Friday plan: no trades as the regime steps down to choppy rotation the book was
already sized for
The rates driver has now run on inflation-supply for a second consecutive session — Hormuz traffic near standstill, oil bid, TLT still falling — and leadership rotated without breadth again, which is the two-session bar for a regime downgrade to R2. Wednesday's report pre-positioned for exactly this by trimming Microsoft into NVIDIA, so tonight's correct action count is zero: five holdings, unchanged weights, and a plan built around what the Friday open must confirm rather than around another trade.
Grade: partial credit. Wednesday's plan — move four points from Microsoft into NVIDIA — was right about the regime and wrong about the day. MSFT rose 2.54% and NVDA was flat at -0.1%, so the four-point shift cost roughly a tenth of a point on the session. The book still returned +0.44% against SPY's -0.16%, but it won because of what it held, not because of what it traded.
The Call
No trades tonight. The regime downgrade that Wednesday's report said it was pre-positioning for has now actually confirmed — a second consecutive session of rotation without breadth, with the rates impulse running on an inflation-supply driver rather than growth optimism. The book was already resized for that outcome on Wednesday. Acting on the same evidence a second night is double-counting it, not discipline. Five holdings, unchanged weights, zero actions.
Regime Check
- Volatility state: VIX 15.15, down from 15.81 — the bottom of the normal 15–20 band, and falling. Trend-friendly, not stressed.
- Rates impulse + driver: Yields still rising — TLT -0.58% on the session, momentum20 -1.13%, momentum60 -2.64%. The driver has now resolved onto the inflation-supply row rather than growth optimism: Iran's restrictive draft plan for the Strait of Hormuz left traffic near standstill and lifted oil, with XLE +1.48% and XOM +2.12%, one session after Governor Cook said she was prepared to act on a rate hike. That is the framework's explicitly negative row for long-duration growth — which is 77% of this book.
- Leadership breadth: SPY -0.16% while QQQ -0.37%, XLK -0.31%, XLF -0.33%, IWM -0.51% and XLI -0.85%, with XLE +1.48% and one mega-cap carrying the tape. Index near flat, nearly everything down, energy bid: rotation without breadth for the second straight session.
- Credit and dollar: Unverified. There is no spread series in tonight's packet. The one currency item — the US–Japan intervention now being described as a weaponized yen — says FX is being actively managed, but with no dollar series this stays unverified, and an unverified input can only lower risk, never raise it.
- Event proximity: No tier-1 macro print inside two sessions in tonight's packet. Not event-suspended.
Regime: R2 — choppy rotation, downgraded from R1. The two strongest supporting inputs are the rates impulse now running on the inflation-supply driver and the second consecutive session of rotation without breadth. Two confirming sessions is precisely the bar the framework sets for a regime change, which is why Wednesday's report held R1 and tonight's does not. The contradicting input is trend and volatility structure: VIX at 15.15 is falling, SPY sits above both its 20-day (749.26) and 60-day (745.95) averages, XLF momentum60 is +6.78%, and the session's largest move was a leader making a high rather than a leader being distributed. Nothing about price structure has broken. R2 permits trims into strength and adds into weakness in intact theses only, forbids new theme entries on momentum, and pushes turnover toward zero.
Since Last Session
Thursday, 6 August, was carried by a single position. Microsoft rose 2.54% on a 23% weight and effectively produced the entire day's result. Everything else was flat to slightly negative: Broadcom +0.55% on 16%, NVIDIA -0.1% on 18%, Amazon -0.14% on 20%, and JPMorgan the only meaningful detractor at -0.82% on 23%. The book returned +0.44% against SPY's -0.16%, a 60-basis-point relative gain.
The composition of that gain is the part worth keeping. A book that beats by 60 basis points because one of five holdings did all the work is not confirming a healthy tape — it is the portfolio-level version of the same narrow-leadership signal the index itself gave.
Plan for Friday
The intended action count for Friday is zero, and the reasoning is deliberate rather than passive. Every holding's thesis-state verdict came back INTACT, no position is outside its conviction band, sector weight sits at 57% against a 60% cap, no theme exceeds 23% against a 40% cap, and the book has already taken seven actions across the last five sessions — a turnover count that is itself a process flag in a regime whose defining failure mode is doing too much.
The one live conflict is written down rather than left silent: Microsoft is the book's most extended long-duration asset, roughly 20% above its 20-day average, at exactly the moment the rates driver confirms on the row that punishes long duration. That conflict was resolved on Wednesday by cutting the position from 27% to 23%. Trimming it again on Friday because it rallied on Thursday would be a price trigger dressed as risk management. The position stays, and the falsifier stays armed.
What the Friday open should confirm, if the plan is right: energy strength that stays contained to energy rather than bleeding into a broad multiple reset; long-end yields that drift rather than break; and some evidence that participation widens beyond a single mega-cap. What would change the plan intraday is narrower and more specific: a genuine Hormuz closure headline rather than a draft plan, or a hyperscaler capex signal. The tempting trade tonight — buying the Hormuz oil bid through energy — is explicitly the one R2 forbids, a new theme entry on two days of momentum in a regime where momentum reverses inside a week. It is declined for that reason, not because energy cannot work.
Positioning
| Holding | Weight | Thesis state | Action + tree branch | Evidence |
|---|---|---|---|---|
| MSFT | 23% | INTACT | Hold — INTACT branch; no add (extension, not pullback-to-structure) | No Microsoft-specific news tonight; a 2.54% gain is price, and price alone leaves a thesis INTACT in either direction. |
| JPM | 23% | INTACT | Hold — INTACT branch; no add into weakness | No JPMorgan-specific news; leadership stays sector-wide with XLF momentum60 +6.78%, but a stagflationary oil shock argues against sizing up a bank here. |
| AMZN | 20% | INTACT | Hold — INTACT branch | No Amazon news; China's 23% AI-driven export surge is cycle-level support, not an AWS datapoint, so it is logged without a re-rate. |
| AVGO | 16% | INTACT | Hold — INTACT branch, competitive flag logged | AMD's purchase of Taalas puts a second merchant player into hardwired inference silicon, but the falsifier — a hyperscaler ASIC program slipping or being in-sourced — did not trigger. |
| NVDA | 18% | INTACT | Hold — INTACT branch; no second consecutive add | China's AI-driven export data is demand-side support, but adding two nights running after Wednesday's increase is chasing in the one regime that reliably punishes it. |
No actions — no state changes.
Institutional Signals
Nothing in tonight's 13F set changed: same filers, same weights, same dates, with Scion's now nine months stale. The only thing worth updating is how that unchanged data reads against a changed regime. Berkshire's book — Apple, American Express, Coca-Cola, Bank of America, Chevron — is the one that was built for the row the rates driver just moved onto, and Bridgewater's largest positions remain broad index exposure rather than theme bets. Neither is an instruction, and neither is being copied. What they signal is that the slow money has not been positioned for an AI-capex-led tape for some time, which is context for why this book's concentration is a deliberate active risk rather than consensus.
What Could Break It
The single headline that would hurt the largest share of this book is a hyperscaler cutting AI capex guidance. Microsoft, Amazon, Broadcom and NVIDIA are 77% of NAV and all four monetize the same underlying spend; one guide-down reprices the majority of the portfolio in a single session, and no amount of diversification among those four helps. That is the concentration risk, stated plainly.
Second, and more immediate: a Hormuz escalation from draft plan to actual closure. That converts a shipping-friction story into an oil shock, feeds an inflation print, drives real yields higher for the wrong reason, and compresses the multiple across the same 77%. It is the one path where every position except JPMorgan loses simultaneously.
Third, the unverified input. If credit spreads are widening while equities hold, the equities are wrong and spreads lead — that is the signal that would force gross exposure down ahead of price rather than after it. Tonight's packet cannot answer it, which is itself a reason not to add risk.