desk: house-delta date: 2026-08-10 forecasts: delta: p_trade: 0.30 direction: long conviction: low
Briefing reference: /Users/dadbot/Desktop/ClaudeBod/projects/dadbrain/Analysis/briefings/2026-08-10-pm.md PM: DELTA
Watchlist/briefing coverage check (this desk's instruments): the briefing captured NVDA (223.34) and AMD (479.47) premarket prints but stripped no ASML/TSM levels — filled via web search (Friday closes ASML 1,740.99 / TSM 420.04, both +Friday, plus premarket prints) and reconciled above. The 10:26 ET watchlist is after-open — S&P -0.08% / Nasdaq -0.04% at the open, INTC -4–5%, oil +1.4% — treated as early-session context, not pre-open pre-commitment. Note on dispatch: the morning plan job ran late; the 09:30–09:55 measurement window has elapsed at filing time. The forecast below stands as the day's forecast for scoring (per the late-dispatch convention), with the after-open tape as context.
Forecast: p_trade 0.30 | direction long | conviction low
Verdict: No clean setup is the base case, but the setup is genuinely live today — the single clean path (NVDA→ASML) has its best structural backdrop in a week, and the tape has fired 3 of the last 4 measured sessions. The constructive NVDA weekend headlines (The Information: NVDA plans up to $3B Lancium investment to expand its Stargate role; Huang in Seoul: "buy at a discount" during the recent AI sell-off; no countervailing weekend AI story) are exactly the input class that powered 08-03's clean winner (NVDA +2.32% first hour → ASML LONG +$184.68), and the briefing now places chips in the constructive tier of the two-tier tape — the first time in days the semis complex isn't the morning's marquee risk. What's missing is the same-day catalyst: on a quiet, flat-futures Monday with no print to anchor the tape, the base case is that first-hour leader moves stay sub-1.5% and nothing fires. Comparison day is 08-04 (double near-miss: AMD +1.04%, NVDA -0.71% — no trade), not the fired days 08-03/08-05/08-06 — but the weekend AI headlines are the delta that makes 0.30 (not 0.20) the honest number.
Why the clean path (NVDA→ASML) is the day's expression — and its one new wrinkle: NVDA earnings Aug 26 is 2+ weeks out (no contamination); the Lancium/Seoul headlines are unambiguously constructive for the AI-capex complex, and the two-tier tape (memory/software weak, chips constructive) supports a NVDA first-hour extension rather than fighting it — the structural backdrop the 08-05/08-06 cascade days lacked. The wrinkle: ASML's premarket prints are ~+2.9% (≈1,790.99). If that gap holds at the cash open, the laggard has already repriced ahead of the leader — the information propagated through the European session, the lead-lag spread this desk exists to capture is partially closed by construction, and a 10:00 entry into ASML would be a late entry at a compressed edge. A laggard leading is not this desk's signal shape. So even the clean path, if it fires, carries compressed edge today unless the ASML premarket gap fades at the open. Confirm the cash-open gaps on all four names; the mechanical runner still executes if NVDA clears +1.5% (plans are non-binding by pre-registered rule — I do not hand-gate a mechanical signal).
The degraded path (AMD→TSM): AMD's first-hour bounce pattern has fired 3 of the last 4 measured sessions (08-05 +2.15% → $183.57 five-minute loss; 08-06 +2.52% → $59.13 hard-flat win; 08-04 +1.04% near-miss) — the AMD_TSM sub-book sits at n=6, WR 33.3%, PF 0.272 (the weakest cell in the book; pre-registered sub-book kill switch not triggered, and it governs — not a plan-level gate). If AMD fires +1.5% today, the mechanical runner takes the AMD→TSM LONG, and it is logged and scrutinized as the same shock-reversal class, not celebrated. TSM's premarket is flat (~419.9) — unlike ASML, the TSM lag is intact, so an AMD fire is the cleaner expression of the two if it happens. Note the morning drag: INTC's -4% offering gap and the residual AMD post-earnings semi-sector drag are the forces that could keep AMD sub-threshold today.
Calibration math (day-conditional): P(signal) ≈ 0.45 — a quiet Monday's default is low first-hour impulse, but (i) chips are in the constructive tier for the first time in days, (ii) the constructive NVDA weekend headlines are a live first-hour impulse source (the 08-03 template), (iii) this tape has fired 3 of the last 4 measured sessions and 08-04 was a double near-miss — it gets close to 1.5% frequently, and (iv) VIX 15.15 is DELTA's best regime. Against: flat futures, no same-day catalyst, INTC offering drag, CPI-eve drift. P(LONG | signal) ≈ 0.65 — the last four signals have all been UP, the weekend headlines skew NVDA up, and the down paths (-1.5% ≈ NVDA 220.6 / AMD 476.1) are disabled SHORTs by construction. Net tradable-signal probability ≈ 0.30. Direction long by construction (LONG-only book), up-skewed by the headlines. Conviction low: the setup is live but the missing catalyst is the difference between 08-03's fire and 08-04's near-miss, and I do not anchor on the recent hot streak to inflate the number.
Sizing vs. event risk: standard 0.75%, single position — no change. The HIGH rating is CPI-in-two-days (Bucket A convention) + Hormuz headline risk, none of which this desk carries: the position hard-flats at 15:30 (Monday's position never survives into Wednesday's CPI by construction), the traded laggards (ASML/TSM) don't report today, and VIX 15.15 is far from the ≥ 30 barrier. The event risk that matters is signal quality (compressed ASML lag, degraded AMD class) — handled by scrutiny and invalidation, not a size cut. Book context: net −$182.42 on 8 trades (4W/4L, PF ≈ 0.70); SPRT CONTINUE (LLR +0.398, mid-band — +2.546 to CONSISTENT-WITH-BACKTEST, −3.343 to DEGRADED); AMD_TSM sub-book n=6 / PF 0.272 (not triggered); SHORT shadow book n=5; 30-LONG gate n=8/30 (a NVDA→ASML fire today is the highest-quality LONG available — that pair is 2/2 live).
Key levels (spread triggers, from Friday closes): NVDA +1.5% ≈ 227.3 → clean NVDA→ASML LONG (compressed-edge caveat if ASML's premarket +2.9% holds); NVDA -1.5% ≈ 220.6 → disabled SHORT (shadow n=6). AMD +1.5% ≈ 490.6 → degraded AMD→TSM LONG; AMD -1.5% ≈ 476.1 → disabled SHORT (shadow n=6/7). Laggard gaps at the open: ASML gapping up big = lag partially closed (late entry); TSM gapping with AMD = spread zero; a laggard not gapping with its leader is the only genuine lead-lag signal shape left today. Index: cash-open gaps < 1.5% keep the window live; QQQ holding the 722.89 prior-close area through the first hour matters for the two-tier thesis.
Invalidation (stand down entirely):
Information cost (documented, not acted on): if either leader cascades (-1.5%: NVDA ≈ 220.6, AMD ≈ 476.1), the disabled SHORT is logged to the shadow book (n=5 → n=6; n=7 if both). LONG-only keeps costing information in down-tape; the shadow book is the accounting, the 30-LONG gate (n=8/30) is the governing re-enablement path, and the 07-24 learnings observation (the LONG gate may be structurally unreachable in a sustained semis bear — shadow book as the alternate evidence path) stays live.
No trades taken.