Jul 8, 2026

Regime & Verdict#

• Regime: UPTREND (marginal/chop-adjacent) · directional_tradable = TRUE · s1_standdown = FALSE
• Vol-state: negative-gamma / vol-expansion-prone — 2nd-moment conditioner only, no direction read
• Breadth: narrow, ~23–33% participation, Tech-concentrated divergence (index strength ≠ broad strength)
• Regime cap: both UW and Finviz independently call for HALF position sizes book-wide on this tape. This is a ceiling on top of (not a replacement for) the excess-implied size — it binds on the OI_FADE HIGH-lane name (which would otherwise cap at half anyway) and is a no-op on the starter-tier names.
• Event risk: June CPI 2026-07-14 + PPI 2026-07-15 (both Tier-1) fall inside the h10 window opened by any short taken this week (~07-08→07-22), and inside the S2 h5 window (~07-09→07-15) even more tightly than last week. Bank earnings cluster 7/14–15 also sits in this window. Applied as defined-risk-through-print on the idiosyncratic single-name OI_FADE/S2 theses (not macro-directional bets), consistent with 07-07 treatment — not an automatic −1 tier.
• On a beta-heavy, negative-gamma, narrow-breadth tape the gate stack matters more, not less (RESEARCH/50 §5.1: ungated additive-confluence realized −8.2%, 0-for-6).

Bottom line: a small, gate-thinned book — 4 sized names (1 half, 3 starter). OI_FADE contributes UDR (half, cleanest persistent build, CONFIRM) and AROC (starter, tail-capped down from the HIGH-lane half ceiling for a decaying build + offsetting revenue growth). NRG is OI_FADE's third clean-build candidate but is VETOed by fundamentals — watch-only, not sized. CGNX is a tape-beta trap (0.71 SPY beta) in a Tech-concentrated narrow-breadth regime — shorting it here is effectively fighting the one thing propping up the index — watch-only regardless of flow. S2 contributes PG and IBM at starter (both CONFIRM). The MOM_SHORT industrials cluster (SPCX/GGG/TXNM/PATK/FDXF, corr≥0.70, collapsed to one theoretical position) is held at watch — no fundamentals verdict on file, flagged back to the orchestrator for next-run spawn. S4's only candidate, AWR, fails the liquidity floor outright ($35.2M ADV < $50M) and is dropped — a correction to the Phase-B "Advisory/half" tag, which did not verify liquidity. MOM_LONG fired no named candidates — basket composition (Financials 21.9% + Real Estate 11.8%, a crowded rate-narrative cluster) is noted but not actionable without tickers, and the narrow-breadth caveat argues against chasing this leg even if it were.

Directional Book (excess-scored)#

Cut to watch/dropped (not sized, detail below): OI_FADE — NRG (VETO), CGNX (tape-beta trap), TDG/SLM/ENSG/INSP/SUI/NESR/VVX/ORA (single-day-print artifacts, not persistent). MOM_SHORT — SPCX/GGG/TXNM/PATK/FDXF (industrials cluster, missing fundamentals verdict). S4 — AWR (liquidity FAIL, dropped). S2 — MU (disqualified upstream, 61.6% two-sided flow, not a concentration event).

Vol Book (non-directional)#

• JNJ (primary) — earnings 2026-07-15 premarket, IVR 86.3, implied move 1.41% ($3.71). IV rich vs. realized (the "kink" the lane targets). Liquidity PASS ($263.40, $2.42B ADV). Small size, wide wings, two-sided tail; net-of-cost expectancy not independently recomputed — do not enter until it prices net-positive.
• NFLX (primary) — earnings 2026-07-16 postmarket, IVR 97.0, implied move 2.00% ($1.51). Liquidity PASS ($75.59, $3.73B ADV). Same structure discipline as JNJ.
• CTAS (primary) — earnings 2026-07-15 premarket, IVR 85.9, implied move 1.64% ($2.96). Liquidity PASS ($180.17, $392.6M ADV). Same structure discipline.
• ASML / ABT / STT (secondary) — IVR 95.3 / 93.7 / 81.9 respectively, implied moves 3.54% / 1.56% / 5.68%; earnings 7/15–7/16, all premarket except NFLX. All clear the $50M liquidity floor comfortably. STT is the one bank name that survives the "banks excluded IVR<80" screen (81.9 clears); it also sits in the 7/14–15 bank-earnings cluster called out in the event calendar — that cluster is the trade for STT, not a separate downgrade.
• 0DTE-VRP (SPY/QQQ): STAND ASIDE both books. VRP negative net-of-cost; GEX fully negative. No edge to sell premium in a negative-gamma tape — the unsampled left tail is not compensated here.

Watch / Stood-down#

• OI_FADE VETO (dropped entirely): NRG — beat-streak + insider buying + bullish AI-power catalyst directly contradicts the fade thesis.
• OI_FADE tape-beta trap (watch, not a fundamentals question): CGNX — 0.71 SPY beta; in a Tech-concentrated narrow-breadth regime, shorting the one mega-beta Tech name in the cohort is a QQQ-beta short in disguise (RESEARCH/50 §5.3 logic extended from the mega-cap OI_FADE case) — held at watch regardless of flow signal quality.
• OI_FADE single-day-print artifacts (not persistent, lane requires a 5-session build): TDG, SLM, ENSG, INSP, SUI, NESR, VVX, ORA. SUI additionally sits at a sub-threshold 0.68 correlation with UDR (noted, not collapsed — gate is ≥0.70 and SUI isn't live regardless).
• MOM_SHORT industrials cluster — held at watch, missing fundamentals verdict (flagged back to orchestrator for next-run spawn): SPCX, GGG, TXNM, PATK, FDXF. All five sit near their 52-week lows (SPCX $148.30 vs. 52w-low $147.11; GGG, TXNM, PATK, FDXF similarly compressed) and correlate ≥0.70 as a single industrials/rate-sensitive theme — collapsed to one theoretical position (cluster_id mom_short_industrials_cluster_0708) that will size at starter (the MOM_SHORT lane ceiling for a +0.81% mean edge) if and only if a fundamentals verdict clears next run. Not S1-crash-gated (s1_standdown=FALSE) — the block here is purely the missing veto check, not the regime.
• S4 sentiment-contrarian — AWR dropped, liquidity FAIL: close $83.58, 20d $-ADV $35.18M, below the $50M fail-closed floor. This corrects the Phase-B "Advisory/half" pre-size, which did not verify liquidity — per the fail-closed rule, a name that cannot be verified liquid is dropped, not merely downsized. With AWR out, the S4 lane contributes nothing today.
• S2 disqualified upstream: MU — 61.6% two-sided options flow, not a genuine one-sided DP concentration event; the mean-reversion mechanism this lane trades does not apply.
• MOM_LONG — no named candidates fired. Composition note only: the near-52w-high universe skews Financials (21.9%) + Real Estate (11.8%), a crowded rate-narrative cluster, and MOM_LONG is basket-only by construction (never per-name HIGH, tail-driven +2.35% mean). The narrow-breadth caveat argues this leg would likely stall if breadth ever expands — not a today problem since there's nothing to size.

Risk#

• Correlation clusters applied: (1) SUI–UDR at 0.68, sub-threshold, no collapse, and moot since SUI isn't live; (2) SPCX/GGG/TXNM/PATK/FDXF at ≥0.70, collapsed to one theoretical MOM_SHORT position, held at watch pending fundamentals. No other pairwise correlations ≥0.70 identified among today's sized names (UDR/AROC/PG/IBM span four different sectors — Real Estate, Energy, Consumer Defensive, Technology — with no shared theme).
• Tape-beta trap gate: CGNX excluded on the mega-cap-QQQ-beta logic extended to single-name narrow- breadth exposure — a reminder that this gate isn't only about literal OI_FADE-cohort mega-cap dominance; it's about any short whose realized behavior is closer to "short the index's one working sector" than "short an idiosyncratic flow signal."
• Event overhang: CPI (7/14) + PPI (7/15) sit inside every h10 window opened this week and inside the S2 h5 window even more tightly than 07-07's read (PPI now lands on day ~5 of a 5-day window that starts closer to today). Applied as defined-risk-through-print on UDR/AROC (OI_FADE) and PG/IBM (S2) — not an automatic downgrade, since none of these are macro-directional bets. Bank-earnings cluster 7/14–15 is the trade itself for STT (vol book), not a separate risk.
• Liquidity floor caught one real miss: AWR ($35.18M ADV) fails the $50M fail-closed floor outright — dropped, not downsized, correcting the Phase-B "Advisory/half" pre-size tag. All four sized directional names and all six vol-book names clear the floor with real margin (lowest is AROC at $94.5M ADV).
• Regime-cap discipline: the book-wide HALF ceiling from UW+Finviz is a no-op here except on UDR, which was already capping at half from its own HIGH-lane excess. No name was sized above half; the hard rule (never up-size above the excess-implied size) held throughout — AROC's tail-cap and AWR's liquidity drop are both downgrades, never upgrades.
• Missing-verdict discipline: the MOM_SHORT industrials cluster is a real, regime-appropriate, correlation-collapsed candidate sitting on the sideline purely because its fundamentals verdict hasn't been run yet — flagged back to the orchestrator, not silently dropped or silently sized.
• Hedge note: with only one short (UDR, half) and one starter short (AROC) against two starter longs (PG and IBM), net book exposure is modestly long and small. In a negative-gamma, narrow-breadth tape this is a deliberately thin book — consistent with invariant #3 ("most days have no directional edge — say so"), even though today does clear the bar for a handful of names.

---

Envelope: decision.json (schema v2, validated — see confirmation below). Conviction write-back persisted to conviction_2026-07-08.json for next run's adverse-flow exit check (UDR, AROC, PG, IBM). Regression gate not run tonight — no lane/threshold changes made.