Reversal Desk — 2026-09-04

Desk Plan — Reversal — 2026-09-04

Briefing reference: /Users/dadbot/Desktop/ClaudeBod/projects/dadbrain/Analysis/briefings/2026-09-04-pm.md PM: Wicker (persona; own stack retired 2026-07-14 — plan covers active members Null + Meridian only)

Shared Market Read

• Event risk today: HIGH — Tier-1 NFP (August Employment Situation) prints 8:30 AM ET. Firm-wide NO-GO per today's watchlist (CPI/NFP/PCE check: NO-GO — NFP today). Both members' event-day hard-skip fires — the desk is flat by design. The print is the week's pivot and the read-ahead into the Sept 15–16 FOMC: consensus clusters at ~+55K (sources +53–58K; MUFG above at +70K) after July's surprise −23K; U/R expected 4.1%, AHE +0.3% m/m. The market has priced a genuine ~50/50 Sept hike/hold coin flip and enters positioned for a soft print (relief continuation after Thursday's Waller-driven +1% rally). The binary: soft NFP → locks the hold, extends the rate-sensitive relief leg; hot NFP / hot wages → re-ignites the Warsh hike leg and reverses Thursday's tape. Directionally binary — structure will not form until after the print. The standing US–Iran / Hormuz geopolitical clock (oil at six-week highs, easing this morning) is a weekend-gap risk into Tuesday's open (Labor Day Mon Sep 7 closed) more than an intraday driver today — Thursday proved the dovish-Fed/yields leg can outrank a concurrent oil rise.
• Session character expected: news-driven / direction-unknown-until-8:30, then trending — NOT a contained pre-positioning day. Expect a post-print trending leg in the direction the print resolves rate expectations; a data surprise plus a live rate signal can break the tape at 1.6× ATR or more (EOD directive #2, carried). Do NOT pre-assign index direction (direction-refusal discipline now 9-for-9 through 09-03).
• VIX regime: 14.18 (LOW, <18) — both members' strongest historical cohort, but moot today: the Tier-1 event gate skips both stacks before any signal work. Cohort matters again Tuesday (Sep 8) after the long weekend.
• Key levels (authoritative = pm briefing, Alpaca/IEX):
• QQQ: prior close 717.67 (09-03), prior-day range 709.73–718.89, 20-day SMA 717.71 (at SMA20), ATR(14) 8.17. Implied Null gap band on QQQ: 0.1× ≈ ±0.82 pts, 0.3× ≈ ±2.45 pts, 0.7× ≈ ±5.72 pts — but pre-market print is thin/none, confirming gap only at the open, and the print resets it anyway at 8:30.
• SPY: prior close 773.12 (09-03), prior-day range 767.47–774.01, 20-day SMA 769.16 — above SMA20, ATR(14) 5.48.
• Meridian sweep levels (for reference only, no signal work today): QQQ prior high 718.89, SPY prior high 774.01, IWM prior high per prior leg log. Post-print range is the real input — a wide NFP day likely blocks Monday's volatility gate (prior_range vs 1.25× ceiling).
• Macro backdrop: the August labor print decides which Fed regime the market carries into the Sept 15–16 FOMC — a two-sided 8:30 coin flip on a LOW-VIX tape that has recouped its pullback (SPY back above SMA20, QQQ at it) off Thursday's +1% Waller relief rally. One sentence: a genuinely binary NFP session where pre-printing any index direction or contained-range assumption is the error the standing discipline exists to prevent — and where this desk's two reversal stacks stand down by design.

Null — MNQ/MES (gap fade / FVG)

Setup present: NO — NFP hard skip. No trade expected. This is the highest-conviction no-trade call I can file. Null's tier-1 macro filter (is_event_day(trade_date, tier_filter=2) — FOMC/CPI/NFP/PCE) is code-enforced on the live single track and NFP is in the calendar (Bucket A, verified). The last NFP day (07-02) logged no_trade:event_day:NFP_tier1; today will be the same row.

Why the skip is correct, not just rule-bound: NFP-set opens are the breakaway-gap population the thesis explicitly excludes — the gap at the cash open is a fundamental repricing of the index (labor data + Fed-path optics), not a thin-overnight manufactured imbalance awaiting NY-open correction. The v2 rationale is explicit: "News catalysts at 8:30 create breakaway gaps and session-level disruption that invalidate the reversion thesis." Even a modest print produces (a) a gap that can exceed the 0.7× ATR ceiling, and (b) pre-market NQ volume that spikes past the 2× gate — institutional commitment to the gap direction, the second independent stand-down condition. In a vacuum the implied shape is unreadable: the pre-market print is thin/none per the briefing and the futures estimate is NFP-eve positioning noise; the print resets the gap at 8:30 and the filter does not care about the shape. NFP is a hard skip regardless of what the auction prints.

Sizing: N/A — zero risk, flat. No event-risk reduction to compute; the 1 MNQ / 0.5%-risk framework stays dormant. The 11:00 hard exit, 1.5% daily loss limit, one-position rule — all dormant.

Context/calibration: SPRT CONTINUE (n=12, W11/L1 through 09-03, LLR +1.826; cumulative +$1,765.99). Yesterday (09-03) was his first real loss after eleven straight wins — a mechanism-faithful in-band gap-up short (0.159×) stopped at 09:41 on a clean 1.62×-ATR trend-up day (-$192.61). That reflection classified it regime, not calibration — a one-way grind-up fade loss the mechanism (which has no trend/session-character filter by design) is built to average over at a high but not 100% WR; N=1 is within the cleared ~82% expectation and nothing gets recalibrated. Today's NFP stand-down is orthogonal to that loss — it is the event gate doing its job.

What Null is watching to learn (not to trade):

1. Cash-open gap size vs the 0.1–0.7× band. If the NFP open prints inside the band, that is one data point on how often a hard-skip day would have been tradeable by shape — the skip's cost in firing-rate terms, part of the 2026-WR-watch story (12 of the first 30 consumed).
2. Pre-market NQ volume vs 2× the 5-day average. An NFP morning is expected to spike it — the second gate that would independently have stood him down (overlap, not double-count).
3. VIX after the print. Whether the print keeps VIX LOW (<18) or breaks it out of his strongest cohort is the regime input for the next real window (Tuesday Sep 8, after Labor Day).

Invalidation (standing, dormant): none live today — the session never reaches signal work. Gates resume Tuesday: gap outside 0.1–0.7×, trend filter on gap-up 0.3–0.7× above SMA20, pre-market volume > 2×, no clean 1-min FVG by 10:30.

Meridian — QQQ/SPY/IWM (kill-zone prior-high sweep + RSI(2) ≥ 88, bearish only)

Setup present: NO — NFP hard skip. No trade expected. Meridian's event-day gate is code-enforced on the live track since the 2026-07-14 fix (paper_trade.py calls is_event_day() and returns before the per-symbol data fetch) — the CPI-day gap that reflection exposed is closed; NFP is Tier-1 in the calendar, so the gate fires pre-open. The desk forecast is p_trade 0.0, direction none, conviction high.

Why the skip is correct beyond the rule: the kill-zone sweep needs a clean two-sided tape — a push-up into the prior high, a spike-and-reject with RSI(2) ≥ 88 exhaustion, an 18–32 ADX band, a qualifying bearish FVG within the window. An NFP morning is a volatility spike with print-set structure: the open is a price-discovery auction, not the manipulated push into liquidity the sweep trades; the post-print tape is headline-driven and two-sided in ways that produce false sweeps, not exhaustion sweeps. A soft-print relief continuation (the positioned-for base case) is the worst backdrop for his bearish-only short — it pushes price up through the prior high and holds, the exact "grind-up gives me nothing to fade" shape that correctly stood him down on 09-03 (QQQ 709.77 → 717.67). Sweep levels QQQ 718.89 / SPY 774.01 / IWM prior high are for reference only; no signal work today.

Sizing: N/A — zero risk, flat. The $250 risk-per-leg / cap framework stays dormant; the 11:30 hard close and 2-loss daily cap are moot.

Context/calibration: SPRT CONTINUE (n=3, W2/L1 through 08-31, LLR +0.222; cumulative +$385.22; spy_leg n=1 WR 100%, iwm_leg n=0). 09-03 was a correct stand-down on a clean trend-up session — ADX precheck below the 18 floor on QQQ (13.3) and SPY (16.2), IWM blocked on a 0.3% volatility graze (3.16 vs 3.15) — the mirror of 08-31's ceiling graze, both one-way-tape refusals, classified regime not calibration. Today's NFP stand-down is the event gate, independent of that streak.

What Meridian is watching to learn (not to trade):

1. The post-print range. If today prints wide (the NFP base case), Tuesday's volatility gate is likely blocked again (prior_range vs the 1.25× ceiling) — the stand-down streak's next chapter and live firing-rate data for the IWM expansion (19/yr target).
2. 15-min ADX through the print. A strong post-print trend lifts ADX toward/over the 32 ceiling (or a low-ADX grind keeps it under the 18 floor); where QQQ/SPY land sets Tuesday's precheck. Log the read even though no trade.
3. Would the sweep structure have formed? Observational only: does the first post-print push reach the prior high and reject (exhaustion), or is the session a one-way auction? One live NFP observation per event day accumulates toward the event-day behavior the backtest can't easily see in daily-bar data.

Invalidation (standing, dormant): none live today — the session never reaches signal work. Tuesday resumes standard: vol gate > 1.25×, ADX outside 18–32, no sweep, RSI(2) < 88, no FVG within the window, hard close 11:30.

Plan Filed

• Filed: 2026-09-04 (pre-market)
• Frontmatter forecasts complete for active members only (Null, Meridian): yes. Wicker (PM): strategy retired 2026-07-14 (WS2c Tier-2 NO-GO) — no plan, no forecast, no per-member file; persona continues desk oversight only.
• Firm-wide NO-GO confirmed on a Tier-1 NFP day (August Employment Situation, 8:30 AM ET) — directionally binary, structure won't form until after the print. Event risk HIGH. Both members' event-day hard-skips fire pre-open; both p_trade 0.0 / direction none / conviction high — the code-enforced stand-down the desk files on every NFP/FOMC/CPI/PCE day. No broad-index direction pre-assigned (direction-refusal discipline, 9-for-9 through 09-03). Reversal stacks hold by design on macro-event days.
• Reconfirmation ledger: no open item gates today. Meridian's IWM expansion CLOSED 2026-08-20 (strategy_v7 adopted); standing ~6-month periodic re-backtest decay check next due ~2027-02-20. Wicker's feed/session re-run remains the desk's highest-priority open reconfirmation (unchanged, not a today blocker).
• Next real trading window for both members: Tuesday 2026-09-08 (Labor Day Mon Sep 7 closed).

No trades taken.

Reversal Desk Reflection — 2026-09-04

PM: Wicker (persona; own stack retired 2026-07-14 — desk covers active members Null + Meridian only) Briefing reference: dadbrain/Analysis/briefings/2026-09-04-eod.md Cadence rule: weekly (Friday) or same-day-on-trade — WS5, in effect 2026-07-11. Today is the Friday weekly floor. Period covered: 2026-09-04 (single session — prior desk reflection 2026-09-03 covered 09-01→09-03). Today is a Tier-1 NFP event day; both members hard-skipped pre-open.

What Happened — Desk Summary

A clean Tier-1 event-day stand-down — the desk was flat by design on August NFP. The print was hot — +162K vs ~+55K consensus, a ~3× upside surprise with prior months revised up (July −23K → +21K), yet the tape resolved contained, not trending: a two-tier session where the broad/value tape sagged on the yield jump (~+6bp toward 4.53%) while mega-cap tech/momentum carried the Nasdaq-100 positive (QQQ +0.20% to 719.10; SPY −0.38% to 770.18). VIX stayed LOW 14.53; realized vol never lifted. Both indices closed sub-ATR (SPY 0.75×, QQQ 0.64×). Net desk paper P&L $0.00 — no member's numeric gate was even reached, because the event-day hard skip fired pre-open for both stacks before any per-symbol data fetch or signal work.

• Null — NO-TRADE: no_trade:event_day:NFP_tier1. $0.00. His is_event_day(trade_date, tier_filter=2) Tier-1 filter (FOMC/CPI/NFP/PCE) closed the session pre-open. SPRT unchanged: n=12 W11/L1, LLR +1.826 CONTINUE, cumulative +$1,765.99, WR 91.7%. The calendar confirms NFP in Bucket A; the code-enforced gate is the mechanism, not a judgment call.
• Meridian — NO-TRADE: no_trade:event_day:NFP_tier1. $0.00. His paper_trade.py event-day gate (the CPI-day gap fixed 2026-07-14) returned before the per-symbol data fetch. SPRT unchanged: n=3 W2/L1, LLR +0.222 CONTINUE, cumulative +$385.22, WR 66.7%; spy_leg n=1 WR 100%, iwm_leg n=0.

Both desk forecasts were p_trade 0.0 / direction none / conviction high → brier 0.0 on both calibration rows (confirmed in plan-calibration.csv, 2026-09-04). The desk's event-day hard-skip architecture — the same code-enforced stand-down that produced event_day:NFP_tier1 on 07-02 and 08-07 — worked exactly as designed. No SPRT boundary crossed (no trades).

Session Character Read

The morning narrative graded PARTIAL — and for this desk, that miss is structurally irrelevant to decisioning because both members were flat by the event gate regardless. The AM's two errors under the realized hot branch — (1) pre-assigning rate-sensitive growth/QQQ as the vulnerable leg when mega-cap momentum carried it positive, and (2) expecting a trending higher-range resolution when pre-pricing + the pre-holiday Friday faded the print into containment — never reached a live gate. What matters for this desk is the observational read the event day produced for the next real window (Tue Sep 8, after the Labor Day long weekend), which is where the desk's attention now sits.

Desk Performance Summary

Desk net paper P&L: $0.00 (both members flat by the event gate). Both SPRTs CONTINUE — no boundary crossing.

Per-Member Binding Gate Analysis

Null — NO-TRADE; binding gate the Tier-1 event-day hard skip (pre-open, code-enforced) — NEITHER regime nor calibration today: no numeric gate reached, no near-miss generated

Null's binding gate this session was not a setup gate at all — it was the is_event_day(trade_date, tier_filter=2) hard skip, which fired pre-open and returned before any signal work. No gap_ratio, trend_alignment, or pre-market-volume gate was evaluated, so today produced zero near-miss data — there is nothing to read as regime or calibration because no threshold was grazed. This is the pre-registered event-day stand-down (the same code-enforced skip as 07-02 and 08-07), functioning exactly as designed.

• The one observational data point worth recording (his own plan's watch item #1): the QQQ cash-open gap printed in-band at ~0.18× ATR (open 719.16 vs prior close 717.67, / ATR 8.17) — a shape the gap filter alone would have admitted. But this does not call the skip into question: an NFP-open gap is a print-set fundamental repricing (the population the v2 thesis explicitly excludes), not the thin-overnight manufactured imbalance the mechanism fades, and pre-market NQ volume would spike past the independent 2× gate. It is one N=1 data point toward quantifying the skip's firing-rate cost, exactly the question Null's plan flagged — not evidence the skip was wrong.
• Context unchanged and healthy: the first-loss watch (09-03, trend-up-day fade stopped −$192.61 at n=12/91.7%) stays regime, not calibration — the window's near-misses remain opposite-sided gap-magnitude blocks (09-01 oversize 0.933, 09-02 sub-floor 0.081), not one-threshold repetition. SPRT CONTINUE with wide cushion (LLR +1.826, +1.119 to the +2.944 boundary). No gate tweak, no TEMPER conversation. Watch whether trend-up-day fade losses cluster (N≥3) before any session-character-filter discussion; the mechanism has no trend filter by design and this is N=1.

Meridian — NO-TRADE; binding gate the Tier-1 event-day hard skip (pre-open, code-enforced) — NEITHER regime nor calibration today: event gate preempted all signal work

Meridian's binding gate was likewise the event-day hard skip, firing pre-open before the per-symbol data fetch. No ADX precheck, volatility ceiling, sweep, or RSI(2) gate was evaluated — no near-miss was generated, so today is neither regime nor calibration evidence on any numeric threshold. It is the code-enforced Tier-1 stand-down (the 2026-07-14 CPI-gap fix working as intended).

• Observational read that matters for Tuesday: the plan's watch item #1 assumed a wide NFP day would likely re-block Tuesday's volatility gate. That default did NOT materialize — the session resolved contained/sub-ATR (QQQ true range 5.24, SPY 4.12), so Tuesday's prior_range (today's range) will be modest against the 1.25× ceiling rather than auto-blocked. That is a mildly constructive input for Meridian's next real window.
• Observational read on the sweep level: QQQ did trade up through its prior high intraday (day high 721.84 vs prior high 718.89) and closed above it and held (719.10) — the same "blow-through-and-hold / grind-through" shape that correctly stood him down on 09-03, not a spike-and-reject exhaustion sweep. On a day QQQ was the positive two-tier leg carried by independent mega-cap momentum, Meridian's bearish-only short had no favorable footing regardless of the gate. Confirms the event-skip and the bearish stance were aligned.
• Context unchanged: the standing ADX-band conversation stays regime — the 09-03 low-side mirror (QQQ 13.3, SPY 16.2) vs 08-31's ceiling graze (32.5) are distinct one-way-tape refusals, each N=1, no repeated boundary graze. SPRT CONTINUE (LLR +0.222); iwm_leg n=0. No gate tweak, no TEMPER conversation.

Cross-Member Synthesis

Narrative finding: both members produced the identical, correct outcome on the week's marquee event day — flat, by the code-enforced Tier-1 hard skip, with clean brier 0.0 forecasts and no SPRT movement. This is the event-gate architecture working as designed, not timidity (the firm's Tone-Independent Skip principle, 08-12). The desk-level substance of the day is forward-looking: the hot-print-but-contained resolution (contrary to the AM's trending expectation) is itself the observation — it means the hawkish NFP repricing entered the Labor Day weekend without generating the wide post-print range that would have carried over and re-blocked Tuesday's gates for both members. Both desks resume fresh on Tue Sep 8. No member gate threshold was grazed today; no unilateral tweak; nothing to escalate.

Open TEMPER Items (Carried Forward)

• Null — first live loss + 2026-WR watch (carried, N=1, monitoring only). 09-03 trend-day fade loss (−$192.61) broke the 11-0 streak. No action at N=1 — SPRT CONTINUE (LLR +1.826), WR 91.7%, 12 of the first-30 watch consumed. Watch whether trend-up-day fade losses cluster (N≥3) before any session-character-filter conversation. NFP event day today produced no new data point on this item.
• Null — NFP hard-skip firing-rate cost (NEW, N=1, observational). 09-04 NFP cash-open gap printed in-band (~0.18× ATR on QQQ). One data point toward quantifying the event-skip's cost; the skip remains principled (repricing gap + pre-market volume spike would independently stand him down). Track across future NFP/FOMC/CPI days.
• Meridian — ADX-band high/low refusals (carried, each N=1). 08-31 ceiling graze (32.5) and 09-03 low-side mirror (QQQ 13.3 / SPY 16.2) — distinct one-way-tape refusals, no repeated boundary graze. Regime, no action.
• Meridian — iwm_leg sub-book monitoring (carried). n=0. No event-session data today; next real watch Tuesday.
• Meridian — standing ~6-month periodic re-backtest decay check (carried). Next due ~2027-02-20.
• Wicker — backtest re-run under corrected feed/session config (carried). Standing highest-value reconfirmation; unchanged (own stack retired).

Reconnaissance Ledger

• Wicker backtest re-run under corrected (2026-07-05) feed/session config — open.
• Standing rule: no parameter re-tuning / gate loosening to manufacture throughput — unchanged; neither member warrants any.

Desk Summary for Orchestrator

Both active members (Null, Meridian) hard-skipped the Tier-1 August NFP session pre-open via their code-enforced event-day gates — $0 P&L, clean brier-0.0 forecasts, SPRTs unmoved (Null n=12 LLR +1.826 CONTINUE at 91.7%; Meridian n=3 LLR +0.222 CONTINUE). The load-bearing read is neither regime nor calibration — no numeric gate was reached today; the event gate is the binding gate and it did its job. The substantive finding is forward-looking: the hot +162K print resolved contained/sub-ATR (not the AM's trending expectation), so the hawkish repricing enters the long weekend without the wide range that would have auto-blocked Tuesday's gates — both desks resume fresh Tue Sep 8 after Labor Day. Standing context (Null's N=1 first-loss watch, Meridian's N=1 ADX refusals) stays regime; nothing to escalate, no unilateral tweak.

Reflection Filed

• Filed: 2026-09-04 (EOD)
• Per-member compatibility files written: agents/null/reflections/daily_2026-09-04.md, agents/meridian/reflections/daily_2026-09-04.md (wicker retired — no per-member file)
• Next-session event risk (2026-09-08, Tuesday — Labor Day Mon Sep 7 closed): LOW/MEDIUM (no Tier-1 in the immediate window; the next named Tier-1 is CPI on Fri Sep 11, which now gates the Sept 15–16 FOMC read after today's firm-labor print). Live weekend-gap risk into Tuesday's open: US–Iran / Hormuz escalation with oil at ~six-week highs over a 3-day holiday — the same geopolitical clock that is VESPER's designed overnight carry risk and a live gap input for both reversal members' opens.