Friday's tape is day 2 of a relief rally, not a follow-through day. Breadth is flat (RSP not confirming). And the audience's behaviour — JP Morgan's record retail-buy print — is exactly what the base-rates data says loses money: median max drawdown 85%, 2.5-year recovery, more than half never come back. Matt initiated AMZN + GOOGL live as long-term positions, set a 4%-risk-defined NVDA entry with a 26-day earnings countdown, and walked through FCUV as the show-don't-tell of "take profits, move stops up."
Thursday's retail print was the highest in a decade — concentrated in NVDA, AMZN, S&P. But the historical record on the 6,500-stock sample from 1985–2024 is brutal: median max drawdown 85%, median recovery 2.5 years, and over half never reclaim prior highs. Matt's read: "The setup is the entire opposite of what the crowd is doing." Two long-term positions initiated live on the call — AMZN on the FTD pivot (258–268 range, target 278), GOOGL on the 20-day reclaim — close the 30 Jul MSFT-initiation arc and build the portfolio on rule-based triggers, not on the headline.
The bounce is still day 2 of an unconfirmed rally attempt — FTD doesn't print until Tuesday at the earliest. The retail record-buy is the trap, not the signal: 85% median drawdown, 2.5y median recovery, half never come back. The portfolio construction is now MSFT + AMZN + GOOGL (three long-term positions, all rule-based), with NVDA risk-defined at 4% and 26 days to earnings. Don't buy the dip. Wait for follow-through.
Matt initiated a long-term position live on the call, sizing into the 258–268 FTD pivot range with target 278. The earnings print confirmed the AI demand story (FCF + AWS growth); the trigger is the technical pivot, not the headline. Stop is defined below the entry range.
Third leg of the long-term portfolio, added on the reclaim of the 20-day moving average. Earnings already confirmed the AI demand thesis. Matt called the trigger live on the call and pulled the trigger on the bounce.
Long-term position initiated 30 Jul at $472.495, stop $415. Still working — Matt called out "hitting new highs" on today's call. This is the Mac 7 anchor; AMZN and GOOGL were added as legs 2 and 3 of the same portfolio construction arc.
"This is the main beneficiary of all this, currently lagging." Matt defined a small, risk-bounded entry with 4% downside and a 26-day countdown to earnings — enough time to work the level without forcing a decision. The AI tape has to confirm before NVDA turns; it's a coin flip until the print.
The complete trade: 9.11 pivot break → 8-day bounce → halted upside at 14.40 / 15.62 → next targets 16.50 / 18.48 → stopped on the 8-day break. Matt narrated the stop-out as the lesson, not the loss. The show-don't-tell of "take profits, move stops up" from this week's calls.
"Punished after earnings" — Matt still wants the Mac 7 slot but is waiting for a better setup. Not initiated today. The fourth-and-watch long-term name: needs a clean pullback / reclaim before it joins MSFT / AMZN / GOOGL in the portfolio.
Same CXMT supply-expansion story as 30 Jul. Hyperscaler earnings confirmed AI demand structurally, but supply is coming into the tape. No re-entry until the chart prints a new base — "one green candle doesn't mean anything" still applies.
Same structural supply story as MU — the memory cohort trades together. SNDK needs a new base before it stops being a falling knife. P/E alone is not the trade; chart structure still rules.
Day 2 of the rally attempt, breadth still narrow but the index is holding. AMZN / MSFT / GOOGL long-term positions initiated. Watchers want this to be the turn — and it might be, on Tuesday. The thesis is "AI demand confirmed by hyperscaler earnings, supply-side anxiety easing."
Thursday's $4.7B retail buy was the highest in a decade. Historical record (6,500 stocks, 1985–2024): 85% median max drawdown, 2.5y median recovery, over half never reclaim prior highs. The audience is buying exactly the setups the data says lose money. Matt's read: "The setup is the entire opposite of what the crowd is doing."
"I'd rather eat the loss than be rewarded for undisciplined trading. Whenever I tried to abandon the process, I just got punished every time. So I'm talking from experience." The Friday tape is day 2 of a relief rally, not a confirmed reversal. Confirmation needs an FTD — earliest Tuesday.
Use confirmation, partial size, defined stops, scaled exits. Don't buy the dip on the retail record-buy signal — base-rates say 85% median drawdown, 2.5y median recovery, half never recover. Move stops up on any 8-day-line-holding trade (the FCUV lesson).
The Follow-Through Day framework is still the gate. Until NASDAQ closes above yesterday's high on volume, every long is a relief bounce, not a reversal. Wait for the trigger. Don't be a hero on the candle.
Matt opened the Friday tape framing the rally as day 2 of an attempt, not a follow-through day. Breadth is staying flat — RSP lagging SPY — so the move is narrow and tech-driven, not broad. Confirmation needs an FTD, which doesn't print until Tuesday at the earliest.
Thursday's retail print was the highest in a decade — concentrated in NVDA, AMZN, S&P. Historical data on the 6,500-stock sample 1985–2024: median max drawdown 85%, median recovery 2.5 years, over half never reclaim prior highs. "The setup is the entire opposite of what the crowd is doing." The anti-FOMO argument with receipts.
Live on the call, Matt initiated a long-term position in AMZN on the FTD pivot: 258–268 entry range, target 278. Earnings confirmed AI demand (FCF + AWS growth). The trigger was the technical pivot, not the headline. "I'm getting a follow-through day here on Amazon, so I'm taking initiate my position here long-term on Amazon."
Third leg of the long-term portfolio, added on the reclaim of the 20-day moving average. "The third addition to my long-term will be Google here, bouncing off, there we go, on the reclaim here of the 20-day line." Earnings already confirmed the AI demand thesis.
"This is the main beneficiary of all this, currently lagging." Matt defined a small, risk-bounded entry: 4% downside, 26 days to earnings. Enough time to work the level without forcing a decision. The AI tape has to confirm before NVDA turns — it's a coin flip until the print.
The 30 Jul long-term initiation ($472.495 / $415 stop) is still working. Matt called out "hitting new highs" today. MSFT is the Mac 7 anchor; AMZN + GOOGL today make it a three-name long-term portfolio with explicit stop discipline on each leg.
9.11 pivot break → 8-day bounce → halted upside at 14.40 / 15.62 → next targets 16.50 / 18.48 → stopped on the 8-day break. Matt narrated the stop-out as the lesson, not the loss. The single best show-don't-tell clip of "take profits, move stops up" this week. Pair with the "I'd rather eat the loss than be rewarded for undisciplined trading" line.
"Punished after earnings" on memory cost pressure. Matt still wants the Mac 7 slot but is waiting for a better setup. "I really want to get an entry on Apple, and really then, it's going to be my Mac 7 here as well." Not initiated today — the fourth-and-watch long-term name.
Same CXMT supply-expansion story as 30 Jul. Hyperscaler earnings confirmed AI demand structurally, but supply is coming into the tape. MU + SNDK = no new base, no re-entry. "One green candle doesn't mean anything." The rule stays the rule.
Source plan: long-term position initiated at $472.495 with a $415 stop. The stop sits well below the structural level so the position can breathe through normal volatility.
Supply is coming into a weak tape. The cheap-P/E story is loudest right before the worst trades.