2026-10-23 14:00 ET — Mega-cap internet/cloud into next week's prints: GOOGL, MSFT, AMZN, AAPL (and META after its print)

Slot: Fri 2026-10-23, 14:00 ET hourly research. Status: complete (§9 second-researcher consult added ~14:25 ET).

TRAI data used: /api/latest (run 2026-10-22_premarket_2026-10-23T12-53-19-641Z, 8:53 AM ET; daily stats = Oct 22 session), /api/events (asOf 2026-10-23, gap-risk window 5 trading days), /api/correlations (60 sessions to 2026-10-22), /api/regime, /api/macro, /api/options (asOf 2026-10-22), /api/scorecard (replay 2026-04-23 → 2026-10-21). All returned 200 at ~14:05 ET. Intraday prices: Yahoo chart API 5-minute bars, last bar ~14:05 ET (my pull); "%" = vs the Oct 22 close. Reaction-day cross-returns: Yahoo daily bars (my calc).

1. Focus and why

• Focus: the five mega-cap internet/cloud names: GOOGL, MSFT, AMZN, AAPL, which all report within seven sessions, and META, which reported 10-21 and is in today's short_momentum bucket.
• Rotation: today's earlier reports covered semis (09:00), software (10:00), telecom/energy (11:00), AI power/industrials (12:00) and payments (13:00). The 09:00 report covered META's print only as the trigger for the semis selloff. None covered the hyperscalers' own setups.
• Why now: GOOGL and AMZN are mean_reversion fade-long candidates today, and both are inside TRAI's 5-trading-day gap-risk window (GOOGL 3 trading days to report, AMZN 4). META is short_momentum (TRAI /api/latest, /api/events). Hyperscaler capex guides are the demand side of the cluster-1 AI-hardware trade (NVDA, AVGO, VRT, ANET, ORCL), and that trade has had two −6% SMH days this month.
• One-line read: next week has two crowded nights. Wed 10-28: FOMC decision at 2 PM, then GOOGL and MSFT after the close. Thu 10-29: AMZN after the close, on the same day as GOOGL/MSFT's reaction session, Q3 GDP/PCE and MA/CAT's BMO prints. Options price ±5.5–7.7% moves. All five names' gaps since Oct 2025 have been ≥ ~2 ATR and rarely filled (except AAPL). Nothing scheduled changes gap risk on today's candidates before Monday.

2. Earnings calendar — verified

TRAI's /api/events dates and timings match all four IR announcements (no isEstimate flags).

Same-window collisions (all with sources):

• Wed 10-28: FOMC decision 2:00 PM ET, press conference 2:30 PM ET (Federal Reserve calendar, Oct 2026). Also VRT and GEV BMO; NOW, KLAC, TMUS, CMG and CVNA AMC (TRAI /api/events).
• Thu 10-29: reaction session for GOOGL and MSFT. Q3 advance GDP and September PCE at 8:30 ET (BEA schedule, per 13:00 report §7). MA, CAT, LLY and MRK BMO; AMZN, INTC, PINS, RBLX and COIN AMC (TRAI /api/events).
• Fri 10-30: AMZN reaction session; XOM, CVX and ABBV BMO (TRAI /api/events).

3. Per-name setup

3.1 GOOGL — Alphabet (Wed 10-28 AMC)

• Short-horizon drivers: Google Cloud growth and margin, TPU external sales, Search growth vs AI-search disruption, capex/FCF and financing.
• Last print (Q2, 2026-07-22): revenue $119.8B (+24%). Cloud +82% to $24.8B. Search & other +17%. Diluted EPS $9.11, including a $98.0B net gain mostly from unrealized gains on equity securities. Q2 FCF was −$5.86B (OCF $39.07B, capex $44.92B). Alphabet raised $49.6B in equity (Class A/C stock plus mandatory convertible preferred) and set up a $40.0B ATM program (Alphabet Q2 release, EX-99.1 on SEC). The 2026 capex guide went to $195–205B from $180–190B. The stock fell ~7% the next day (CFO Dive, 2026-07-30); TRAI has the 2026-07-23 gap at −6.13% and close-to-close at −7.13%.
• Expectations for Q3:
• Zacks consensus (via Yahoo): EPS $2.94, revenue $111.64B, Cloud $29.16B (+92.4%), Search & other $64.99B (+14.9%), YouTube ads $11.34B (+10.6%), TAC $16.53B (Yahoo/Zacks, Oct 2026).
• BofA: revenue $111.4B and GAAP EPS $3.07 vs Street $111.0B / $2.98. Cloud +93% vs ~89% consensus; Search ~15% in line, but BofA's checks suggest some deceleration. 2026 capex ~$200B (Street $202B) and ~$300B for 2027, "with a bias to the upside" (Proactive, Oct 2026).
• The buy-side GCP bogey is +105–115% vs Street ~+90% (Finvaulta summary of a J.P. Morgan desk note, 2026-10-21; secondary summary).
• Benzinga quotes revenue est. $128.20B (Benzinga, Oct 2026). This conflicts with the $111–113B from other sources, so I treat it as unreliable.
• Implied move: TRAI /api/options ±6.21% (Oct 30 straddle; 0.98× its historical avg |gap|). EarningsMoves ±6.25% (EarningsMoves, data 2026-10-22 3:53 PM). Pepperstone cites ±5.07% vs a 4.02% historical average (Pepperstone, 2026-10-22); the methods differ.
• TRAI history (last 4): avg |gap| 6.37% (2.45 ATR), max 6.90%, fill 0/4, continuation 50%, avg day-1 in gap direction −1.41%. Events: +6.20% (2025-10-30), −6.25% (2026-02-05), +6.90% (2026-04-30), −6.13% (2026-07-23).
• News today: Google Cloud made an Anthropic-managed Claude deployment option generally available on Google Cloud (Google Cloud blog, 2026-10-23). This is product news; no price impact found. GOOGL was −0.66% at 14:05 (open −1.03%, low −1.25%, high +0.10%).

3.2 MSFT — Microsoft (Wed 10-28 AMC, fiscal Q1 FY27)

• Short-horizon drivers: Azure growth vs guidance, capex and leases, FCF, and Microsoft Cloud gross margin. Gaming rarely moves the stock.
• Last print (FQ4, 2026-07-29): revenue $90B (+18%), Microsoft Cloud $59.3B (+27%), Azure +43%. The CY26 capex expectation was revised to ~$175B from $190B only because of an accounting change (data-center useful life 15 → 25 years moves some leases from finance to operating). Underlying CY26 capex plans were "unchanged" (CFO Amy Hood). The stock rose +15.5% the next day (CFO Dive, 2026-07-30); TRAI has gap +12.13% and close-to-close +15.51%.
• Guide for this quarter: Azure +45% cc, revenue $89.85–90.95B, capex ~$50B (vs ~$36B last quarter), EPS est. ~$4.70. Last quarter's FCF was −23% y/y. In January Azure grew 38%, above guidance, and the stock still fell ~10% (Tradingpedia, 2026-10-23); TRAI has 2026-01-29 close-to-close at −9.99%. EarningsMoves consensus EPS is $4.71.
• Implied move: TRAI /api/options ±5.45% (0.84× historical); EarningsMoves ±5.47% (EarningsMoves).
• TRAI history: avg |gap| 6.51% (2.62 ATR), max 12.13%, fill 0/4, continuation 4/4, avg day-1 in gap direction +1.53%. MSFT's gaps have extended on day 1 every time in the sample. Drift days 2–5 were +5.45%, +9.19%, −3.19% and +10.81%; the first two came after down-gaps, so they partly reversed.
• Positioning context: the stock is ~33% above its pre-July-print level and near its 52-week high of $553.72 (Tradingpedia, above). TRAI Oct 22: −2.23%, beta60 2.13, ATR 2.18%, no bucket signal. At 14:05 it was +0.63%, the strongest of the five today.

3.3 AMZN — Amazon (Thu 10-29 AMC)

• Short-horizon drivers: AWS growth and margin, capex/FCF, retail operating income vs guidance, advertising growth, and Anthropic mark-to-market in GAAP EPS.
• Last print (Q2, 2026-07-30):
• Sales $200.6B (+20%); operating income $27.5B (13.7% margin).
• AWS +37% to $42.2B at a 39.4% margin, including $551M of energy-contract gains; ads +26% to $19.8B.
• Net income $62.6B (EPS $5.75) included a $53.4B pre-tax gain, primarily the Anthropic revaluation.
• Capex was $53.1B and the 2026 plan went to ~$220B from $200B. Trailing FCF −$7.6B.
• The stock rose 15.3% the next day (Pepperstone, 2026-10-23; Investing.com, 2026-10-19). TRAI has gap +12.53% and close-to-close +15.32%.
• Guide and expectations:
• Guide: Q3 sales $197–202B (Prime Day moved into Q2, a ~400 bp headwind to Q3 growth); operating income $22.5–26.5B.
• Consensus EPS ~$1.96 on ~$202B revenue, which sits at the top of the sales guide. AWS consensus ~35%.
• Anthropic stake carried at $190.4B at 6/30. The $21.3B invested in OpenAI after 6/30 sits in investing cash flow (Pepperstone, above).
• Other AWS views: consensus "30–35%" (Investing.com, 2026-10-19); J.P. Morgan desk buy-side bogey ~42% vs Street ~37% (Finvaulta, above). The bar is unclear and the range is wide.
• Anthropic linkage: Anthropic is reportedly targeting a November IPO. A leaked S-1 (per WSJ) says it struggles to get enough capacity from AWS, Google and Microsoft, which increasingly see it as a competitor (TechCrunch, 2026-10-20). An IPO pricing event would re-mark AMZN's (and possibly GOOGL's) equity stakes. The size of GOOGL's Anthropic stake is not verified here.
• Implied move: TRAI /api/options ±7.74% (0.82× historical); EarningsMoves ±7.76%; Pepperstone ~±7.4%.
• TRAI history: avg |gap| 9.40% (3.51 ATR, the largest of the five), max 12.53%, fill 1/4, continuation 1/4, avg day-1 in gap direction −1.65%. AMZN's big gaps hold (unfilled) but day-1 tends to give a bit back.
• TRAI Oct 22: −4.57%, z1d −2.62, relVol 1.37, ATR 2.84%, beta60 1.75 → mean_reversion fade-long, cluster flag "Consumer Disc. ×3" (with HD/LOW, which S-001 shows are uncorrelated with AMZN). At 14:05: −0.58% (low −1.57%, high −0.15%), so no rebound so far.

3.4 AAPL — Apple (Mon 11-02 AMC, fiscal Q4)

• Short-horizon drivers: iPhone revenue and supply constraints, Services growth, gross margin and the holiday-quarter guide. This is the first call under new CEO John Ternus (Reuters via Yahoo, 2026-10-23).
• Guide for the quarter (given 2026-07-30):
• Revenue +9–11% y/y; gross margin 47–48%.
• iPhone growth "mid-teens", net of FX and supply constraints, which are expected to "increase significantly sequentially"; FX is a ~2.5-pt headwind.
• Services growth similar to the June quarter after FX.
• Source: 9to5Mac, 2026-07-30. The quarter (6/28–9/26) has only launch-week iPhone 18 Pro sales and no iPhone Duo (MacRumors, 2026-10-06).
• Consensus: EPS $2.00 from 8 estimates (EarningsMoves). A revenue consensus of ~$113.3B appears in search snippets but was not verified on a page I could read.
• Last reaction (2026-07-31): Services had its first sequential decline since 2022 and the company warned on supply. The stock was ~6.5% lower after hours (9to5Mac, above). TRAI: gap −8.58% (3.54 ATR), close-to-close −7.35%.
• News today: the iPhone Duo (first foldable, ~$2,000) went on sale. Supply is rationed, deliveries are slipping to late January, reviews are mixed, and IDC expects Apple to lead foldables by 2027 (Reuters via Yahoo, above). AAPL hit an all-time high in the morning (IBD, 2026-10-23 11:20 ET). It is a sell-the-news day so far: −0.49% at 14:05 after a high of +1.01%.
• TRAI: AAPL fails the liquidity filter (ATR% 1.75% < 2%), so it is not tradeable in TRAI today. History: avg |gap| 3.65% (1.62 ATR), fill 2/4, continuation 1/4, the most "fadeable" of the five. Implied move: TRAI ±5.05% (Nov 6 expiry, so it includes non-event days; 1.38× history); EarningsMoves ±4.76%.

3.5 META — Meta (reported 10-21 AMC; corrections to the 09:00 note)

• Primary-release facts (Meta IR, 2026-10-21):
• Revenue $64.60B (+26%), with ad impressions +10% and price per ad +14%.
• Costs +68%, including a $9.89B charge for the teen-mental-health settlement with 52 state AGs. GAAP EPS $3.83, or $7.40 ex-charge.
• Q3 capex $34.91B; FCF −$2.12B.
• Guidance: Q4 revenue $73–76B; 2026 capex $135–145B, "narrowed" from $130–145B; 2026 Reality Labs loss $18–19B, ~$15B in 2027. No 2027 revenue guide.
• "A number of youth-related trials scheduled in the coming months."
• Correction: the "new shares" are a new non-voting Class C stock, effective 2027-01-19. It will be distributed as a one-for-one stock dividend, so it has the economic effect of a 2-for-1 split. The record and payment dates are not set yet. Meta intends to use Class C for future capital raising, compensation and M&A (Meta IR, above; CFO Susan Li on the call per TechCrunch, 2026-10-21). No offering is scheduled. The 09:00 note's "plans to issue new shares" is right in spirit but needs this qualification.
• TRAI data issue: /api/events records the print as an EPS "surprise" of −29.91% ($3.83 vs $5.46 est.). Because of the one-off charge, that number is not comparable (see S-040).
• Today: META stabilized (Yahoo Finance live, 2026-10-23 2:05 PM ET). At 14:05 it was −0.53% (open −1.65%, low −2.40%, high −0.12%). For the short_momentum candidate, the round-trip pattern of S-027 is visible again.

4. Linkages

4.1 Correlations (TRAI /api/correlations, 60d to 2026-10-22)

• GOOGL–AMZN ρ 0.72 (cluster 20), the only mega-cap pair that clusters. Today's two fade-long candidates are effectively one position under the 1-per-cluster cap (S-001), and they report on consecutive nights.
• MSFT is a singleton (cluster 36). Top ρ: AMZN 0.50, ORCL 0.49, SOFI 0.48, SNOW 0.47, TSM 0.42, LRCX 0.42. MSFT trades with the AI-infrastructure complex more than with internet names (META −0.02).
• META is a singleton (cluster 31). Top ρ: CVNA 0.45, QCOM 0.42, HOOD 0.35, GOOGL 0.35.
• AAPL clusters with LIN (cluster 23, 0.50) and is negatively correlated with the AI-capex trade: AMZN −0.47, VRT −0.36, ORCL −0.32, ANET −0.32, NVDA −0.22, GOOGL −0.22. It acted as the hedge on Oct 22 (+1.23% while QQQ was −2.86%).
• Low 60d ρ to semis: GOOGL–NVDA 0.12, MSFT–NVDA 0.22, AMZN–VRT 0.32, META–AVGO 0.03. S-004's ρ ≥ 0.5 peer-event rule therefore would not flag AI hardware around hyperscaler prints, even though the reaction days below show large read-through.

4.2 What hyperscaler reaction days did to each other and to AI hardware (Yahoo daily, close-to-close %, my calc)

The sample is small (n = 13 distinct days) and the days are heavily confounded, because several mega-caps report on the same night.

Reads (anecdotal):

1. Dispersion between same-night reporters is large. On 2026-04-30 GOOGL was +9.96%, MSFT −3.93% and META −8.55%; on 2026-01-29 MSFT was −9.99% and META +10.40%. Reporters do not move together on their reaction day, so being long one and short another is two independent ~2.5-ATR bets, not a hedge.
2. The reporter's direction does not set the sign of AI-hardware read-through. On 2026-02-06 AMZN fell 5.55% (its capex guide; exact figure not checked here) while VRT rose 10.03%, NVDA 7.87% and AVGO 7.22%. On 2026-07-23 GOOGL fell 7.13% on its capex raise, and semis were roughly flat to down. On 2026-10-22 META fell 6.98% on capex/funding, and SMH fell 5.99% together with Lam (09:00 report). What matters is how the capex number is read: "more orders" versus "funding stress". This is not tradeable as a simple sign rule.
3. MSFT is the only reporter with a clean positive read-through to ORCL/SMH (2026-07-30: MSFT +15.51%, ORCL +8.34%, SMH +6.88%), consistent with MSFT–ORCL ρ 0.49.
4. AAPL's own days barely move the others. AAPL moves against AI-capex shocks (2026-10-22, 2026-07-31).

4.3 Replay check: TRAI buckets on mega-caps and around earnings (TRAI /api/scorecard replayTrades, net of 5 bp + impact, open→close, my calc)

• Mega-caps, all buckets: n 50. mean_reversion n 18, mean −0.46%, hit 38.9%, t −1.38. short_momentum n 13, mean −1.14%, t −1.38. long_momentum n 19, mean +0.18%.
• Whole universe by distance to own earnings reaction day:
• Trade on the reaction day itself: n 54, mean −0.15%, sd 4.34% vs 2.46% for non-event trades (n 1,009).
• Trade 1–5 sessions before the reaction day: n 64, mean −0.00%.
• Trade 1–5 sessions after: n 204, mean −0.31%, t −1.57.
• Reaction-day trades carry ~1.8× the dispersion with no evidence of edge.
• Caveats: the events file only holds each name's last 4 reports, survivorship applies, and the VWAP is a daily-bar proxy.

5. Macro and tape context

• TRAI regime: neutral (score 17.5), VIX "transitional" (19.98, VIX9D/VIX 0.94, mild contango), trend choppy (ADX 13.5). The 10y is 4.96%; HYG/IEF is −2.49 pts over ~1 month. XLK is −5.9% and XLC −5.3% vs SPY over 20d, both in the "lagging" quadrant (TRAI /api/regime, /api/macro, asOf 2026-10-22).
• Oct 22: Nasdaq −2.6%, its worst day since June 5. Apple (+1.2%) was the only Magnificent-7 gainer. The Nasdaq is ~5.8% below its Oct 6 high (Al Jazeera, 2026-10-23).
• Today 2:05 PM ET: Dow −1.2%, S&P −0.4%, Nasdaq ~flat. NVDA and AMD edged up; META, TSLA and SpaceX stabilized. Brent was ~$112 and WTI ~$103 on Middle-East tension (Yahoo Finance live, 2026-10-23 2:05 PM ET).
• Intraday (Yahoo 5m, ~14:05 ET): GOOGL −0.66% · AMZN −0.58% · META −0.53% · MSFT +0.63% · AAPL −0.49% · NVDA −0.12% · AVGO +0.45% · VRT −0.09% · ORCL +1.60% · SMH +1.00% · QQQ +0.22% · SPY −0.40%. The mega-caps are lagging a semis bounce, which is the reverse of Oct 22's leadership.

6. Actionable for today (Fri 10-23) and the weekend gap — explicit

• No catalyst found today changes the scheduled gap risk on a current TRAI candidate. All four upcoming dates match TRAI's /api/events.
• GOOGL and AMZN (fade-long candidates): both are inside the 5-day window ("elevated"), but the first reaction session is Thu 10-29. TRAI is flat by 3:55 PM, so no weekend or earnings gap is held from today's entries. GOOGL–AMZN ρ 0.72 means at most one of them under S-001. Neither has bounced (−0.66% / −0.58% at 14:05). Mega-cap fade-longs have a weak replay record (§4.3). These are tie-breakers, not vetoes.
• META (short_momentum): no event inside the window. It is round-tripping toward flat (low −2.40% → −0.53%), so S-027 applies. Youth-related trials are "scheduled in the coming months" (Meta IR). No specific weekend date was found.
• Weekend, unscheduled:
• Anthropic IPO headlines (reportedly targeting November; TechCrunch 2026-10-20) could move AMZN/GOOGL through their equity-stake marks.
• Oil and the Middle East (Brent ~$112).
• No mega-cap scheduled event over the weekend.
• Monday 10-26 heads-up: AAPL enters the 5-day window (still fails liquidity). GOOGL/MSFT drop to 2 trading days. Monday is the last full session before the FOMC/GOOGL/MSFT Wednesday.

7. Strategy implications (proposals; see strategy.md)

• S-038 (proposal): reaction-day risk sizing. On a name's own earnings reaction session, size off the realized opening range instead of ATR(14) alone. Reaction-day replay trades have sd 4.34% vs 2.46%.
• S-039 (proposal; refines S-010/S-004): S-010 already maps customers → suppliers for peer-event flags. New evidence: S-004's ρ ≥ 0.5 test misses these links, and the read-through sign doesn't follow the reporter. On hyperscaler reaction days, assume no direction for flagged AI hardware, and run no fades until the first 30 minutes set the SMH direction.
• S-040 (data fix): flag EPS-surprise values that are dominated by non-operating marks or one-off charges (GOOGL +214%, AMZN +215%, META −30%). Extends S-034.
• S-041 (proposal): crowded-night rule. When ≥ 2 universe mega-caps report the same night, or a reaction session coincides with 8:30 macro or the FOMC, treat that session as one event: max one position among reporters plus their ρ ≥ 0.5 peers. No pair trades between reporters.
• S-042 (question): do mega-caps belong in the mean_reversion bucket at all, especially in the T−5 window? The replay is weak (n 18, mean −0.46%).
• S-043 (question): AAPL as an anti-AI-capex hedge. Its ρ is negative to AMZN, VRT and ORCL. Does AAPL outperform on AI-capex shock days? It currently fails liquidity, so this matters mainly for factor and residual work.
• S-044 (question; whiggity): does the average GOOGL+MSFT overnight gap predict AMZN's open→close on AMZN's own report day (Thu before its AMC print)? See §9.

8. Unknowns / open questions

• GOOGL Q3 revenue consensus: $111.0–112.7B across sources; Benzinga's $128.2B looks wrong. GOOGL's Anthropic or other stake sizes were not verified.
• AWS consensus ranges from 30–35% (Investing.com) to ~37% Street and ~42% buy-side (J.P. Morgan desk via Finvaulta). The real bar is unclear.
• AMZN's Q4 2025 (Feb 2026) capex guide figure behind the 2026-02-06 reaction was not checked this hour.
• AAPL revenue consensus (~$113.3B) was not verified on a readable page.
• META's next earnings date has not been announced. The Class C record and payment dates are TBD.
• Whether TRAI's earnings-gap stats are robust: n = 4 per name.

9. Research partner (whiggity) consult

I consulted Kieran's Monica research bot whiggity in a new chat at ~14:00 ET and gave it this report's draft summary. I treat its reply as input from a second researcher, not as a source. Items are tagged [whiggity] and grouped by what I could check against my own cited sources (TRAI /api/events, Yahoo daily bars, the IR pages in §2–§3; accessed 2026-10-23 ~14:10–14:25 ET). I did not open or verify whiggity's own citations. Some were generic pages and one was a Facebook post. Where its claims are confirmed below, the confirmation comes from my own sources, not from its links.

Verified (kept, with my citations)

1. [whiggity] META reported Tue 10-21 after the close, with revenue $64.6B (+26% y/y): verified (Meta IR, 2026-10-21; §3.5). Correction: whiggity said Meta "raised capex". The release says the 2026 capex range was "narrowed" to $135–145B from $130–145B: the low end moved up $5B and the top did not change. The market's worry was funding (the $9.89B charge, Class C stock, FCF −$2.12B), not a capex raise.
2. [whiggity] The META short_momentum signal comes from the post-earnings day: verified. TRAI /api/events has the reaction session as 2026-10-22. That day's gap was −3.04% (0.94 ATR) and day 1 continued −4.07% in the gap direction (close-to-close −6.98%). Today's signal (daily stats = Oct 22) is therefore a day-2 post-earnings short, which is the setting where S-027's round-trip filter applies (§3.5, §6).
3. [whiggity] GOOGL and MSFT report Wed 10-28 after the close, and AMZN Thu 10-29 after the close: verified. This matches the IR announcements and TRAI /api/events (§2).
4. [whiggity] Oil hit a five-month high on Oct 22: verified on Yahoo daily (BZ=F front-month). Brent closed at $113.17 on Oct 22 (high $113.49). That was its highest close since 2026-05-04 ($114.44), about 5½ months earlier. Brent was ~$111.5 today, in line with the ~$112 in §5. WTI closed at $99.47 on Oct 22 and was not at a multi-month high on my bars. Front-month roll effects are not adjusted.

Contradicted by my data

1. [whiggity] AAPL reports Thu 10-29 (it flagged the date as unconfirmed): wrong. Apple IR lists Mon 2026-11-02 after the close (§2, §3.4). AAPL enters TRAI's 5-day window on Mon 10-26, not this week.
2. [whiggity] MSFT's gap is "two-stage" and "frequently reverses" the 4:05 PM direction during the call guidance: counter-evidence on the part I can measure. In TRAI's last 4 events, MSFT's reaction session continued in the gap direction 4/4 (+0.89%, +1.48%, +0.74%, +3.01%) and filled 0/4 (§3.2). Whatever happens after hours, the close→open gap has been a reliable read on the direction of day 1. Kept as an open question (below), limited to the after-hours path (4:05 PM → call → 8 PM). That is not tradeable in TRAI (flat by 3:55 PM, no after-hours trading), and daily bars cannot test it.
3. [whiggity] Typical gap ranges (no source given): only partly consistent with TRAI's measured last-4 |gaps|.

Each of whiggity's ranges contains the options-implied move, but the realized gaps fall outside them more often than inside. MSFT, AMZN and AAPL have fat tails both ways: small gaps (2–4%) and very large ones (8.6–12.5%). This supports §3 and S-038 (size off the realized opening range on the reaction day). It does not support using whiggity's ranges as stop or sizing bands.

Consistent with my findings, but whiggity's framing goes further than the data

1. [whiggity] Two read-through regimes: "capex divergence" (hyperscaler down, NVDA/AVGO/VRT up) vs "capex exhaustion" (both down, like Oct 22). Both happened in §4.2: divergence on 2026-02-06 (AMZN −5.55%, VRT +10.03%, NVDA +7.87%) and exhaustion on 2026-10-22 (META −6.98%, SMH −5.99%). Several reporter-down days fit neither: on 2026-01-29 (MSFT −9.99%) NVDA/AVGO/VRT were +0.52/−0.75/+0.69; on 2026-07-23 (GOOGL −7.13%) NVDA −1.56 and VRT +0.96; on 2026-04-30 NVDA −4.63 and VRT +7.29. The labels describe days after the fact. No pre-print variable has been shown to predict which regime occurs. This agrees with S-039: assume no direction for AI hardware and wait for the first 30 minutes. It is a useful label, not a rule.
2. [whiggity] AMZN's Thursday pre-print session (10-29) will trade on the GOOGL/MSFT cloud read-through, not on mean reversion. META's strong ads raise the bar for GOOGL/AMZN. This is plausible and consistent with GOOGL–AMZN ρ 0.72 (§4.1). A quick check on Yahoo daily bars (my calc; n = 3, anecdotal) gives mixed results:
3. 2025-10-30: GOOGL +6.20% / MSFT −2.04% → AMZN gap −1.41%, open→close −1.85%.
4. 2026-02-05: GOOGL −6.25% → AMZN gap −3.47%, o→c −0.99%.
5. 2026-07-30: MSFT +12.13% → AMZN gap +2.96%, o→c +0.92%.

The sign of AMZN's gap matched the reporters' average gap 2/3 times, and AMZN's open→close matched it 2/3. This becomes the S-044 test. It affects Thu 10-29, not today: TRAI is flat by 3:55 PM, so today's AMZN fade-long carries no exposure into next week. "META's ads raise the bar" is interpretation; I have no source for the GOOGL/AMZN ad consensus beyond §3.1 (YouTube ads est. $11.34B).

Open questions (not verified; not used as facts)

• [whiggity] Sympathy-squeeze risk on META if GOOGL/AMZN ads are strong. No source. META–GOOGL ρ is 0.35 and META–AMZN is not in META's top-ρ list (§4.1). This matters from Thu 10-29 (GOOGL/MSFT reaction), not today. Any META short_momentum signal that week would sit inside S-041's crowded-night rule.
• [whiggity] MSFT's after-hours path (4:05 PM print → call guidance) often reverses. Not tested. Daily close→open data says day 1 continues (item 6). Testing it needs after-hours 1-minute bars around the last 4–8 prints.
• [whiggity] Pre-earnings order books thin out 24–48 h before a print. No source, and TRAI has no depth data. relVol and spread data from TRAI's quote feed could be a proxy. Not tested.
• [whiggity] The mega-caps' share of S&P 500 earnings growth is "falling to ~9%". Its source was a Facebook post (EarningsWhispers), which I did not open or verify. Not used.
• [whiggity] Three test questions were filed in strategy.md:
• (1) Does mean reversion in GOOGL/AMZN fail during T−3..T−1 after an SMH ≤ −5% / QQQ ≤ −2.5% shock? Appended to S-042. Oct 22 met the shock condition (SMH −5.99%, QQQ −2.86%), so today is the first live day, n = 1.
• (2) Does the average GOOGL+MSFT overnight gap predict AMZN's open→close on its own report day? New S-044, NEEDS DATA.
• (3) After an SMH ≥ 5% five-day selloff, do NVDA/AVGO/VRT morning sympathy gaps on GOOGL/MSFT reaction days revert or trend into the close? Appended to S-039. Quick look (Yahoo daily, my calc): the condition held before 2026-02-05 (SMH 5d −8.31%) and 2026-07-30 (−14.09%). On 02-05 all three reverted (NVDA +0.42% gap → o→c −1.74%; AVGO +2.04% → −1.22%; VRT −3.43% → +0.83%). On 07-30, NVDA (+1.81% → +0.82%) and AVGO (+2.71% → +1.97%) trended, and VRT reverted (+6.35% → −4.09%). n = 2, mixed. SMH's 5-day return to the Oct 22 close was −6.35%, so the condition may be live for 10-29 depending on the next four sessions.

Did the consult change the verdict?

No. "Nothing actionable today" stands. Everything whiggity raised that checks out applies to next week's reaction sessions (Thu 10-29 / Fri 10-30), not to today's TRAI candidates:

• GOOGL and AMZN (fade-longs) and META (short_momentum) are all flat by 3:55 PM and carry no event exposure over the weekend.
• Its two date claims checked out (GOOGL/MSFT 10-28, AMZN 10-29) or were wrong (AAPL). Its META capex wording needed correcting.
• Its gap ranges understate the tails for MSFT, AMZN and AAPL, which strengthens S-038 rather than changing today's trades.
• The verified five-month Brent high is already in §5/§6 as weekend headline risk.

_Screening/research, not investment advice._