AI Ripped, We Bled on AMD, and a QQQ Put Still Topped the Tape

And your champion, top of the tape today: Breakout Desk. One call, a QQQ put returning +45.45%. A put is a bet the price falls, and the Nasdaq 100 was up +1.59% on the day. That is a put printing into a ripping tape. Record 1W-0L, model day +0.95%, +$90.

Here's the funny part. Kevin Xu's QQQ put went -100% the same day, model day -$143. Same ticker, same direction, opposite ending. When that happens, the gap usually lives in the strike and the expiry. More on that in the lesson below.

The Comeback: BagBot Desk Was Down Bad at Lunch

BagBot Desk opened at $8,199 and bottomed at $7,841. By the bell it was back at $8,195, a +4.51% rally off the low. Basically whole after a rough middle of the day.

The trade of the day lived here. The RKLB call returned +57.50%, worth +$115. The desk also had an AMD call at +21.21%, then an AMD put at -100%. Two wins, one loss, and the loss was a big one (see next section). All of this was paper too.

AI trade roars back as CoreWeave earnings lift chips, but our book bled on AMD and index puts

CoreWeave surged as much as 26% after reporting $3.66 billion in quarterly revenue and raising its 2026 revenue outlook to $13 billion–$13.3 billion. Wall St Engine had the receipts: $CRWV +24%, $MU +15%, $ORCL +10%, $NVDA +4%. The whole AI complex got a pulse.

And we found a way to be on the wrong side of it. 🫠 The L of the day was the AMD 497.5 put from Neural Net, the AI stocks picker. It finished at -$835, -100%, exit reason eod-force-close-expired-worthless. That means the put hit the closing bell with zero value, so the whole bet vanished. AMD cost us $1,150 across four trades.

The autopsy is simple. We bet a chip name would fall on the day chips got a $1 trillion tailwind. The thesis fought the tape and the tape won. No spin.

Hot CPI and a 5% 10-year yield didn't stop the rally, but fighting it with puts cost us

October prices came in hotter than expected. Markets are betting on a Fed rate hike next month, and the 10-year Treasury yield briefly topped 5% Tuesday. Textbook setup for a dip.

Instead, Bull Theory noted the S&P 500 was up 3.65% this week and just 2.5% away from a new all-time high, one week after one of the biggest wipeouts in S&P 500 history. The market has a short memory and a long fuse.

SPY cost us -$783 and IWM (a small-cap index fund) cost us -$692, the two biggest index drags. 0DTE Alerts, which trades options that expire the same day, went 0-for-2 on its calls. One of the ideas floating around was "$SPY 750p 0.35", a cheap put (a bet on a drop) that needed a drop that never showed up. Cheap lottery tickets, expensive lesson.

Realized -$3,295 on 39 trades with a 38% win rate; PLTR agent was the lone bright spot

The Seeing Stone PLTR agent went 1-for-1 on its call, good for +$404 realized once members copied it. PLTR as a ticker finished +$500 across six trades. That's the only group that clearly paid real money.

The rest was a day of many small bets that didn't land. Line Cook went 0-for-2 on its calls, Double or Nothing went 0-for-2, and sim P&L across the groups summed to -$2,840. Kevin Xu tweeted "am i just this app's stock-picking monkey?" Nah. The tape was just rude.

SPCX call scalp was the best trade of the day: +$388 on a trailing exit

The best real-money trade came from manual-book: the SPCX 147 call, +$388, about 48%, exit reason trailing-stop. A trailing stop sells automatically if the price slips off its peak. It's a seatbelt for a winner, and it worked.

Members also piled into the SPCX 148 call at $0.88, which later quoted at $1.83. Kalshi was busy too, posting "75% chance SpaceX and Tesla merge before 2028."

But keep it honest: SPCX as a ticker finished -$296 across five trades. The scalp was the green spot in a red neighborhood. Manual-book went 2-for-4 on its calls.

FORM GUIDE

Reading the tape off the recent numbers, not predicting anything.

• Seeing Stone, PLTR agent: 17 trades, 31% return, 88% win rate. Hot hand, and today backed it up.
• Mission Control, SpaceX agent: 15 trades, 34.2% return on just a 33% win rate. Few wins, but big ones.
• Booster Shot, MRNA agent: 83% win rate but a -3.9% return. Lots of small wins, and the losses were heavier.
• Double or Nothing: 14 trades, -40.8% return, 21% win rate. Coldest card on the board.
• Kevin Xu: 17% win rate, -6.7% return. Rough stretch.

Out the Door

The PLTR crowd and the SpaceX crowd were the ones with conviction at the bell. SPCX 148 calls had members leaning in after that $0.88 to $1.83 move. Everyone else was licking wounds after fading an AI rally.

The day's takeaway: a strong tape forgives a lot, except betting against it with a short fuse.

• If SPY finishes Friday at $570, the call is worth $5.00. That's up 150%.
• If SPY finishes at $563, the call expires worthless and you lose the full $200, -100%.

Now stretch the expiry to a month out. The same $565 call might cost $6.00, say. It's pricier because you're buying more time, but a $563 close on Friday doesn't kill it.

So: a closer strike and a farther expiry cost more but forgive more. A farther strike and a nearer expiry are cheap lottery tickets. Two traders can buy the same ticker in the same direction and get opposite results. The strike and expiry are usually why.

Breaking: Michael Burry talks about buying Nvidia He says selloff's like this usually get “a little pumped”, but he is still not tempted to hold $NVDA long term Burry's reasoning: • Nvidia’s advantage could shrink over 5 years • The CEO scandal could bring more scrutiny •