The Nightcap White-Cap

Built ~5:15 PM ET · 2026-07-01 · validates the 7/1 regular-session close (Massive /v2 grouped daily + /v3 snapshot, COALESCE(session_close,session_price)) anchor: system-reminder 2026-07-01 == Bash date (agree) · static — regenerate to refresh

01Session Scorecard — 2026-07-01

9 setups in the open backlog; 6 finalized this run — 4 on today’s 7/1 session (Curd + Midday) plus a 2-setup 6/30 Midday backfill (missed on the 6/30 Nightcap) — and 3 weekly Sunday Sundae setups left interim (window active through Thu 7/2 June-NFP, ahead of the Fri 7/3 Independence Day close). Tally: 2 FIRE / 3 VOID / 1 MIXED / 0 NO_EVIDENCE. Today was a sharp semis/AI-hardware unwind (SMH −5.40%, MU −10.57%, KLAC −11.77%, ASML −7.36%) that funded a rotation into megacap-software (META +8.81%, MSFT +3.02%), financials (XLF +2.18%) and equal-weight — RSP +0.30% > SPY −0.14% > QQQ −1.52%, 7 of 11 SPDRs green. The semis-rotation short (SRT) FIRED; the two continuation longs into semis correctly stood aside (never armed); the rotation-long into small-caps MIXED (tagged target intraday, faded into the close).

Setup	Outcome	Evidence	Δ-ATR	Actual
SRT S SMH,QQQmd-260701-1209-SRT-SMH-S	FIRE	Sector-rotation-top short in the crowded QTD-leader semis. QQQ failed its 729.52 VWAP reclaim (high 731.92 didn’t hold) and closed 725.17 — below the 727.64 target; SMH −5.40% (620.46, below 632), MU −10.57% with no oversold snapback and no semis-bid return. No kill. Levels are QQQ-denominated (primary inst SMH), so Δ-ATR / realized-R are n/a by construction — scored on the QQQ close + semis confirmation.	—	QQQ 725.17
MS L QQQ,SMHmm-260701-MS-SEMI-L	VOID	Momentum-continuation long, armed only on a fresh opening-range reclaim of 737.62. QQQ gapped down (open 729.19, below the trigger), high just 731.92 — never reclaimed — and closed 725.17 below 730; the semis complex unwound (SMH −5.40% lost its opening range, MU −10.57%, TSM −6.98%). Kills fired (QQQ lost 730; semis rolled red; IWM lost 300, closed 299.32). Long never armed — the “NOT a chase after +4% in two sessions” discipline sidestepped the unwind.	0.72	QQQ 725.17
LRT S QQQ,SPYmm-260701-LRT-QQQ-S	VOID	Conditional tag-and-reject short, armed only on a tag of the 740–745.45 shelf. QQQ opened weak (729) and fell on the semis-only unwind — high 731.92, never reached the shelf. And the modeled mechanism (lagging megacaps + equal-weight fail to confirm) resolved opposite: equal-weight was green (RSP +0.30%) and megacaps led (META +8.81%, MSFT +3.02%). Right index direction (QQQ −1.52%), wrong mechanism, no trigger — VOID by the never-armed convention.	1.17	QQQ 725.17
MS L IWM,RSP,XLFmd-260701-1209-MS-IWM-L	MIXED	Broadening-rotation long in the cohort absorbing the semis unwind. IWM tagged its 301.86 target intraday (high 302.72) — the right leg early — but the “hold into power hour” thesis broke: IWM reversed to close 299.32, below the 301.19 entry and 300.74 stop, as RSP lost its 213.70 VWAP and SPY lost 746.00 VWAP (2 of 3 kills fired into the close). Target hit, then thesis failed — MIXED.	0.35	IWM 299.32

Backfilled — 6/30 Midday (window closed at the 6/30 bell; missed on the 6/30 Nightcap, self-healed this run)

Setup	Outcome	Evidence	Δ-ATR	Actual
MS L QQQ,SMHmd-260630-1209-MS-SEMI-L	FIRE	Midday momentum-continuation long into the quarter-end markup. QQQ broke the 734.7 midday high, tagged the 737.62 session high and closed 736.40 — above the 736.13 target; SMH +3.19% held above 647. No kill (QQQ held its 730 VWAP, leaders closed near the highs).	0.10	QQQ 736.40
LRT S QQQ,SPYmd-260630-1209-LRT-QQQ-S	VOID	Conditional midday exhaustion-fade short, entry 738. QQQ high 737.62 stalled 0.4pt shy of the entry and never armed; it held near the highs into the close on a rising VWAP (733.82) — no rejection (the “no-rejection / VWAP-rising” kill triggered). Quarter-end markup; short correctly stood aside.	0.09	QQQ 736.40

Still open — interim (weekly Sunday Sundae 6/28, window through Thu 7/2 June-NFP; not finalized)

Setup	Interim	Status as of the 7/1 close	Δ-ATR	Actual
OMR L QQQsun-260628-OMR-QQQ-L	OPEN	Working. Oversold mean-reversion long, entry ref 716 — QQQ closed 725.17, ~+1.3% above the trigger despite today’s semis fade; no kill (QQQ<702 not hit; session low 724.6). Finalize at the Thu 7/2 NFP gate.	—	QQQ 725.17
UCR L SPYsun-260628-UCR-SPY-L	OPEN	Working. Undercut-and-reclaim long, entry ref 737 — SPY closed 745.76, ~+1.2% above and held >716.58; no kill. Today’s broadening rotation supports the broad-index long even as QQQ lagged. NFP Thu gate.	—	SPY 745.76
EXF S XLVsun-260628-EXF-XLV-S	OPEN	Leaning against (WATCH). Exhaustion-fade of stretched defensives — XLV closed 159.54 (+0.55%), catching a small bid on the semis-risk-off tick; the fade is not confirming (defensives firm, not rolling). Low-conviction, no exec; no kill decisively hit. NFP Thu gate.	—	XLV 159.54

LensA clean dispersion day for the book: the semis short fired, and the two continuation longs into semis correctly never armed — the MS-SEMI long’s “no-chase” discipline sidestepped a −5% to −12% single-day unwind (capital protected is the win here, not a hit). The one blemish is the IWM rotation-long MIXED: right that money would rotate to small-caps/equal-weight, but it faded the tag into a re-narrowing close. The 6/30 backfill closed cleanly (1 FIRE / 1 VOID). Weekly Sundae longs still working (QQQ/SPY above entries, unkilled); the XLV fade is the one leg leaning offside — all resolve at Thursday’s NFP.

02Calibration

0.2477 Rolling Brier (last 50) up from 0.2418 (6/30) · 0.25 = random

42% Hit rate (rolling 50) 41.0% over all 83 eligible

83 Calibration-eligible 6 finalized this run · 0 NO_EVIDENCE · 3 weekly interim

Rolling Brier ticked up (worse) 0.2418 → 0.2477 — still fractionally under the 0.25 random line, but today’s batch hit only 2 of 6, and the one clean win was under-confident: the SRT-SEMI short fired at prob 0.40 (Brier 0.36 — a correct call the model wasn’t sure enough about), while the MS-IWM continuation long missed at prob 0.55 (Brier 0.30 — over-confident into a fade). No data-feed gaps this run: every finalized leg scored from confirmed Massive closes; no leg was gated on $S5FI/$S5TH/AAII (the breadth kills referenced IWM/RSP/SPY VWAPs — entitled ETFs), so the breadth read used confirmed price action, never inference.

Pattern performance — Component 13 (calibration-eligible; n≥3)

Pattern · Dir	n	Hit%	Mean prob	Mean Brier	Evidence class
LRT · S	25	52%	0.37	0.323	lore_pending
MS · L	22	36%	0.51	0.254	lore_pending
SRB · L	7	71%	0.41	0.299	lore_pending
SBD · L	6	0%	0.37	0.138	primary_source
GFD · L	5	60%	0.39	0.257	practitioner_backtest
MS · S	4	75%	0.44	0.318	lore_pending
SBD · S	4	0%	0.28	0.079	primary_source
LRB · L	3	33%	0.43	0.216	lore_pending

Plus six thin cells (SRT-S now n=2 at 50% after tonight’s FIRE, and n=1 singletons FBR-S, GFU-S, VAB-L, VBR-L, VSR-L) — too few to read. The two workhorses both softened tonight: LRT-S 56% → 52% over 25 (two more conditional shorts that voided rather than fired — the levels didn’t set up), and MS-L 37% → 36% over 22 (one FIRE, one never-armed VOID, one MIXED). SRB-L (71% / 7) stays the cleanest high-n edge; the 0%-hit SBD cells (primary-source) remain the standing red flag.

Calibration trend — Component 14 (rolling Brier by validation date)

06-230.272

06-240.261

06-250.2435

06-260.2411

06-290.2399

06-300.2418

07-010.2477

Mechanism leak · C15 11/36 right-direction / wrong-mechanism (rate 0.306, up from 0.300), still led by LRT-S (6). Both of tonight’s LRT shorts are counted: each closed below its short level (mechanically “right direction”) but neither fired on its designed tag-and-reject mechanic — the recurring LRT signature. Caveat: the metric keys on close-vs-level, so it flags the 6/30 stand-aside (QQQ actually rose) alongside the 7/1 never-armed VOID.

LensA flat-to-slightly-worse night, not a giveback — Brier ~0.248, still under the 0.25 random line. The recurring calibration signal is unchanged and worth internalizing: the model is under-confident on its level / rotation-short work (SRT fired at 0.40; the level shorts keep landing right-direction) and over-confident on continuation longs carried into the close / event risk (the 0.55 IWM long faded). And the LRT-S remains the repeat right-direction/wrong-mechanism leak — the edge is fading failed reclaims, so it correctly voids when a level never sets up, but the pattern’s design still leaves R on the table.

02bMacro-Prediction Calibration · thinktank-v2 predictions, read-only

Macro-prediction calibration unavailable (thinktank-v2 absent — resolve_predictions returned “unavailable (thinktank-v2 absent)”). This is a separate calibration lens from the setup track-record above — an external macro/cross-asset forecast log, never combined with the setup Brier/hit-rate — and its absence does not affect the validation core. Nothing scored or written (DB read-only, not mutated).

03Tape & Rate Backdrop

Index	Close	Day %	Read
S&P 500 (SPY)	745.76	−0.14%	~flat — rotation offset the semis drag
Nasdaq 100 (QQQ)	725.17	−1.52%	semis-heavy — the day’s laggard
Equal-weight (RSP)	213.41	+0.30%	> SPY > QQQ — broadening
Russell 2000 (IWM)	299.32	−0.38%	green midday (hi 302.72), faded late
Dow (DIA)	522.40	0.00%	flat — value/cyclical ballast

Dominant driver = a semis/AI-hardware positioning unwind, not a macro shock. The QTD leaders (SOX +87.8% in Q2) mean-reverted hard — SMH −5.40%, MU −10.57%, KLAC −11.77% — and the proceeds rotated into megacap-software, financials and equal-weight. SPY barely moved; QQQ took the hit.
Econ (confirmed, midday reconcile): ADP +98K vs ~110K consensus — soft; ISM Manufacturing 53.3 vs 53.0 — a slight beat, still expansion. A cool-labor / firm-factory mix.
Rates: 10Y ~4.4% (est., build-ref) — yields firmed despite the soft ADP (TLT −1.04% confirms), a mild quarter-start/supply tilt rather than a growth-scare bid. VIX ~17.5 (est., web intraday; prev close 16.45) refresh-required — a modest lift on the semis air-pocket, not a fear spike.
Realized regime = rotation / dispersion (broadening breadth, equal-weight leadership) — exactly the Midday Frappe’s “rotation_dispersion” read. Stated as what happened, not a forward call.

LensThe index-level calm (SPY −0.14%) masks a violent under-the-hood rotation: the single most-crowded complex de-rated 5–12% in a session while the tape broadened. That is a healthier internal than last week’s narrow melt-up — but a 10%+ down day in the QTD-leader group is the kind of positioning event that can spill over if it continues into Thursday’s NFP.

04Cross-Asset

Asset	Proxy	Close	Day %	Read vs equities
US Dollar	UUP	28.49	+0.28%	firm — mild headwind for risk
Crude	USO	103.27	−2.98%	soft — XLE red, no energy bid
Gold	GLD	370.60	+0.60%	small bid; GDX −0.50% — muted safety demand
Long bonds	TLT	85.52	−1.04%	down (yields up) — no flight-to-safety
Credit (HY)	HYG	79.59	−0.48%	mildly soft — no credit stress under the unwind
Bitcoin	IBIT	34.00	+2.13%	up — risk appetite intact; diverged from semis

LensCross-asset says this was not a macro risk-off: long bonds sold (yields up, the opposite of a safety bid), gold caught only a small tick, HY credit stayed calm, and bitcoin rose +2.13%. The de-risking was confined to the crowded semis trade, not the whole tape — corroborating the rotation (not liquidation) read.

05Sector & Breadth (realized)

XLCComm+2.44

XLFFins+2.18

XLYDiscr+0.69

XLVHealth+0.55

XLBMaterials+0.37

XLREREITs+0.34

XLPStaples+0.28

XLEEnergy−0.56

XLIIndus−1.01

XLUUtils−1.26

XLKTech−2.57

7 green / 4 red — a broadening tape, the mirror image of last week’s 4-green narrow melt-up. Comm (XLC +2.44, META-led) and Financials (XLF +2.18) led; even defensives (XLV/XLRE/XLP) were green.
Tech was the sole heavy loser (XLK −2.57) — the semis unwind isolated to the one sector, not a broad growth de-rate.
Equal-weight vs cap-weight: RSP +0.30% > SPY −0.14% > QQQ −1.52% — broadening, the advance de-concentrating away from the megacap/semi complex. IWM −0.38% gave back a green midday.
Breadth internals ($S5FI/$S5TH/$ADRN/$TICK/$TRIN) refresh-required — Massive 403s index symbols and BarChart isn’t reachable headless; the read leans on the sector strip + RSP/SPY/IWM (no leg was gated on internals this run).

LensUnder a −0.14% SPY sits one of the cleaner broadening prints in weeks: money rotated out of the single crowded complex and into seven of eleven sectors plus equal-weight. Constructive internals — provided the semis air-pocket doesn’t turn into a broader deleveraging.

06Single-Name Movers

Illustrative of the day’s leadership only — never trade plans. From confirmed Massive closes.

Name	Close	Day %	Note
META	612.91	+8.81%	the standout — gapped up (open 607.91) and held; single-name catalyst move
MSFT	384.28	+3.02%	megacap-software bid absorbing the semis exit
AAPL	294.38	+1.73%	green — megacaps led, not semis
AMZN / GOOGL / TSLA	241.70 / 361.21 / 425.30	+1.4 / +1.1 / +1.1%	broad megacap participation
NVDA	197.58	−1.25%	held up best of the semis — lost $200
SMH (semis ETF)	620.46	−5.40%	the epicenter of the unwind
AVGO / AMD / TSM	369.34 / 540.88 / 444.23	−2.2 / −6.9 / −7.0%	broad semis de-rate
ASML / MU / KLAC	1843.04 / 1032.28 / 266.19	−7.4 / −10.6 / −11.8%	equipment/memory hit hardest

LensA textbook rotation within growth: capital left the crowded AI-hardware/memory leaders (KLAC −11.8, MU −10.6, ASML −7.4) and rotated into megacap-software/platforms (META +8.8, MSFT +3.0). This dispersion — leaders down double-digits, the broad tape green — is exactly the tape the midday SRT-semis short was built for, and why the continuation-into-semis longs were right to stand aside.

07Morning Lens vs Reality

The Early Bird Curd framed 7/1 as “digest-the-rip after a two-day semis-led melt-up” — explicitly flagging semiconductors as “the fundamental leader but the most extended cohort” into an event-heavy, holiday-shortened week. That extension flag was the day’s decisive read: the semis didn’t just digest — they unwound (SMH −5.40%, memory/equipment −7 to −12%) while the rest of the tape rotated higher. The Midday Frappe upgraded the call to “rotation_dispersion” in real time — and that is precisely how the session closed.

What the lens got right drove the scorecard: (1) the “most-extended semis” caution kept the MS-SEMI continuation long from chasing — it required a 737.62 reclaim that never came, sidestepping the unwind; (2) the midday rotation read put the SRT-semis short and the IWM rotation-long on the correct thematic side (the short fired; the long faded its tag); (3) the LRT-QQQ short had the index direction right (QQQ red) but its mechanism (broad risk-off, megacaps failing) was wrong — this was a rotation, megacaps led. The honest miss is granular, not thematic: the book was right about semis being over-extended and about the rotation; it was over-confident that small-caps would hold the rotation bid into the close.

LensLoop closed cleanly: the morning’s “extended semis / holiday-week caution” framing and the midday rotation upgrade both held, and the no-chase discipline was the day’s real edge — protecting capital from a 5–12% single-day unwind matters more than the two FIRES. The open question into Thursday’s NFP: does the semis air-pocket stay a rotation, or spread.

08Execution Debrief — realized execution on closed setups, not a forward trade plan

Setup	Profile	Entry	Stop	Target	Realized R
MS · L — md-260630-1209-MS-SEMI-L	day	734.70	733.75	736.13	+1.79R
LRT · S — md-260630-1209-LRT-QQQ-S	day	738.00	738.95	736.58	+1.68R
MS · L — mm-260701-MS-SEMI-L	swing	737.62	711.00	790.86	−0.47R
LRT · S — mm-260701-LRT-QQQ-S	swing	745.45	772.08	692.21	+0.76R
MS · L — md-260701-1209-MS-IWM-L	day	301.19	300.74	301.86	−4.16R*

Realized R = (close − entry) / (entry − stop), self-signed, exit = the validated close. *The IWM −4.16R is a close-as-exit artifact on a day setup: the 301.86 target was tagged intraday (IWM high 302.72), so a real day-fill would have booked ~+1.5R before the fade — the close (299.32) overstates the loss. The two never-armed swing shorts (−0.47R / +0.76R) are notional — no live tag, no fill. SRT-SEMI-S carries an exec block but its levels are QQQ-denominated against an SMH primary, so realized-R is omitted by construction (not mis-marked). Aggregate (Component 16, all 36 closed setups carrying exec): n=36 · mean realized R +0.57 · win-rate 66.7% · stop-rate 11.1% — tonight’s −4.16R outlier drags the mean.

LensThe honest execution read: the two 6/30 day setups paid clean (the MS long tagged its target into the markup; the LRT short’s notional +1.68R reflects a fill that never triggered), while the 7/1 continuation longs were right to sit out — the swing MS-SEMI would have carried a −0.47R paper mark into the semis unwind, but its no-arm discipline meant zero real exposure. Realized execution on closed setups — not a forward trade plan.