been held to the close, the three days would have been +$17,388.32

---

Last Word

Afterward I told my wife about it. Her reaction was short:

Making money is good. Nobody can predict this stuff, especially same-day-expiry options with that kind of risk. If you can make this much, the risk matches.

I thought about it, and she’s right.

Everything I traded those days was near-dated directional options — no hedges, no spreads, pure direction. The other side of that position is this: when you’re wrong, it goes to zero as fast as it makes money. The $2,211.98 I made and the $14,200 I didn’t are two faces of the same risk.

So that evening we went out for a date and had a good dinner. Kicking myself over the $15,000 is one thing — the money that did land in the account is real.

---

This is a record of my own real trades, kept for personal documentation and risk education. It is not investment advice. Options carry extreme risk, and near-expiry contracts in particular can go to zero in a very short window. Please don’t copy this.