⚠️ Allocation Check: Day 34 Closes the Book on Day 7 — Kill-Switch Still Locked, Floor Still Cash
Day 34 closes reading exactly like Day 33, Day 32, and most of the week before it. Seven full sessions parked in 100% Cash, and across that stretch QQQ has done one thing: climb. Today it tacked on another record — $719.79 on the consolidated close, +0.58% past Wednesday's $714.71 settle — and the override didn't so much as twitch. The EMA 25 kill-switch is still armed in the above_7 zone, distance closed at +7.3%, and the floor holds at cash.
This is the part that grinds: the rubber band is exactly as stretched tonight as it was this morning. Distance opened the session at +7.19%, closed at +7.3% — call it flat, after a day where QQQ ranged a tidy $717.81 to $721.81 and settled in the middle. No compression. No reset. Just another vertical-ish session that kept the 25 EMA from catching up.
Closing state:
- Score: 4.69 (Thursday 4:09 PM ET close) · Extreme Risk-Off · no bear-stretch boost · drifted -0.05 from the 9:09 AM 4.74 open · 12th straight sub-4.95 reading
- QQQ: $719.79 consolidated close (fresh record) · intraday range $717.81–$721.81 · +0.58% on the day
- EMA 70: $638.30 · price +12.8% above · trend filter buried deep into override territory
- EMA 25 distance: +7.3% · zone above_7 (kill-switch armed) · flat on the day from this morning's +7.19%
- Final allocation: 100% Cash · unchanged since the May 6 11:08 AM trim · Day 7 of zero long exposure complete
- Re-entry trigger: distance below +5% · roughly QQQ $704 with the EMA 25 near $670.80 · spot closed ~$16 above the trigger
Another record, another shrug from the override. Seven days of cash, and the only thing that's changed is the price the model is watching from the sidelines.
📊 The Score's Intraday Chop: A Profit-Taking Engine With No Profits to Take
The score spent the whole session doing what it's wired to do, just in a vacuum. It opened at 4.74, and every time QQQ pushed toward the highs it subtracted: down to 4.68 at 11:18 AM with QQQ at $721.10, back to 4.73-4.74 by early afternoon as price faded toward $719, and down again to 4.68-4.69 into the close as QQQ drifted back over $720. Net drift on the day: 4.74 → 4.69. A six-bp chop, all of it deep in Extreme Risk-Off, none of it within shouting distance of the 0.07 rule.
That subtract-at-the-highs behavior is the price-level engine working — QQQ crossing above key levels makes the macro read register the tape as expensive relative to what the data supports. On a normal day, that's the score taking profits into strength. The problem today, and for the last seven sessions, is that there are no profits to take. The position is flat. The score is running its expensive-zone math on a portfolio that's entirely in cash.
For the record on the mechanics, since this is where it gets confused: the score isn't the thing holding cash. A 4.69 raw score is a 100% SQQQ reading. But QQQ is +12.8% above its 70 EMA, so the EMA 70 override stomps that to a 100% QQQ floor — then the EMA 25 stretch overlay sees price +7.3% above the 25 EMA, past the +7% kill-switch, and trims the floor all the way to cash. Three layers, and the bottom one has been calling the shot for a week. Until distance drops below +5% — that ~$704 line, a full reset, no half-steps — nothing about that changes.
📈 What the Tape Wanted: Chips, Cisco, and a Summit in Beijing
The macro data Thursday morning was, frankly, a little soft — and the market didn't care for a second. Retail sales came in +0.5% for April, a touch under the +0.6% consensus and a sharp cooldown from March's downwardly-revised +1.6%. Initial jobless claims ticked up 12K to 211K, above the 205K consensus and the prior week's 199K. Consumer's slowing, labor's loosening a hair. On a different week that's a story. This week it was a footnote.
Because the tape had bigger fish. The headline driver: Washington cleared Nvidia's H200 AI chips for sale to roughly ten major Chinese firms — Alibaba, Tencent, ByteDance, JD.com — each cleared to buy up to 75,000 units. Whether the chips actually ship is another question (Beijing's own guidance has Chinese firms hesitating), but the market traded the headline, not the nuance. On top of that, Cisco ripped ~15% after a Q3 beat — $15.8B revenue, +12% YoY, $2.1B in AI orders, full-year AI revenue guide bumped to $9B — even with a 4,000-job AI restructuring attached to it. And Trump landed in Beijing with a delegation of tech royalty (Huang, Musk, Cook) for a summit investors are reading as a trade-tension de-escalator.
Add it up and you get the Dow pushing through 50,000, the S&P notching another record, the Nasdaq grinding +0.27% higher, and QQQ closing at its fresh ATH. It was a narrow, headline-driven, AI-led tape — exactly the kind of melt-up the macro engine is built to be skeptical of, and exactly the kind that's been steamrolling that skepticism all week.
🎯 My Take: Seven Days of Cash and a Melt-Up the Model Sat Out
I'm not going to dress this up. The kill-switch trimmed to cash on May 6 with QQQ around $695.75. It's now $719.79. That's roughly $24, about 3.5%, of upside the model has watched go by from the sidelines — and it kept watching today. Seven sessions, a steady grind of records, and the override hasn't given an inch. If you're long, this week felt great and the model felt like a wet blanket. Fair.
But here's the case for the wet blanket, and it's a real one. The EMA 25 +7% kill-switch isn't a macro call — it's a blow-off-top filter, and it's backtested. Its entire job is to refuse to chase a market that's stretched this far above its own short-term trend, because the historical edge says the giveback from these distances is steeper than the last leg of the melt-up. The cost of that discipline is exactly what you're seeing: it's early, it leaves money on the table, it looks dumb while the rally runs. The payoff only shows up when the rubber band snaps — and at +7.3% distance, it's wound about as tight as this system ever lets it get.
The honest framing: this is the model being early, not the model being wrong. Early and wrong feel identical right up until they don't. I'd rather be sitting in cash with a re-entry plan than be the guy adding leverage seven days into a vertical move on AI headlines. But I'm also not going to pretend Day 7 of cash during a record run is comfortable. It isn't. It's the price of the rule.
⚠️ Bottom Line: The Only Exit Is Down or Sideways
Nothing about the override unlocks on green. A push to fresh records tomorrow just re-widens the stretch and digs the kill-switch in deeper. The two paths back to long exposure haven't changed: a ~2% QQQ pullback toward that ~$704 line, or a long enough sideways grind to let the 25 EMA climb into range. After a week of one-way tape, neither is knocking on the door.
Day 8 of cash starts Friday. The model has a re-entry plan and the patience to wait for it — the market has a melt-up and no interest in slowing down. One of them blinks eventually. The kill-switch is betting it's the tape.