Weekly review · Week 30, 2026

Regime trajectoryCHOP

Mon 07-20) -> PULLBACK (Thu 07-23, deepened Fri 07-24). Verified independently against data/prices.parquet: SPY ret5 -0.587%, ret10 -2.122%, 15d drawdown -2.122% as of 07-24. Daily path: Mon CHOP (ret5 -0.95/ret10 -1.22, breadth 32.6% green, VIX 18.65) -> Tue/Wed compression (VIX 16.64 Wed trough, breadth recovering to 48.9) -> Thu regime FLIP to PULLBACK on the Mag-7 gap (ret10 crossed -1.5%, VIX +12.4% to 18.70, SPY GEX -2.47B) -> Fri deepened (ret10 through -2%) but ret5 STABILIZED (-1.67% -> -0.587%): a grind-down, not a crash. Vol-state ROUND-TRIPPED rather than net-expanded: VIX 18.65 -> 16.64 -> 18.70 -> 18.58, ending -0.4% vs Monday. GEX stayed FULLY_NEGATIVE (short-gamma amplification) on both SPY and QQQ every single session; VRP stayed FAIR all week (no premium-selling edge opened). directional_tradable=TRUE and s1_standdown=FALSE on ALL FIVE sessions -- the crash guard never fired, consistent with a grind. NOTE: the Phase-B momentum lane independently reported this week as CHOP with ret10 -1.37%; that read is WRONG and was discarded -- the panel gives -2.122%.

Sector leadership

Extreme dispersion inside a mildly-down index -- SPY -1.09% masks a ~8.6pp spread. Leaders (5-session, 07-17->07-24, all measured off data/prices.parquet): XLE +3.36% (Red Sea/crude), XLU +2.48%, XLI +1.81%, XLB +1.44%, XLRE +1.17%, XLV +0.92%. Middle: SMH +0.84%, XLK +0.17%, XLF +0.09%. Laggards: XLP -1.24%, XLC -3.93% (GOOGL/META), XLY -5.22% (TSLA, worst). NOTE: an earlier live-feed read of XLC lagged a session (last bar 07-23) and produced -6.45% over a 07-16->07-23 window; the panel figure of -3.93% over the correct 07-17->07-24 window supersedes it. The lag was transient and did not reproduce on re-query. The rotation is defensives + energy + industrials OVER mega-cap growth. uw sector_rotation money flow confirms: OUT of Technology (-$392.5M), Consumer Cyclical (-$160.6M), Industrials (-$50.3M); IN to Consumer Defensive (+$7.5M), Energy (+$2.2M). Critically, the near-52w-HIGH cohort has rotated OUT of tech entirely -- it is now industrials/insurance/rails (OII, TRV, FAF, PKG, WAB, CSX, RTX, BAC, SPG).

Notable flow

Dark-pool premium, week of 07-20 (top by $): SPY $52.1B (61k prints), MU $46.2B (144k prints), QQQ $33.8B (65k), SNDK $26.8B (107k), NVDA $25.6B (87k), AMD $18.0B (66k), TSLA $16.9B (62k), AAPL $14.9B (34k), INTC $11.1B (43k), GOOGL $11.0B (29k). MU printing $46B on 144k prints -- MORE prints than SPY and QQQ combined, and second only to SPY in dollars -- is the single most anomalous flow item of the week; SNDK at $26.8B/107k prints is likewise far above its market-cap weight. The memory complex absorbed institutional-scale two-way volume into its melt-up and then reversed hard on Friday. DIARY ONLY, 0 signal: raw flow measured as beta in research/20 and 70 and is not scored here.

Catalysts this week

2026-07-20/2026-07-24NO Tier-1 US macro print landed this week (no CPI/PPI/PCE/FOMC/NFP). Tier-2 only: Chicago Fed NAI (Wed), Existing Home Sales (Thu), flash S&P Global PMIs + New Home Sales (Fri).The week's entire structure formed WITHOUT a macro catalyst. This is a clean observation for the pre-registration accrual: the down-week candle is attributable to a single-name earnings cluster, not a macro print.
2026-07-22Alphabet (GOOGL) Q2 earnings -- strong quarter, but FY26 capex guidance RAISED to $195-205B from $180-190BGOOGL -7.1% Thursday 07-23 (342.09 -> 317.69), -7.80% on the week. Sold DESPITE a fundamental beat -- the market repriced AI capex intensity, not earnings quality. This is the week's dominant mechanism.
2026-07-22Tesla (TSLA) Q2 earningsTSLA -14.5% Thursday 07-23 (374.01 -> 319.69), -17.81% on the week -- its worst earnings-reaction day on record. Drove XLY -5.22% for the week.
2026-07-23Mag-7 aggregate repricing on AI-capex skepticismBloomberg Magnificent 7 Index -4.8% intraday, ~$767B of market value erased in one session. META -3.4%, AMZN -4.6% the same day on no company-specific news -- pure read-through. Nasdaq -1.66%, S&P -0.86%, Dow -0.73%. The regime flipped CHOP -> PULLBACK on this session.
2026-07-23Red Sea tanker attacks lift crudeXLE +3.36% on the week, the top sector -- an exogenous geopolitical bid that is orthogonal to the equity capex story and partly masks the index-level damage.
2026-07-20/2026-07-23Memory/storage melt-up on DRAM-NAND shortage (ongoing theme, no dated print)MU +8.48%, STX +8.13%, SNDK +6.03% on the week -- but ALL THREE reversed hard on Friday 07-24 (MU -7.0%, SNDK -10.8%, STX -6.8% from Thursday's close). A melt-up that gave back a third of its gain in the final session. Q1'26 DRAM contract prices +95%; MU fiscal-Q3 revenue +346% YoY.

Weekly candles

Each candle is normalised to its own weekly range — the source is cross-sectional at incompatible price scales, so a shared axis would misrepresent it. Real prices are in the tooltip and the table below.

Weekly candles — data table

Weekly candles: symbol, open, high, low, close, return
Symbol	Open	High	Low	Close	Return
SPY	747.06	750.02	735.21	738.93	-1.09%
QQQ	702.16	710.05	682.48	684.23	-2.55%
IWM	295.01	296.75	290.17	291.17	-1.30%

SPY-1.09%

Open: 747.06
High: 750.02
Low: 735.21
Close: 738.93

LOWER LOW, follow_through TRUE (confirms prior week's down direction). Second consecutive down week (-1.22% then -1.09%). Close in the bottom quartile of the weekly range. Not inside, not outside, no hammer/star. Closed below both 20-SMA (746.15) and 50-SMA (745.07).

QQQ-2.55%

Open: 702.16
High: 710.05
Low: 682.48
Close: 684.23

LOWER LOW, follow_through TRUE. Closed at 6% of the weekly range -- effectively on the dead low, the weakest close_pos of the three indices. Second consecutive down week (-3.12% then -2.55%), cumulative -5.6% over two weeks. The mega-cap-growth de-rating is concentrated here.

IWM-1.30%

Open: 295.01
High: 296.75
Low: 290.17
Close: 291.17

LOWER LOW, follow_through TRUE. Fourth consecutive down week, but the shallowest of the four (-0.59/-0.35/-1.30). Small-caps declined LESS than QQQ this week -- an unusual relative-strength inversion driven by the mega-cap-specific capex repricing rather than broad risk aversion.

Weekly technicals pre-registered · 0 points · not sized

Triple-index confirming lower-low week (SPY+QQQ+IWM all lower_low AND follow_through=TRUE)All three indices made a lower weekly low, closed in the bottom quartile of the weekly range (close_pos 0.25/0.06/0.15), and confirmed the prior week's down direction. No inside week, no outside week, no hammer/hanging, no star/inverted on any of the three.would signal: A unanimous confirming down-week with weak closes across large/mega/small caps WOULD read as trend-continuation into the following week -- i.e. a short-side tilt or a stand-aside for longs.
QQQ close_pos 0.06 -- close on the dead low of the weekly rangeQQQ closed at 684.23 vs a weekly low of 682.48 and high of 710.05 -- 6% of the range, the weakest of the three indices, with body_frac 0.65.would signal: A close on the weekly low with a dominant real body WOULD read as unresolved selling pressure carrying into the next week's open, and is the classic setup the continuation literature claims for.
Reversal-after-catalyst (Tier-1 macro arm)NO OBSERVATION THIS WEEK. Zero Tier-1 macro prints landed (no CPI/PPI/PCE/FOMC/NFP). The reversal-after-CPI arm accrues nothing from 2026-W30.would signal: n/a -- the arm requires a dated Tier-1 print to condition on.
Reversal-after-catalyst (single-name earnings-cluster arm)The week's entire candle structure was produced by ONE session (Thu 07-23) following a two-name after-close earnings cluster (GOOGL+TSLA on Wed 07-22). Thursday did NOT reverse the catalyst move -- it extended it, and Friday held the loss (SPY +0.10%, no bounce). No reversal-after-catalyst.would signal: A catalyst-day move that HOLDS through the following session WOULD read as genuine repricing rather than an overreaction to fade -- the opposite of the reversal setup. Distinguishing 'held' from 'faded' catalyst gaps is the discriminator this arm exists to test.
IWM outperforming QQQ in a down week (relative-strength inversion)IWM -1.30% vs QQQ -2.55% -- small-caps fell HALF as much as mega-cap tech, inverting the usual risk-off ordering where small-caps lead the downside.would signal: A down week where small-caps outperform mega-caps WOULD read as a factor/valuation repricing rather than a broad risk-off impulse, and WOULD argue against treating the pullback as a beta event. Untested -- this panel has no power for it.

Directional book

No sized calls this week.

Lane status3

laneOI_FADENo name clearedShortregime fit0.70

NO SCORED SINGLE-NAME. Regime is the most favorable this lane has seen in weeks (PULLBACK, ret10 -2.12%, short-gamma, defensives-over-growth rotation) and TWO names cleared every mechanical lane gate -- but both were already adjudicated and killed in the 2026-07-24 daily scan, which the lane agent did not carry. (1) ALLE: net_10d +4,308 (call +4,531/put +223), rel_build 2.539, persistence 0.412 (genuinely organic -- build CONTINUED post-print at +880 on 07-23 and +653 on 07-24), ADV $223M PASS, earnings 2026-10-22 clear of h21 PASS, not an ETP. FUNDAMENTALS VETO STANDS: Q2 EPS $2.40 vs $2.2559 (+6.39% surprise), revenue +13% YoY, management RAISED FY26 guidance, and TWO post-print PT raises on 07-24 (JPM $150->$170, Baird $175->$190) both 11-24% ABOVE the $153.36 close. The gap held (-0.78%) while the tape sold off = idiosyncratic strength. Nothing over the weekend changes this. Plus a PEAD tail cap: shorting into a held positive-surprise gap runs against documented post-earnings drift. (2) SSNC: mechanically top-ranked (rel_build 2.493) but INVALIDATED by catalyst resolution -- independently verified, the entire build (+335/+3,070/+731 on 07-20/21/22) landed BEFORE its 07-23 after-close print, then went +0 and +1 on 07-23/07-24 while the stock gapped +10.35% (66.95 -> 73.88). That is resolved pre-print positioning that already paid, not a crowd left to fade. NOTE: earnings_gate.py PASSES SSNC (next print 2026-10-22) because it only checks FORWARD earnings -- it cannot see a catalyst that already landed inside the build window. That is a real coverage gap in the helper, not a lane error. (3) ACVA: strictly worse than last week -- still fails the $50M floor ($25.6M/day), now ALSO earnings-blocked (08-10 inside h21), and last_call_net turned negative (calls genuinely closing). Also cut: AVTR/MWH/HRI/AVIR/AARD/SCM/CACI (earnings inside h21); SLGN/BKE/FHN/OSIS/TASK/DLLL/EQPT (persistence >=0.85, single-block artifacts); BITQ/IBX/CBRG (ETP). FOMC 07-29 + Core PCE 07-30 also sit inside any 2-4wk hold opened now.

ALLESSNC

laneMOM_LONGBasket watchLongregime fit0.30

BASKET-ONLY, never per-name -- and off this week. Regime-fit stays at ~0.30 (unchanged from W29): a long-near-52w-high basket into a confirmed PULLBACK (ret10 -2.12%), with all three indices closing the week on lower lows and QQQ on its dead low, is fighting the tape. 20 names pass full hygiene (ETP exclusion, $50M floor, trading-day earnings gate): OII, TRV, FAF, PKG, WAB, AIT, WSBC, APGE, BAC, RNR, CSX, SON, UNF, OVV, ATAI, HOMB, SPG, RTX, UBSI and one more. The COMPOSITION is the notable diary item -- the near-52w-high cohort now contains essentially no mega-cap technology; it is industrials, insurance, rails and regional banks. The cohort survived the week precisely because it is not what got de-rated. Prior remains +0.18% mean / -1.06% MEDIAN, tail-driven (n=457) -- the negative median is why this never sizes per-name. Off until a trending-up regime re-establishes (ret10 back >+1% for >=2wk).

OIITRVFAFPKGWABBACCSXRTXSPG

laneMOM_SHORTStood downShortregime fit0.60

STOOD DOWN FOR NEW STARTERS BY STANDING POLICY, NOT BY THE CRASH GUARD -- the distinction matters. s1_standdown = FALSE on all five sessions this week, and affirmatively so on Friday: ret10 -2.12% is past the -2% confirmed-downtrend boundary, so this is explicitly NOT the unconfirmed-V-bounce setup the guard exists to catch. The lane is nonetheless capped watch-only for new starters under CLAUDE.md invariant #6, because its re-baselined excess is NEGATIVE: -0.03% mean, n=281. The hit rate (0.48 vs 0.27 base, +20.6pp) is NOT the binding read and must not be cited as one -- this lane wins often and loses big. Six names pass full hygiene at the weekly horizon: ORCL (already held, see carry-forward), ACI, CCI, CMCSA, OKLO, AZO. None sized. Regime-fit 0.60 is genuine (PULLBACK + short-gamma is the supportive case for a short lane) but moot under the cap. CORRECTION LOGGED: the Phase-B momentum agent assigned this 0.60 off a CHOP label with ret10 -1.37%; the panel gives PULLBACK at -2.122%. The fit lands in the same place for a different and better reason.

ACICCICMCSAOKLOAZO

Lane glossary (3)

OI_FADE: OI-flow fade — fades persistent multi-day net call-OI building; heavy call-OI build precedes underperformance. The most robust lane measured (short, h10).
MOM_LONG: Momentum, long leg — near-52w-high relative strength. Tail-driven; basket-only, never sized per-name (h10).
MOM_SHORT: Momentum, short leg — near-52w-low relative weakness. Regime-gated against momentum-crash/squeeze rebounds (h10).

Bottom line: an AI-capex repricing week disguised as a quiet one. SPY fell only 1.09%, but that number hides a ~9pp sector spread and a single-session Mag-7 shock. No Layer-2 call is scored. Two OI_FADE names cleared every mechanical gate and both were killed on grounds the lane cannot see by itself.

1. The Week in Pictures (DIARY — documentation, 0 signal)#

Weekly candles#

	Open	High	Low	Close	w_ret	close_pos	body	Structure
SPY	747.06	750.02	735.21	738.93	−1.09%	0.25	0.55	lower low, follow-through, 2nd down week
QQQ	702.16	710.05	682.48	684.23	−2.55%	0.06	0.65	lower low, follow-through, closed on the dead low
IWM	295.01	296.75	290.17	291.17	−1.30%	0.15	0.58	lower low, follow-through, 4th down week

All three made lower lows, closed in the bottom quartile of their range, and confirmed the prior week's direction. No inside week, no outside week, no hammer or star on any index. SPY closed below both its 20-SMA (746.15) and 50-SMA (745.07). QQQ is now −5.6% over two weeks.

Catalysts, anchored to price#

No Tier-1 macro landed all week — no CPI, PPI, PCE, FOMC or NFP. The entire week's structure came from a two-name earnings cluster.

When	Event	Price reaction
Wed 07-22 AH	GOOGL Q2 — beat, but FY26 capex guidance raised to $195–205B from $180–190B	GOOGL −7.1% Thu, −7.80% on the week. Sold despite a beat
Wed 07-22 AH	TSLA Q2	TSLA −14.5% Thu — worst earnings-reaction day on record; −17.81% on the week
Thu 07-23	Mag-7 capex de-rating	Mag-7 index −4.8% intraday, ~$767B erased in one session. META −3.4%, AMZN −4.6% on no company news — pure read-through. Nasdaq −1.66%. Regime flipped CHOP → PULLBACK here
Thu 07-23	Red Sea tanker attacks lift crude	XLE +3.36%, top sector — exogenous, orthogonal to the equity story
Mon–Thu	Memory/storage melt-up (DRAM contract prices +95% in Q1'26; MU fiscal-Q3 revenue +346% YoY)	MU +8.48%, STX +8.13%, SNDK +6.03% on the week — all three reversed hard Friday (MU −7.0%, SNDK −10.8%, STX −6.8%)

The mechanism worth remembering: Alphabet was sold on a good quarter. The market repriced AI capex intensity, not earnings quality — which is why the damage was surgical (mega-cap growth) rather than broad.

Regime trajectory#

	Regime	ret5	ret10	breadth %green	VIX	SPY GEX	s1_standdown
Mon 07-20	CHOP	−0.95%	−1.22%	32.6	18.65	−1.94B	FALSE
Tue 07-21	CHOP	—	—	45.5	17.05	less neg	FALSE
Wed 07-22	CHOP	−0.98%	+0.27%	48.9	16.64	short-γ	FALSE
Thu 07-23	PULLBACK	−1.67%	−1.80%	42.4	18.70	−2.47B	FALSE
Fri 07-24	PULLBACK	−0.587%	−2.122%	72.2	18.58	−1.556B	FALSE

A grind-down, not a crash. ret10 deepened through −2% while ret5 stabilized (−1.67% → −0.59%). Vol round-tripped rather than expanded — VIX 18.65 → 16.64 → 18.70 → 18.58, ending −0.4% on the week — but GEX stayed FULLY_NEGATIVE on both SPY and QQQ every session, and VRP stayed FAIR throughout (no premium-selling edge opened). The crash guard never fired on any day.

Friday's breadth print is the week's best single tell: 72.2% green while SPY closed +0.10%. The median stock rallied; the cap-weighted index was dragged by its largest members. Options-flow breadth was only 33% bullish over the same session — a hedging tell, two-sided, not a short-book confirmation.

Sector leadership#

XLE +3.36% · XLU +2.48% · XLI +1.81% · XLB +1.44% · XLRE +1.17% · XLV +0.92% — then SMH +0.84% · XLK +0.17% · XLF +0.09% — then XLP −1.24% · XLC −3.93% (GOOGL/META) · XLY −5.22% (TSLA, worst). An ~8.6pp spread inside a −1.09% index. Money flow confirms: out of Technology (−$392.5M) and Consumer Cyclical (−$160.6M), into Consumer Defensive and Energy.

Corrected. An earlier live-feed read had XLC a session stale (last bar 07-23) and gave −6.45% over a 07-16→07-23 window. The panel figure is −3.93% over the correct 07-17→07-24 window. The lag was transient and did not reproduce; all sector figures above are now measured off data/prices.parquet.

The structural item: the near-52w-high cohort now contains essentially no mega-cap technology. It is industrials, insurance, rails and regional banks (OII, TRV, FAF, PKG, WAB, CSX, RTX, BAC, SPG). That cohort survived the week precisely because it is not what got de-rated.

Notable flow (diary only — raw flow measures as beta here, `research/20`, `70`)#

Dark-pool premium: SPY $52.1B (61k prints) · MU $46.2B (144k prints) · QQQ $33.8B (65k) · SNDK $26.8B (107k) · NVDA $25.6B (87k) · AMD $18.0B · TSLA $16.9B · AAPL $14.9B.

MU printed more individual prints than SPY and QQQ combined and ranked second in dollars to SPY alone. The memory complex absorbed institutional-scale two-way volume into its melt-up, then reversed on Friday.

2. Validated Weekly Setups (SCORED — excess)#

`calls[]` is EMPTY. No Layer-2 call is scored this week.#

Lane	Disposition	Fit	Why nothing sized
OI_FADE	NO_NAME_CLEARED	0.70	Two names cleared every mechanical gate; both already killed — see below
MOM_LONG	BASKET_WATCH	0.30	Basket-only always; long-near-high into a confirmed PULLBACK fights the tape
MOM_SHORT	STOOD_DOWN	0.60	Watch-only cap for new starters (invariant #6) — not the crash guard

OI_FADE — the most favorable regime this lane has seen in weeks, and still nothing to trade.

ALLE — net_10d +4,308 (call +4,531 / put +223), rel_build 2.539, persistence 0.412, ADV $223M, earnings 10-22 clear of h21. Genuinely organic: the build continued after the print (+880 on 07-23, +653 on 07-24). Fundamentals VETO stands — Q2 EPS $2.40 vs $2.256 (+6.4%), revenue +13% YoY, FY26 guidance raised, and two same-day PT raises (JPM $150→$170, Baird $175→$190) both above the $153.36 close. The gap held while the tape sold off = idiosyncratic strength. A PEAD tail cap applies on top.
SSNC — mechanically top-ranked (rel_build 2.493) but invalidated by catalyst resolution. Independently verified: the entire build (+335 / +3,070 / +731 on 07-20/21/22) landed before its 07-23 after-close print, then went +0 and +1 on 07-23/07-24 while the stock gapped +10.35% (66.95 → 73.88). That is resolved pre-print positioning that already paid — not a crowd left to fade.
ACVA — strictly worse than last week: still fails the $50M floor ($25.6M/day), now also earnings-blocked (08-10 inside h21), and last_call_net has turned negative (calls genuinely closing).
Also cut: AVTR/MWH/HRI/AVIR/AARD/SCM/CACI (earnings inside h21) · SLGN/BKE/FHN/OSIS/TASK/DLLL/EQPT (persistence ≥0.85, single-block artifacts) · BITQ/IBX/CBRG (ETP).

Helper-script coverage gap found. earnings_gate.py passes SSNC, because it only checks forward earnings — it cannot see a catalyst that already landed inside the build window. Nothing was mis-sized (the 07-24 scan had caught it independently), but the gate does not protect this case on its own.

MOM_SHORT — the stand-down reason matters. s1_standdown = FALSE on all five sessions, affirmatively so on Friday: ret10 −2.12% is past the −2% confirmed-downtrend boundary, so this is explicitly not the unconfirmed-V-bounce the guard exists to catch. The lane is capped anyway by standing policy because its re-baselined excess is negative (−0.03% mean, n=281). The +20.6pp hit-base is not the binding read. Six names pass full hygiene (ORCL held, plus ACI, CCI, CMCSA, OKLO, AZO); none sized.

3. Weekly Technicals — PRE-REGISTERED (documented, 0 points, NOT sized)#

Feature	This week	Would signal	Status	Bar
Triple-index confirming lower-low week	SPY+QQQ+IWM all lower-low, all follow-through TRUE, all closed bottom-quartile	Trend continuation → short tilt or stand-aside for longs	PRE-REGISTERED	PR-WT (≥2yr)
QQQ close on the dead low	close_pos 0.06, body_frac 0.65	Unresolved selling into next week's open	PRE-REGISTERED	PR-WT
Reversal-after-catalyst (Tier-1 macro arm)	No observation — zero Tier-1 prints landed	n/a — arm needs a dated macro print	PRE-REGISTERED	PR-WT
Reversal-after-catalyst (earnings-cluster arm)	Thu 07-23 gap extended, did not reverse; Friday held the loss	A catalyst move that holds = genuine repricing, not an overreaction to fade	PRE-REGISTERED	PR-WT
IWM outperforming QQQ in a down week	−1.30% vs −2.55%, inverting the usual risk-off ordering	Factor repricing rather than broad risk-off; argues against treating it as a beta event	PRE-REGISTERED	PR-WT

None of these earned a point and none may be sized. The macro arm accrues nothing from W30 — worth noting so the pre-registration count isn't overstated later.

4. Carry-forward#

Held book — two shorts, both working, both already managed on 07-24.

	Lane	Entry	Mark	Gross	Excess	State
ORCL	MOM_SHORT	131.54 (07-13)	114.99	+12.58%	+4.31%	Trimmed 50% on 07-24; remainder on tightened invalidation
PLNT	OI_FADE	55.20 (07-20)	53.48	+3.12%	+1.48%	HOLD, hard exit 2026-08-05

ORCL — hold the remaining half. Live invalidation is close > $123.50 with 2-day confirmation, currently +7.40% away. The entry-era stop (RSI>45 AND close>$144.22) went vestigial once the position ran >10% in its favor — held to it, the trade would have surrendered the entire gain before any documented exit fired. ORCL printed a fresh 52-week low ($114.75 intraday) on Friday. No adds — the MOM_SHORT tail cap forbids new sizing. Reassess by the 08-07 NFP boundary; own ER 09-08 is clear.
- Correction logged: the Phase-B momentum lane read ORCL against the superseded entry-era 52w-low (134.57) and called the stop ~19.6pts away. The live stop is $123.50 — roughly half that distance.
PLNT — hold to the deadline, do not extend to the weekly horizon. earnings_gate.py confirms BLOCKED at h21: the print has drifted to 08-06, inside any 2–4 week hold. Signal itself is intact (post-entry call_net +830/+509/+179/+173, no unwind) and is not vestigial at +1.48% excess. Exit 08-05.

Into next week: FOMC 07-29 and Core PCE 07-30 both land inside any new 2–4wk hold, with MSFT+META (07-29) and AAPL+AMZN (07-30) reporting into the same window — on FULLY_NEGATIVE GEX, where dealer hedging amplifies a surprise in either direction. Vol did not compress into that cluster. Against a book that is 100% short and a lane set producing nothing, the posture is: manage the two held shorts to their documented levels, add nothing pre-event.

Watch: ALLE re-examines only if the raised FY26 guidance is walked back, the PT raises reverse, or the 07-23 gap fills without the call-OI build unwinding. SSNC is done — its catalyst resolved.

Verification notes#

Regime figures recomputed directly from data/prices.parquet (SPY ret5 −0.587%, ret10 −2.122%). The Phase-B momentum lane independently reported CHOP / ret10 −1.37%; that read was wrong and discarded.
preflight.py --date 2026-07-24 clear: 5/5 panel groups, truth set reaches the trade date.
oi_build.py, earnings_gate.py, liquidity.py, chart.py, held_book.py used for all per-name figures; SSNC/ALLE day-by-day OI and the ALLE/SSNC price paths were re-derived rather than taken from the lane.
validate_decision.py --file analyses/weekly/2026-W30/decision.json → VALID.

Sources for the week's catalysts: TheStreet, Jul 23 · Motley Fool, Jul 23 · Mag-7 $767B · CNBC, Jul 23 · 24/7 Wall St — memory rally · 24/7 Wall St — memory pricing