Our Score Says Strong Buy on AMD, Micron and Palantir. Our Own Archive Says Be Patient.

6 Nov 2026 · MODEL AUDIT
Our Score Says Strong Buy on AMD, Micron and Palantir. Our Own Archive Says Be Patient.

Updated November 6 with the November 5 close and the October jobs report.

On November 4 our board was full of AI names scored 98: AMD, Micron, Palantir, Cloudflare, Snowflake, Dell. That's what our score is built to do. It rewards stocks that have pulled back hard below their 20-day average.

So we checked how that exact setup has worked in our own archive, and the answer isn't flattering. The next day made the point for us (see the update below).

The test

We took every daily snapshot from June 29 to November 5, 2026 and picked out stocks that were:

  • scored 90 or higher (Strong Buy), and
  • down 10% or more over the previous 5 sessions.

Then we checked what each one did over the next 5 and 10 sessions, and compared it with two groups: Strong Buys after a calmer week (down less than 5%), and every stock we track.

GroupNext 5 sessionsUp after 5Next 10 sessionsWindows graded
Strong Buy after a 10%+ weekly drop−2.99%32%−6.35%174 / 151
Strong Buy, calm week−0.32%55%−0.37%424 / 218
Every tracked stock+0.04%53%+0.56%6,689 / 5,184

Against SPY over the same 5-session windows, the crash bucket trailed by 2.59 points on average. Since November 5, even Strong Buys after a calm week are slightly negative on average.

Now the important caveat

Almost all of those crash-bucket readings are from this autumn:

PeriodWindowsAvg next 5 sessionsUp after 5
Jun 29 – Aug 2117+6.87%65%
Oct 12 – Nov 5157−4.05%29%

In the summer, the few stocks that got a Strong Buy after a sharp drop bounced well. Since October, falling stocks have kept falling. That matches what we found in October's score tiers, where our highest scores did the worst.

So the fair conclusion isn't "the signal is broken." It's narrower: in a market that rewards momentum, a Strong Buy right after a crash has been early, often by more than a week. Seventeen summer windows is far too few to say the bounce pattern will come back on schedule.

Update: what those 22 stocks did the very next day

At the November 4 close, 22 stocks were scored 90+ after a 10%+ five-session drop: AMD, Micron, Palantir, Cloudflare, Snowflake, Cerebras, Dell, Elastic, Twilio, Intel, Natera, Arm, ASML, Applied Materials, Lam Research, SanDisk, Lumentum, Marvell, Datadog, Super Micro, CoreWeave, Reddit.

On November 5, all 22 fell. They lost 11.8% on average, against −4.4% for SPY and −5.5% for the average stock we track.

Biggest one-day dropsNov 5
SanDisk−22.3%
Lumentum−18.7%
Micron−17.5%
AMD−17.3%
Marvell−16.4%
Lam Research−16.4%
Dell−14.9%

The mildest was Twilio at −5.9%. One day proves nothing about the long run, and some of these may well be higher in a few months. But it's exactly the pattern the table above describes, and we said so the day before. After the drop, 38 stocks now sit in the same bucket.

Most of them are AI hardware or AI software, the exact group the market is repricing right now (why).

How we'd use this

  • A Strong Buy is a level, not a timing call. Our score says the price is far below its recent average. It doesn't know about Alphabet's cash flow or CoreWeave's credit spreads.
  • Wait for the falling to stop. In this tape, a name that holds its low for a few sessions has done better in our data than one bought on the way down. The falling-knife flag on the Rankings page exists for this reason.
  • Size smaller. If you buy a 98 that just dropped 14% in a week, assume the next week can look the same.
  • Check our Receipts. We log every Strong Buy and grade it later, wins and losses alike.

Data: ZonedInOut score-history archive, Jun 29 – Nov 5, 2026, with a gap from Aug 22 to Oct 11. Windows overlap, so they aren't independent samples. Not investment advice.

Scores in this piece come from the ZonedInOut model. How the score works →

Frequently asked

Is AMD a buy after the November 2026 selloff?
Our score rated AMD 98 (Strong Buy) at the Nov 4, 2026 close after a 13.7% five-session drop; it fell another 17.3% on Nov 5. In our archive, stocks scored 90+ right after a 10%+ weekly drop averaged −2.99% over the next 5 sessions (32% higher), almost entirely because of October–November readings. The score measures how far price is below its average; it does not time the bottom.
Does buying the dip work in a correction?
It depends on the regime. In our data, Strong Buys after a sharp drop averaged +6.87% over 5 sessions in the summer of 2026 (17 cases) but −4.05% from Oct 12 to Nov 5 (157 cases), when momentum dominated. All 22 names in that bucket on Nov 4 fell the next day, by 11.8% on average.
Why does ZonedInOut show Strong Buy on falling stocks?
The score rewards distance below the 20-day average, adjusted for volatility and trend, so a sharp drop raises it. It is designed to flag discounts, not to predict when a decline ends. A separate falling-knife flag marks high scores on stocks still trending down.

Get the Sunday Brief

One email a week, before the open: what the board did, which calls worked, which didn't, and what changed. Free. No spam, unsubscribe anytime.