Plan
desk: house-mneme date: 2026-07-28 forecasts: mneme: p_trade: 0.25 direction: none conviction: low trace: p_trade: 0.20 direction: none conviction: low
Desk Plan — House Mneme — 2026-07-28
Briefing reference: /Users/dadbot/Desktop/ClaudeBod/projects/dadbrain/Analysis/briefings/2026-07-28-pm.md PM: MNEME
Shared Market Read
- Event risk today: HIGH (FOMC tomorrow — July 29 rate decision + press conference; PCE July 30)
- Session character expected: Pre-FOMC compression / event-driven / choppy. The morning may see continuation of the AI/semi rout (day 2), but Consumer Confidence (10:00 AM) adds a mid-session catalyst. Afternoon likely consolidates as the market positions for tomorrow's binary event.
- VIX regime: 19.02 (MID). Both members' backtests showed healthy PF in VIX <20 (MNEME 2.39, TRACE 2.46) — vol level is not the binding constraint today.
- Key levels: SPY prior range 735.88–745.52 (below SMA20 at 746.62). DIA prior range — confirm at open. QQQ 682.13 below SMA20 710.45 — two-tier divergence is the dominant structural feature.
- Macro backdrop: AI/semi rout (NVDA financing, China lithography, KOSPI circuit breaker) is the dominant narrative. Dow futures +0.25% while Nasdaq -0.89% — extreme sector rotation. Pre-FOMC positioning adds a caution layer. Consumer Confidence at 10:00 AM is the only macro data point.
MNEME — SPY KNN
Phase 3 record: 3W/4L (42.9% WR, tracking dev holdout almost exactly). SPRT LLR +0.838 (CONTINUE, trending CONSISTENT). Session-character bifurcation persists: 3W/0L clean sessions, 0W/4L narrative-overlay sessions.
Today's call — LOW probability of a trade (p_trade 0.25). The setup environment is the worst possible for MNEME:
- Two-tier divergence is a documented failure mode — TEMPER's 2026-07-24 finding confirmed that rotation-driven rallies during Window A/B observation reverse by the execution window. The KNN library (2020-2024) has insufficient structurally-divergent sector-rotation sessions to correctly weight this fade risk.
- Narrative-overlay session — MNEME is 0W/4L on sessions dominated by a single narrative driver (geopolitical, chip rout, re-inflation). Today's AI/semi rout + FOMC eve is a textbook narrative-overlay day.
- Pre-FOMC compression is hostile to KNN — range-bound sessions with tight intraday ranges produce whipsaw. The KNN finds "similar" patterns in the bar structure, but the macro event context tomorrow invalidates the comparison for those historical matches.
- SPY direction is neutral — two-tier divergence means SPY lacks a clean directional signal. The S&P 500 is being carried by Dow components (industrials, airlines, energy beneficiaries) while tech drags. This creates a mixed-price-action tape that KNN will interpret as noisy rather than directional.
- Window B is the only realistic entry (12:00 entry, 14:00-16:00 execution) — the afternoon session is the most likely to consolidate in pre-FOMC positioning, which is precisely the KNN-hostile range-bound character. Window A (10:00 entry) faces a Consumer Confidence catalyst in its prediction window.
- Hard skip check: FOMC day is NOT today (FOMC is July 29) — so the pre-registered hard skip does not block. But the spirit of the skip (macro events create patterns with no historical precedent) applies to the eve as well, even though the letter does not.
If a signal fires anyway: The strongest-vote window is likely to be Window B (afternoon prediction, 12:00 entry), which enters into the pre-FOMC positioning window. Consider a tighter mental filter on max_similarity — any signal near the gate boundary (max_sim ~0.50-0.55, vote ~62-64%) should be viewed with skepticism as a "best of a weak set" match on a structurally unusual session type.
Setup present: Unlikely. The KNN may fire a Window B signal if the Consumer Confidence print at 10:00 produces a clean morning pattern, but the execution window is FOMC-eve afternoon — the worst possible character for pattern replication.
TRACE — DIA KNN
Phase 3 record: 1W/6L (14.3% WR). SPRT LLR -1.350 with -1.595 headroom to DEGRADED (-2.944). One more qualifying loss will likely breach the DEGRADED boundary. This is the most urgent risk flag in House Mneme.
Today's call — LOW probability of a trade (p_trade 0.20). TRACE should NOT trade today if it can help it:
- Two-tier divergence has been TRACE's specific failure mode — the 7th qualifying trade (2026-07-27) was a DIA long at 523.31 stopped out at 522.32 (-$142.28) on exactly this character: Dow futures positive, KNN signal fired, session opened high and faded. The pattern is consistent: KNN finds high-confidence similarity (max_sim 0.9245 on the 6th trade) on morning data, but the session's open-high-fade character reverses the entry direction. TRACE on DIA is theoretically the beneficiary of the two-tier divergence, but the KNN entry timing (morning observation → afternoon execution) has been consistently wrong.
- SPRT at DEGRADED boundary — a loss today pushes the LLR below -2.944, activating the DEGRADED protocol (desk PM review within 1 session, TEMPER conversation within 5). The pre-registered protocol is the correct response — no pre-emptive action — but this means every signal today carries disproportionate downstream cost. TRACE's own K=10 library may genuinely identify DIA patterns on two-tier divergence days (DIA has been the relative strength story), but the afternoon fade has killed every DIA trade so far.
- Pre-FOMC compression risks DIA as well — while DIA is holding up better than SPY/QQQ, the FOMC-eve character means even industrial names consolidate into the close. DIA's 30-stock blue-chip composition is the least sensitive to tomorrow's rate decision, so the KNN's historical patterns on pre-FOMC days may be more reliable for DIA than for SPY. This is a weak silver lining, not a reason to trade.
- The Dow's relative strength is a trap for TRACE's mechanics — DIA +0.25% futures suggests an open near or slightly above prior close. If the morning observation shows upward drift, the KNN library finds similar "upward" patterns. The issue is that the prediction window (afternoon) is FOMC-eve consolidation, not continuation. The KNN cannot distinguish between "up then more up" sessions and "up then fade" sessions using 3-feature vectors alone.
If a signal fires anyway: Only Window B (12:00 entry, 14:00-16:00 prediction) has any chance of working — Window A (10:00 entry) has Consumer Confidence at 10:00 in its prediction window, and Window C (14:00 entry) enters into the dead zone of FOMC-eve afternoon. A Window B signal with max_sim significantly above gate (>0.75) and vote > 70% would be the only configuration worth noting — and even then, the SPRT cost of a 7th loss is severe.
Setup present: Very unlikely. The KNN could theoretically fire a Window B signal if the Consumer Confidence data supports the Dow's relative strength theme, but the FOMC-eve afternoon consolidation risk overrides any pattern-based entry signal.
Plan Filed
- Filed: 2026-07-28 07:30 ET
- Frontmatter forecasts complete for every active member: yes
- GHOST: FROZEN (Phase 2) — not planned for