House Echo — 2026-08-28

Desk Plan — House Echo — 2026-08-28

Briefing reference: /Users/dadbot/Desktop/ClaudeBod/projects/dadbrain/Analysis/briefings/2026-08-28-pm.md PM: ECHO

Shared Market Read

• Event risk today: MEDIUM — the Warsh Jackson Hole keynote (10:00 AM ET) is the day's single live risk. A Jackson Hole keynote is a speech, not a decision, so per calibration it rates MEDIUM (literal-HIGH is reserved for a same-day/next-day CPI/NFP/PCE/FOMC decision). But it is the week's central-bank pivot and a genuine coin-flip — CNBC Fed Survey: 45% expect a low-key no-guidance reiteration, 32% somewhat hawkish, 19% neutral; futures price ~30% hike / ~70% hold at 3.50–3.75%. No Tier-1 data today (next: NFP Sep 4). Final August Michigan consumer sentiment (10:00 AM, same window) is secondary/telegraphed. Firm-wide GO with a 10:00 AM event window.
• Session character expected: contained / pausing into the 10 AM policy speech; two-tier; direction UNKNOWN — a coin-flip day. Pre-market: futures flat-to-modestly-down, Dow holding a small gain, Nasdaq-100 dipping — a rotation toward value/Dow defensiveness ahead of the speech (the inverse of Thursday's tech-led extension). The AI/momentum leg (Thursday's leadership, NVDA +9%) is giving back pre-market — flag it as fragile, not pre-assigned as the weak side (5-for-5 coin-flip discipline: do NOT pre-assign direction).
• VIX regime: 14.49 — LOW (<18), flat at the multi-session easing low. This is the desk's weakest cohort. Across the LOW-VIX stretch ECHO's vol_ratio gate has blocked wide (0.55–1.19 vs the 1.2 floor) and SURGE's first-hour signal has routinely printed 0.10–0.25pp under its ±0.30% floor. Treat "no clean program / no impulse" as the base case for both.
• Key levels: SPY prior close 771.18 / prior range 767.16–772.32 / SMA20 768.06 (price above) / ATR(14) 4.77. QQQ prior close 721.20 / prior range 714.59–721.32 / SMA20 716.49 (price above) / ATR(14) 7.97. Both closed up Thursday (QQQ +1.39%, SPY +0.68%) and sit above their 20-day SMAs; yesterday was above-ATR for both (QQQ 1.26×, SPY 1.34×) but the trailing ATR stays elevated into today.
• Macro backdrop: Restrictive-Fed / elevated-yield undercurrent (10-yr ~4.69%, just off multi-year highs) ahead of the Jackson Hole pivot. A hawkish Warsh on that backdrop is a genuine downside catalyst; a low-key no-guidance reiteration is a potential non-event or relief. The AI leg's pre-market give-back is the day's flagged-fragile leg.

ECHO — dynamic universe, close-leg exhaustion fade (15:30 ET)

No trade is the base case, but today is structurally more tradeable for me than yesterday — SPY is a legitimate candidate. Unlike 8/27 (where SPY was ineligible because the catalyst was constituent-level NVDA earnings), today's catalyst is a broad-market, index-level policy event (Warsh keynote) — exactly the class my eligibility screen accepts. SPY is the default ECHO candidate on the watchlist.

• SPY — genuine candidate, direction UNKNOWN. The pre-market is a coin-flip (Dow up / Nasdaq down), so I will not pre-assign the fade direction. The event itself resolves at 10:00 ET — hours before my 15:30 entry — so the close-leg timing premise holds: whatever program the speech sets in motion runs the morning, and if it is a real one-way, gate-confirmed move (VWAP persistence, ADX ≥ 20, vol_ratio ≥ 1.2, exhaustion_score ≥ 0.80 at 15:20) it is exhaustible and fadeable CONTRA at 15:30. That is the if; the well-calibrated base case is the LOW-VIX vol_ratio wide-block — all month SPY has printed 0.55–1.14 vs the 1.2 floor (yesterday 0.848 on a +0.29% signal).
• MRVL — non-qualifying single-stock. The watchlist flags it as an ECHO single-name DOWN complement (beat-and-raise, -2% give-back after 187% YTD). The 7/24 INTC lesson governs: single-stock post-earnings reversals complete in the first ~90 minutes, not through a full institutional program — no close-window exhaustion to fade. Do not reach in.
• XLE — no active program; XBI — no active program. Oil ticking down on the Iran–Oman Hormuz de-escalation is context (with-sentiment energy driver), not a fresh sector-level directional program — and fading a strengthening own-driver sector is the 8/10 loss template regardless. XBI has no US-listed FDA catalyst today (the 177Lu-edotreotide PDUFA is a foreign sponsor). Neither is an ECHO candidate today.

Call: No trade is the base case (p_trade 0.22). Direction: none — the session is a coin-flip and I will not pre-assign the fade leg (this is only scored on traded days anyway). If the engine fires, sizing stays standard 0.75% per instrument / 2.0% session budget — no event sizing, because I hold no overnight risk (hard flat 15:58) and the event is resolved hours before entry. Invalidation: a forced SPY trade without all four gates passing at 15:20 (a big morning signal alone is not a setup); any single-stock AI-name reach (MRVL); any XLE/XBI reach. On VIX 14.49 the wide vol_ratio block is the expected, well-calibrated outcome.

SURGE — SPY/QQQ, VWAP pullback resumption (10:00–12:00 ET)

No trade expected — the 10:00 AM event window lands directly inside my entry window, which puts this in the catalyst/price-discovery class I am 0/2 on.

• This is NOT a hard-skip. NVDA reported after Wednesday's close; per the event-day dating convention (earnings date + next-calendar-day adjacency), the QQQ_EARN window was yesterday (8/27), not today. So the engine should evaluate normally. That makes the stand-down a deliberate one, not a calendar artifact.
• The structural block: the event overlaps the entry window. My qualifying first-hour signal is computed on 09:30–10:00 — i.e. pre-Warsh drift, which is unreliable as a trend-establishment read when a policy speech fires at exactly 10:00. And any move that follows the speech is price-discovery, not a clean three-phase trend → the 7/24 learning is direct that the three-phase structure is invalid on catalyst/event sessions. The 8/24 QQQ short (a trend-qualified first move on a divergent tape, whipsawed through VWAP for -$108.75) is the freshest evidence the session-character quality bar is real. On a coin-flip policy day whose direction is explicitly not pre-assignable, there is no with-trend thesis to buy.
• VIX 14.49 LOW is my weakest cohort — the first-hour signal has routinely printed 0.10–0.25pp under its ±0.30% floor on this tape (yesterday SPY +0.28%, QQQ trend-qualified but no VWAP-touch retrace). Even absent the event, no impulse is the base case.

Call: No trade expected (p_trade 0.12). Direction: none. Sizing: standard $187.50 / single position / hard flat 14:00 — but not invoked; no adjustment is meaningful on a stand-down day. Invalidation: if, against the event-window read, a clean three-phase structure genuinely forms (a trend-qualified first move after the speech resolves that retraces to VWAP with contracting volume and a resumption bar within 10:00–12:00), it would be a direction-neutral mechanical entry — but that is not the forecast; the base case is no_signal or the event-day price-discovery denial.

Plan Filed

• Filed: 2026-08-28 04:33 PDT
• Frontmatter forecasts complete for every active member: yes

No trades taken.

House Echo Desk Reflection — 2026-08-28 (Friday)

Cadence rule header: Weekly/Friday or same-day-on-trade. Today is Friday 08-28 — a scheduled reflection day. Period covered: 2026-08-25 → 2026-08-28 (since the last desk reflection, 2026-08-24). Four sessions; neither member traded on any of them.

Plan reference: desks/house-echo/plans/2026-08-25..28-plan.md EOD briefing: dadbrain/Analysis/briefings/2026-08-28-eod.md

Shared Read (what the tape actually did over the period)

A four-session LOW-VIX stretch (VIX 14.43–15.85, LOW <18) with two hard-skip binary sessions bookending the week and two "no impulse" sessions inside. The week's dominant thread was the hawkish-Warsh Jackson Hole resolution: the 08-28 keynote (10 AM ET) was the day's catalyst, resolving hawkish (rate-hike door open, 2-yr +~12bp, 10-yr ~4.73%) into a choppy/ranging two-tier down close — QQQ -0.66% vs SPY -0.25%, a large intraday range (SPY 1.38× ATR) but a small close and flat last hour (no clean directional drift). 08-26 was a firm-wide NO-GO (July PCE at the open, NVDA Q2 after the close); 08-25 and 08-27 were contained/no-impulse days around the binary cluster. Across all four sessions the desk's LOW-VIX "no clean program / no impulse" base case held literally: neither expression of Family 2 found a qualifying setup on any day.

Desk Performance Summary (period)

Desk net P&L this period: $0.00 (no trades). Running desk book is the sum of the two members' cumulative books (ECHO -$206.75 + SURGE +$602.83), not a pooled number. Both SPRT trajectories healthy and comfortably mid-channel — neither approaching a boundary this period.

Period Outcome Table (per member, per date)

Plan calibration this period was clean on both members — every forecast was no-trade at low p_trade and no-trade occurred (Brier 0.0036–0.09, all well-calibrated; no under/over-forecast, no plan-vs-engine divergence). The 08-26 PCE NO-GO was honored by both stacks.

Per-Member Binding Gate Analysis

ECHO — binding gate: vol_ratio ≥ 1.2 → regime, not calibration

The close-leg exhaustion fade was blocked on vol_ratio ≥ 1.2 on every non-skip session of the period, and blocked wide on essentially every leg: SPY 0.550 / 0.848 (and a morning_signal 0.07% block on 08-28), XLE 0.868 / ADX-16.8, XBI 0.329 / 0.800 / 1.146. The closest graze was XBI 1.146 on 08-25 (0.054 under the floor, on a +2.27% morning signal) — a moderate near-miss, but not accumulating into a repeated-threshold pattern: XBI's vol_ratio has ranged 0.33–1.19 across the month (1.195 on 8/21, 1.082 on 8/24, 1.146 on 8/25, 1.010 on 8/26, 0.329 on 8/27, 0.800 on 8/28) with no consistent cluster, and XBI has not re-qualified since 8/20. On a LOW-VIX tape (14.43–15.85) no exhaustible institutional program forms — there is nothing for the close-leg fade to fade, so the vol_ratio gate correctly blocks wide. This is the same regime reading the 8/14 and 8/24 reflections established; this period is a clean confirmation, not a new signal.

No fix, no tweak. The gate is positioned correctly for the regime. XBI sub-book remains n=2, WR 0.0%, PF 0.0 — not triggered (min_n=15, threshold 0.5); the XBI exclusion-review condition is nowhere near. The standing plan/engine enforcement gap stayed latent this period (the engine agreed by blocking — no high-conviction stand-down was ever stressed against a qualifying signal), so it was neither resolved nor re-stressed.

SURGE — binding gate: the complete three-phase structure → regime, not calibration (with a calibration-flavored sub-thread to monitor)

For with-trend SURGE the binding gate is the complete three-phase structure (first-hour signal → VWAP pullback → resumption), expressed through the ±0.30% first-hour signal floor and the pullback_resumption formation. This period's no-trades resolve through both:

• First-hour signal floor (±0.30%) blocked on 08-25 (SPY +0.25%, 0.05pp under), 08-27 (SPY +0.28%, 0.02pp under), and 08-28 (SPY +0.12%, QQQ -0.12%). The 08-27 SPY 0.28% and 08-25 SPY 0.25% are genuine near-grazes — within 0.02–0.05pp of the floor.
• Pullback_resumption was the operative block on 08-27: QQQ printed a trend-qualified +0.84% / ADX 25.8 first move — the signal cleared — but never retraced to VWAP within the window, so the structure never completed.

The load-bearing read: even where the first-hour signal did clear (QQQ 08-27), the pullback didn't form — and where the pullback did form (the 08/24 QQQ trade, covered in the prior reflection), it whipsawed through VWAP for -$108.75. On a LOW-VIX sub-impulse tape, no clean three-phase structure is available in any combination — this is structural absence (regime), not a mis-tuned gate. No fix warranted.

Calibration-flavored sub-thread to carry into the standing conversation: the ±0.30% first-hour signal floor is being repeatedly grazed on SPY — +0.25% (08-25), +0.28% (08-27), and earlier +0.29% (08-05, 08-21) — all inside one LOW-VIX regime on the same instrument. That is a recurring close-graze cluster against one numeric threshold. It does not change the regime verdict this period (the operative block is the three-phase completion, not the signal floor alone), but it is a named data point for SURGE's standing ADX/session-character-filter conversation (open TEMPER item): if the SPY 0.28–0.29% graze cluster persists past the LOW-VIX stretch or ever blocks a genuinely trend-qualified day, the floor's positioning is worth a TEMPER look. Flag for learnings.md (desk-blind).

Cross-Member Synthesis

Narrative finding: ECHO's vol_ratio gate and SURGE's first-hour signal are two expressions of the same LOW-VIX "no impulse" phenomenon — the desk's two complementary sides of Family 2 both stand down for the same underlying reason: on a VIX 14.43–15.85 tape no institutional program forms (ECHO) and no clean trend-impulse-then-pullback forms (SURGE). The desk's base case across the whole LOW-VIX stretch — "no clean program / no impulse" — held literally on all four sessions. Both members' gates are doing their jobs; neither is being repeatedly grazed on a threshold it was built to hit (ECHO blocked wide; SURGE's operative block is the structural pullback, not the signal floor). The plan/engine gap stayed latent for ECHO; SURGE's only recent trade (08/24, a loss) was a session-character quality issue, not a gate issue. This is the desk's well-calibrated stand-down side, now extending its clean no-trade forecasting record.

Open TEMPER Items (Carried Forward)

• ECHO — plan/engine enforcement gap (wiring plan-level / driver-sector stand-downs into gate logic, or excluding high-conviction XBI fades): flagged 08-19/08-20, remains open — latent this period, neither resolved nor re-stressed. No action this session.
• ECHO — XBI exclusion review (WR < 50% over 15+ live trades): sub-book n=2, WR 0.0% — NOT triggered (needs n≥15). Carried forward.
• SURGE — ADX re-tune / session-character-filter conversation (moderate-trend ADX 22–30 outperforms strong-trend; the 08/24 two-tier-divergent-tape loss is the freshest evidence the session-character quality bar is real): carried forward. This period adds the recurring SPY first-hour-signal graze cluster (0.25–0.29% vs ±0.30%) as a data point for that conversation.
• SURGE — event-day skip diagnostic (Tier 1 reconfirmation item): resolved 2026-07-17 (MIXED), no calendar deadline — carried, not open.
• SURGE — 2023 PF 0.95 watch: rolling 6-month PF monitor, no trigger this period.
• SURGE — event-day skip disposition (QQQ_EARN adjacency): the 08-27 NVDA-earnings QQQ_EARN hard-skip was correctly dated to 08-27 (next-day adjacency), and the 08-28 session was correctly evaluated normally (not a skip) — both consistent with the convention.

Desk Summary for Orchestrator

Four-session LOW-VIX stretch (VIX 14.43–15.85), zero desk trades, zero desk P&L — the desk's well-calibrated stand-down side held literally: ECHO blocked on vol_ratio ≥ 1.2 (widest 0.33, closest graze XBI 1.146 on 08-25, not accumulating) and SURGE on the three-phase structure (first-hour signal ±0.30% grazed at 0.25/0.28% on SPY; QQQ signal cleared 08-27 but no VWAP pullback formed) — both read regime, not calibration (no impulse/no program on LOW VIX). No unilateral tweak; the only action is to carry SURGE's recurring SPY first-hour graze cluster (0.25–0.29%) into its standing ADX/session-character TEMPER conversation as a named monitoring data point (Flag for learnings.md). Both SPRT trajectories CONTINUE, comfortably mid-channel (ECHO +0.184, SURGE +1.278); no boundary risk. Plan calibration clean across the period (all no-trade forecasts honored, Brier 0.0036–0.09). Open TEMPER items (ECHO plan/engine gap, XBI review n=2 not triggered, SURGE session-character conversation) carried forward, none advanced or re-stressed.