2026-11-02 10:00 ET: Software day 2. Friday's losers split: the no-news, pre-print consumption names (DDOG, NET, MDB) keep falling into this week's prints, while the earnings flush (RBLX) and APP bounce. IGV lags QQQ. ISM beats (55.5 vs 54.5–54.7) and the market barely moves

Slot: Mon 2026-11-02 10:00 ET hourly research. Status: DRAFT, written ~10:12–10:45 ET. Screening/research, not investment advice.

TRAI data used:

• /api/latest: premarket run 2026-10-30_premarket_2026-11-02T13-58-09-807Z (8:58 AM ET; marketSessionDate 2026-10-30), read ~10:13 ET.
• /api/events (asOf 2026-11-02), /api/correlations (60d, asOf 10-30), /api/scorecard, /api/regime, /api/macro (asOf 10-30). All returned 200.
• Quotes: Yahoo chart API, regular-session 5-min bars, latest bar 10:15 ET, read 10:15 ET. 1-min bars were used around the 10:00 release (Yahoo chart API). Unofficial, may be delayed. All statistics are my own calculations unless a source is cited.

Previous slots: 09:00 today covered oil/OPEC+, PSX, TMUS, CHKP and MAR. Friday's 10:00 covered the app-vs-consumption software split, and 11:00 the semis rout. This slot doesn't repeat them. It covers why each of Friday's software/internet losers fell, their next catalysts, how day 2 is trading, a day-2 base rate, and the 10:00 releases.

1. Focus and why

1. Four of the eight short-momentum names (bucket suspended) are software/internet: RBLX −18.1%, DDOG −11.5%, APP −10.4%, TEAM −6.1%. Two mean-reversion fade-longs (active) are software and report within one session: PLTR −6.3% (tonight AMC) and SHOP −5.6% (Tue BMO). All figures from /api/latest.
2. Only one of the six fell on its own news (RBLX's Q3 print). The other five fell on the AI-spending/software sell-off, with no company release, and four of them print this week (§2). That difference shows up in day-2 trading (§3).
3. 10:00 releases: ISM Manufacturing (Oct) and construction spending (Sep) (§5).

2. Why each fell on Fri 10-30, what it does, and what's next

Sources for this table:

• Cluster and gap columns: TRAI /api/events, /api/correlations.
• Business lines: company notes.
• Dates: Datadog IR, AppLovin IR (10-07), Atlassian IR, Palantir via Business Wire (10-12), Shopify IR.

(a) RBLX: the only true earnings flush

• Q3:
• revenue $1.57B (+16% y/y)
• bookings $1.70B (−12%)
• DAU 138M (−9%)
• engagement hours ~34B (−15% y/y, +17% q/q)
• Q4 guide:
• bookings $1.89–2.00B, a 10–15% y/y decline, vs ~$2.01B consensus
• adj. EBITDA loss $10–70M
• FCF −$105M to −$190M
• Analyst targets:
• Morgan Stanley $55→50 (Overweight)
• UBS $53→50 (Neutral; cited GTA VI and 2027 operating leverage)
• Roth $60→54
• Wedbush $55→52
• Freedom Capital raised $55→60
• Sources: Yahoo/Stocktwits, 2026-10-30. StoneX $32 and Jefferies $35 appear in search summaries only (Investing.com); not read in full.

(b) DDOG: multiple compression, not estimates

• Closed $235.27 (−11.50%) "with no company announcement behind the drop". Next-12-month EPS estimate unchanged at ~$2.65 (9-30 vs 10-30). Forward P/E ~103× → ~89×. Wells Fargo raised its PT to $325 on 10-28 (Yahoo Finance/TIKR, 2026-10-31).
• Same source: the reported ~$4B Upwind deal "now looks unlikely" (TechCrunch 10-28: Upwind walked away; unconfirmed by either company).
• All ten software stocks in SaasBuzz's count fell Friday, led by DDOG −11.5% and NET −9.5%; IGV −3.4% on the day and −5.6% on the week (SaasBuzz, 2026-10-30).
• Q3 setup (11-04 BMO):
• guide: revenue $1.135–1.145B, non-GAAP EPS $0.63–0.65
• consensus: ~$1.14B / $0.64 (i.e., at the guide midpoint)
• Earnings Whispers "whisper": $0.69
• FY26 revenue guide: $4.45–4.47B after two raises
• known headwind: the largest customer (an AI company) renewed but with fewer seats from this quarter, so headline growth is expected to slow from 36% in Q2
• Source: Earnings Whispers via Finviz, 2026-10-27.

(c) APP: no news. Sell-off plus a litigation calendar

• "The company released no announcements, earnings reports, or filings." Close $228.50 was its lowest of the past year (high $733.60 on 2025-12-22). The stock has had six one-day falls at least this large in 365 days (Trefis, 2026-11-01).
• Q3 revenue guide: $2.055–2.085B (same source). Q2 revenue was +52.8% to $1.92B, slightly below ~$1.94B consensus; management said the "pace of meaningful model improvement" was lighter than normal (Tickeron, 2026-10-30; secondary).
• Securities class actions: class period 2026-02-12 → 08-05; lead-plaintiff deadline 2026-11-16 (Tickeron, above; BestStocks, 2026-10-30). BestStocks' consensus EPS of $4.04 is unverified.

(d) TEAM: no company news found

• 10-30 close $191.72 (−6.07%), open $202.50 = high, low $188.43 (exa daily, 2026-10-30).
• The explanations I found (AInvest) are AI-generated and inconsistent (e.g., wrong prior close), so I am not using them. Consensus for FQ1 is unverified. Cause: sector sell-off (IGV −3.4%) plus pre-print de-risking; it's in cluster 22 with CRM (ρ 0.55), with ρ 0.60 to NOW and 0.68 to PLTR.

(e) SHOP: sympathy, prints tomorrow BMO

• Listed by StockStory among stocks that "fell in the afternoon session after underwhelming earnings reports from tech giants Amazon and Intel" (StockStory, 2026-10-30).
• Q3 consensus (aggregated in search summaries, not read in full): revenue ~$3.74–3.75B (~+31%), EPS ~$0.45–0.47, GMV ~$118.7B (MarketWatch preview; Yahoo/Zacks preview). Treat as approximate. Company guide: revenue growth in the low thirties % (SHOP note).

(f) PLTR: prints tonight

• Guide and consensus are in the 10-30 12:00 report. Revenue guide $2.160–2.164B vs consensus ~$2.18B; options-implied ~11%. Not repeated here.
• Cluster 5 now has PLTR (11-02), DDOG (11-04) and NET (11-05) reporting within four sessions. That meets the S-085 cluster earnings-window condition. PLTR–DDOG ρ 0.67, PLTR–NET 0.69, DDOG–NET 0.80.

3. Day 2 so far (open → ~10:15 ET, Yahoo 5-min, unofficial)

• Software is lagging the bounce. QQQ is +0.8% but IGV is −0.4%, and the weakness is concentrated in cluster 5 / consumption (DDOG, NET, MDB, CRWD, SNOW). Seat/app software (NOW +2.7%, CRM, ADBE) and the ad/gaming names (APP, RBLX, U +5.9%, TTWO +1.7%) are up. It's the same factor split as Friday (see software note), with the difference that the cluster-5 names report this week.
• No new company news found for DDOG or NET today (searched ~10:20 ET); coverage still points to last week's AI-spending sell-off (TIKR). Yahoo's live blog notes Anthropic filed an amended S-1 this morning (Yahoo live, 2026-11-02 10:07). No source ties that to software moves; that link is unverified.
• Short-momentum read (bucket suspended): 2 of 4 SM software names are continuing (DDOG, TEAM) and 2 are reversing (RBLX, APP). The reversing ones are the earnings flush and the name that already had the deepest drawdown. The continuing ones are pre-print no-news names. Small sample (n 4, one morning).

4. Base rate: day 2 after a ≥ 10% post-earnings drop (software/internet, my calc)

• Method:
• events: TRAI /api/events reaction days in Software + Internet/Media with reaction close→close ≤ −10%, last four prints per name (2025-11 → 2026-10), all non-inferred dates
• excluded: RBLX 10-30 (today is its day 2)
• n = 20, 13 of them Software-sector
• day-2 bars: Yahoo daily, with QQQ for excess
• Results:

• Read:
• No reliable edge. Day 2 is a coin flip, with a small positive mean driven by a few large reversals (APP 2026-02 +7.0%, RBLX 2026-05 +5.9%, NOW 2026-04 +5.1%).
• RBLX's own two prior cases were both up on day 2 (+5.90%, +3.06% o→c), but then lagged QQQ by −3.9 and −2.6pp over days 2–5.
• This fits the scorecard: the short_momentum replay is n 415, net mean −0.45%, hit 41.7%, t −3.62 (/api/scorecard, 05-01 → 10-29). Shorting after big down days loses on average, but that doesn't make the opposite trade validated.
• It also fits the software note base rate (≤ −15% days, next o→c +0.28%, n 46).
• Caveats:
• n is small; one year only; TRAI's four-print window per name.
• The reaction day is defined by TRAI's event dates.
• The universe is survivors.

5. 10:00 ET releases

(a) ISM Manufacturing PMI, October 2026: beat.

• Headline: 55.5 vs 54.5 in September (prior). 10th straight month of expansion (ISM via PR Newswire, 2026-11-02 10:00 ET; ISM).
• Consensus: 54.5 per Investing.com and investingLive; 54.7 per MetaTrader calendar (the 54.7 matches the FHLBNY figure in the 16:00 report). Beat by 0.8–1.0 pt either way.
• Components:
• New Orders 55.5 (from 55.3)
• Production 59.0, highest since Nov 2021
• Employment 52.4 (from 52.7)
• Supplier Deliveries 59.1
• Inventories 51.4 (from 48.6)
• Prices 75.2 (from 77.9, still very high)
• Backlog 54.6 (from 56.4)
• Customers' Inventories 41.6
• All five PMI sub-indexes are in expansion. ISM maps 55.5 to ~2.7% annualized real GDP (ISM, above).
• Read: strong activity with input inflation easing but still high. That leans against rate cuts at the margin, a headwind for long-duration software.

(b) Construction spending, September 2026: weak.

• Total $2,186.1B SAAR, −1.9% (±0.7%) m/m vs revised August $2,227.4B, and −2.7% y/y.
• Private −2.4%: residential −5.5%, nonresidential +1.1%. Public −0.1% (not significant) (Census, 2026-11-02).
• Consensus: not found in the sources I checked.

(c) Market reaction: muted.

• Gold slipped 3,974.8 → 3,966.9 by 10:02, which fits a firmer-data move.
• Yahoo's ^TNX 1-min feed stalled at 09:59 (4.88%), so I can't measure the 10y move precisely.
• Net: the ISM beat and the construction miss roughly offset; no regime-relevant move. Software didn't trade off on the print. IGV was already lagging before 10:00.
• Regime: VIX 18.79, VIX9D 18.88 at 09:58 (Yahoo) → live ratio 1.005, just above 1. On Mondays the ratio runs higher by +0.044 on average (S-112), so this is borderline, not stress. Unchanged from the 09:00 read (S-102).

6. Implications for today's buckets (research, not advice)

• PLTR / SHOP (MR fade-long, active):
• Both are above VWAP, inside OR30, and up modestly vs the prior close.
• TRAI is flat by 3:55 PM ET, so tonight's PLTR and tomorrow's SHOP prints carry no gap exposure for an intraday trade.
• Pre-print sessions are the S-106 open question.
• PLTR's cluster-5 peers are the weakest group today (DDOG −4.6%, NET −4.4%). That cuts against a PLTR fade-long working cleanly (ρ 0.67–0.69).
• SM names (suspended): research only. DDOG/TEAM continuation vs RBLX/APP reversal is logged as S-120.
• Cluster-level print risk this week: cluster 5 meets the S-085 condition (PLTR, DDOG, NET within four sessions). Cluster 12 has APP (11-04 AMC) and QCOM (11-04 AMC) on the same night.

7. Research partner (whiggity)

Consulted ~10:15 ET in a new Monica chat with whiggity, given the hour's focus (independent research requested). Its reply added little this hour: two of its searches errored and it found no drivers for DDOG, APP, TEAM, PLTR or SHOP, no catalyst dates and no ISM/construction figures, deferring those to the Catalyst Scout and Risk Auditor bots.

• From whiggity, consistent with this report: Roblox reported Q3 2026 after the close on Thu 10-29 with a 16:30 EDT call (StreetInsider/Business Wire, 2026-10-29). Its other RBLX items (a tweet calling results "a mixed bag") are unverified and not used.
• Nothing contradicted the drivers, the IR-verified print dates or the ISM/construction numbers above.
• Open questions it raised (worth testing, not answered): (1) day-2 intraday return distribution after 10:00 entry following 5%–18% single-day software drops, without look-ahead; (2) peer contagion vs idiosyncratic drift, i.e. does an own-earnings drop (RBLX) spill into no-news peers the next day, under TRAI's rules (no entries 09:30–10:00, 1 ATR stop / 2 ATR target, 60-min time stop, flat by 15:55); (3) how to rank mean-reversion longs vs continuation in the software cluster with one position per correlation cluster while short momentum is suspended. (1) overlaps this report's §6 calc (n 20, no edge); (2) and (3) feed S-119/S-120.

8. Strategy proposals (see strategy.md)

• S-119 (proposal): tag each short-momentum signal by cause (own-earnings reaction vs no-news/sympathy vs macro) and split the replay that way before any re-activation decision.
• S-120 (needs data): day-2+ behavior of no-news "pre-print sympathy losers" vs earnings flushes.
• S-106 gets a live case (PLTR/SHOP MR on print eve); S-085 gets cluster 5 this week.

9. Open questions

• Does the cluster-5 weakness extend into PLTR's reaction (Tue) and DDOG's print (Wed BMO)?
• TEAM consensus from a primary or reliable aggregator (not found).
• Construction-spending consensus (not found); the 10y reaction with a working feed.
• Is there any confirmed link between this morning's Anthropic S-1 amendment and software weakness? (Unverified; no source.)