MU: a trillion-dollar run, a 41% drop in five weeks, and three recipes that paid in cash
The Market Wave Algorithm for Equities turns Micron’s swings into realized profit in three windows, beating buy and hold and DCA each time.
Micron: the AI memory maker that has more than tripled this year, never in a straight line.
Why this case study, and why now
Micron has more than tripled this year, and for anyone who held, every dollar of that gain is still on paper. This study asks whether a rule can do the selling instead, and turn Micron’s swings into cash as they happen.
Three windows, three recipes of the Market Wave Algorithm for Equities:
The weekly paycheck
Strategy+44.2%
- Buy and hold
- +17.6%
- DCA
- +10.5%
The patient buyer
Strategy+45.9%
- Buy and hold
- +3.3%
- DCA
- +12.5%
The 2026 scoreboard
Strategy+461.7%
- Buy and hold
- +247.1%
- DCA
- +100.3%
Every run starts from the same hypothetical $100,000 in cash and zero shares.
Backtested results. Past performance does not guarantee future results. For educational purposes only.
Micron’s year in seven dates
Micron makes memory and storage chips for data centers, PCs, and phones, including the high-bandwidth memory that feeds AI processors. AI demand sold out its 2026 high-bandwidth memory and more than quadrupled its quarterly revenue in a year.
- Jan 2Up more than 10% on the year’s first trading day, as Bernstein raises its target.
- Mar 18Quarterly revenue nearly triples, to $23.86 billion. The stock falls 15% in four days.
- May 26Micron becomes a trillion-dollar company.
- Jun 4 to 5A fifth of its value disappears in two sessions, to $864.
- Jun 25A record high of $1,255, the day after revenue more than quadrupled to $41.46 billion.
- Jul 29A low near $739, 41% below the peak, as the memory bear market deepens.
- Sep 23Near $1,080, a week before earnings on September 30.
Here is the catch that matters for this study: Micron paid holders on paper, then took back 41% of the peak in five weeks. Only gains already sold were safe.
The investor’s dilemma
Micron is a nerve test, and nerve is not a strategy.
Taken back from the peak in five weeks
The paper gain trap
A January buyer had roughly quadrupled their money at the June 25 high. Five weeks later 41% of that peak value was gone, and none of the gain had been realized.
The too high freeze
Every entry looked like the top. Buyers who waited for a pullback got one on June 4 and 5, a fifth off in two days, hesitated, and watched a record three weeks later.
Narrative whiplash
On June 23 Micron fell 13.3% on a Korean regulator’s warning about memory ETFs. Two days later it hit a record on quadrupled revenue. Every headline explained the last move, never the next.
Dollar cost averaging deserves credit. It doubled this year without a timing decision. But it buys blind, never sells, and never realizes a dollar, so it rode the 41% drop with every share it had bought.
A rules-based algorithm restores what Micron takes away: a way to get paid while the stock swings.
The algorithm
Ride confirmed moves
It enters when price meets a preset rule, and keeps adding while the move holds. It never needs a headline.
Exit on rules, not nerves
It sells a set share of the position at a profit target. It sold Micron into the June run.
Step aside in chop
Without a signal, it waits. The patient recipe held cash for eight weeks, through the peak and the drop.
Each recipe sets three levers: scope (0.5 micro to 20 macro), thresholds, and trade size.
The weekly paycheck runs scope 1 and sells 83% of its shares per exit.
The patient buyer runs scope 7.5 and only buys below the wave.
The 2026 scoreboard runs scope 10 at full size.
Backtested results. Past performance does not guarantee future results. For educational purposes only.
“The weekly paycheck”
June 1 to September 18, 2026 · Weekly chart · Micro recipe, sells 83% of its shares on each exit
+44.2%
- Buy and hold +17.6%, outperformance +26.6 percentage points
- DCA +10.5%, outperformance +33.7 percentage points
- Ending value
- $144,231
- Realized profit
- $37,906
- Unrealized profit
- $6,325
- Shares held
- 133 at $968.25 avg cost
- Cash
- $9,129
Open full-size chart
Backtested results on the weekly chart, valued at the September 18, 2026 weekly close. Past performance does not guarantee future results. For educational purposes only.
“The patient buyer”
June 1 to September 23, 2026 · Four hour chart · Buy-below recipe, scope 7.5
+45.9%
- Buy and hold +3.3%, outperformance +42.6 percentage points
- DCA +12.5%, outperformance +33.4 percentage points
- Ending value
- $145,932
- Realized profit
- $28,531
- Unrealized profit
- $17,401
- Shares held
- 99 at $900.23 avg cost
- Cash
- $39,408
Open full-size chart
Backtested results on the four hour chart, valued September 23, 2026. Past performance does not guarantee future results. For educational purposes only.
“The 2026 scoreboard”
January 1 to September 23, 2026 · Four hour chart · Full-size recipe, scope 10
+461.7%
- Buy and hold +247.1%, outperformance +214.6 percentage points
- DCA +100.3%, outperformance +361.4 percentage points
- Ending value
- $561,731
- Realized profit
- $416,560
- Unrealized profit
- $45,171
- Shares held
- 519 at $994.37 avg cost
- Cash
- $485
Open full-size chart
Backtested results on the four hour chart, valued September 23, 2026. Past performance does not guarantee future results. For educational purposes only.
Side by side
Read the realized row first. Buy and hold and DCA realized nothing in any window. Every dollar of their return is still on paper.
| The weekly paycheck | The patient buyer | The 2026 scoreboard | |
|---|---|---|---|
| Window | Jun 1 to Sep 18, 2026 | Jun 1 to Sep 23, 2026 | Jan 1 to Sep 23, 2026 |
| Chart | Weekly | Four hour | Four hour |
| Recipe | Scope 1, 93% in, 83% out | Scope 7.5, buy below the wave | Scope 10, 100% in, 90% out |
| Strategy return | +44.2% | +45.9% | +461.7% |
| Buy and hold | +17.6% | +3.3% | +247.1% |
| Outperformance | +26.6 pts | +42.6 pts | +214.6 pts |
| DCA | +10.5% | +12.5% | +100.3% |
| Outperformance vs DCA | +33.7 pts | +33.4 pts | +361.4 pts |
| Ending value | $144,231 | $145,932 | $561,731 |
| Realized profit | $37,906 | $28,531 | $416,560 |
| Share of gain realized | 85.7% | 62.1% | 90.2% |
| Shares held | 133 | 99 | 519 |
| Avg cost basis | $968.25 | $900.23 | $994.37 |
| Cash | $9,129 | $39,408 | $485 |
All three start from a hypothetical $100,000 in cash and zero shares.
Backtested results. Weekly window valued at the September 18, 2026 weekly close, four hour windows valued September 23, 2026. Past performance does not guarantee future results. For educational purposes only.
Why the results differ
One algorithm, three questions.
Can a volatile stock pay you as it swings?
The weekly recipe sold most of its shares into each rally and bought back each drop, realizing $37,906 in one summer.
Can you skip the top without predicting it?
The patient recipe waited eight weeks for price to reach the wave, then bought the drop.
What did this year actually pay?
The 2026 recipe realized $416,560 on $100,000.
Buy and hold decides once and waits. DCA decides once and repeats.
The edge is not prediction. None of these recipes knew June 25 was the top.
Backtested results. Past performance does not guarantee future results. For educational purposes only.
The real lesson
- 01
Volatility can be income, if something sells
Micron swung double digits again and again this summer. Holding turned each swing into stress. Selling by rule turned them into $37,906 of realized profit.
- 02
Selling is the only way to keep a gain
Micron gave back 41% of its peak in five weeks, and holders kept none of the top. Every recipe here sold into strength by rule.
- 03
Cash is a position, not a missed trade
The patient recipe held cash for eight weeks, through the record and the 41% drop, then bought near the low. It still beat buy and hold by 42.6 percentage points.
The client decision is which job you want done: harvest every swing, wait for the big drop, or compound a year of profit-taking. The Market Wave Algorithm for Equities has a recipe for each.
Backtested results. Past performance does not guarantee future results. For educational purposes only.
A path forward
Arch Public runs the Market Wave Algorithm for Equities in your own account.
Zero custody. Full control.
Rules, not reactions.
What do you want your account doing the next time a stock drops 41% in five weeks?
On paper, a January buy and holder is up
247.1%
The same $100,000 under the 2026 recipe is worth
$561,731
with $416,560 realized.