The Nightcap White-Cap

Wednesday, 06-24-2026

Evening market read · validation & recap

The Milkman
OuroTaurus
Built 5:16 PM ET · 2026-06-24 · validates the 6/24 regular-session close (Massive /v3/snapshot, COALESCE(session_close,session_price); cross-checked vs /v2/aggs/grouped — exact match) anchor: system-reminder 2026-06-24 == Bash date (agree) · static — regenerate to refresh

01Session Scorecard — 2026-06-24

5 setups in the open backlog; 3 finalized this run0 FIRE / 1 VOID / 2 MIXED. The session was a faded bounce: green at midday (SPY +0.72%, IWM to a fresh record, equal-weight beating cap-weight) then the entire move round-tripped to a flat close (SPY 733.24, -0.05%). One number decided all three reads — SPY closed above the 731 50-day but below the 736.92 VWAP. Two weekly Sundae setups remain interim (arm on Micron FQ3 tonight / PCE Thu).

SetupOutcomeEvidenceΔ-ATRActual
LRB L SPY,RSPmm-260624-LRB-SPY-L MIXED Level-rejection-at-bottom long: the 50-day ~731 held (low 730.84 wick, closed 733.24 above) and the defensive/equal-weight exposed basket closed green — RSP +0.71%, XLP +0.86, XLV +0.77, XLU +1.04 — with yields down (TLT +1.37%, the y10>4.60 kill never threatened); Micron AMC after the close, no override. But the predicted oversold reclaim did not materialize (SPY -0.05%, faded from the +0.72% midday high) and confirmed liquid-name A/D rolled to 0.73 (<1, the named breadth-confirmation kill). Swing thesis intact (close >> 713.39 stop). +0.19 733.24
LRB L SPY,IWM,RSPmd-260624-1209-LRB-SPY-L VOID Midday VWAP-hold continuation long — invalidated by its own primary kill: SPY lost the 736.92 VWAP, closing 733.24, below the 735.28 stop (close-based -2.26R); and IWM failed back under the 299.49 breakout (296.69), the second kill. The broad/cyclical basket closed green (XLY +1.15, XLI +1.16) but the specific intraday continuation (hold VWAP into the close) failed on the late-day fade 739.95 → 733.24. +0.31 733.24
FBR S IWM,QQQmd-260624-1209-FBR-IWM-S MIXED Failed-breakout / negative-divergence fade short (latent, prob 0.22). Both arm conditions fired — IWM failed the 299.49 record breakout (299.69 high → 296.69 close) and SPY lost VWAP — and the short’s 297.76 target was hit (+2.49R close-based); raw liquid A/D 0.73 confirms negative breadth, vindicating the fade. But the literal rsp_over_spy breadth-broad kill held (RSP +0.71% > SPY -0.05%), so the setup’s own “do not short while breadth holds” gate never cleared. Mechanism fired; gate conflicted. (QQQ carried a benign STALE-FLAG: anchor 736.27 vs close 710.62 — a real multi-session tech decline, not a data error.) +0.41 IWM 296.69

Still open (interim — window not closed, carried forward not scored): sun-260621-LRT-SPY-S (weekly level-rejection short — no arm, no kill; SPY 733.24 sits ~17–23pt below the 750–757 rejection zone the setup arms on, and the week’s catalyst, May PCE, prints Thu 6/26) and sun-260621-MS-AIInfra-L (weekly semis / AI-infra long — arms on a strong Micron reaction; MU FQ3 reports AMC tonight, and semis closed red into the print, SMH -0.50% / XLK -0.62%). Both hinge on binaries that have not resolved.

LensOne level, 733.24, scored the whole slate. It sat above the 731 50-day — so the curd’s wider level-hold long (LRB) paid in the equal-weight/defensive basket (MIXED, the breadth caveat the only blemish) — but below the 736.92 VWAP, so the midday’s tighter continuation long VOIDED, and that same VWAP loss plus IWM’s failed breakout is exactly what armed the FBR fade into its target (MIXED, gated by the rsp>spy caveat). A bounce that round-tripped: broadening held at the median-stock level (RSP beat SPY, IWM green) while cap-weight gave it all back into Micron and PCE.

02Calibration

0.261Rolling Brierlast-50 validated · lower = better · 0.25 = random · cumulative all-validated 0.275 (n=62)
41.9%Hit rateFIRE / calibration-eligible (26/62) · last-50 window 44%
62Eligible sampleexcl. 3 NO_EVIDENCE / unevaluable
Pattern performance (Component 13) — calibration-eligible, by pattern × direction
PatternnHit%Mean probMean BrierEvidence class
LRT · S1765%0.350.370lore_pending
MS · L1331%0.490.263lore_pending
SBD · L60%0.370.138primary_source
SRB · L667%0.390.315lore_pending
GFD · L475%0.410.298practitioner_backtest
SBD · S40%0.280.079primary_source
LRB · L333%0.430.216lore_pending
MS · S3100%0.420.341lore_pending

Singletons (n<2) omitted: FBR·S (today’s first-ever entry — prob 0.22, MIXED), GFU·S, SRT·S, VAB·L, VBR·L, VSR·L. The 3 legacy name-schema Sundae setups (sun-260621-*) are excluded from the rollup (missing the p/dir/inst fields — same schema gap as the xlsx build; ISS-20260622-01).

Calibration trend (Component 14) — rolling-50 Brier by validation date
06-08.333
06-09.326
06-10.311
06-12.301
06-15.289
06-16.283
06-18.278
06-22.271
06-23.272
06-24.261
Mechanism leak · C15Canonical count 6 (right-direction / wrong-mechanism across the record), rate 0.333. Today added 2 — both of today’s MIXED outcomes: the curd LRB·L (SPY closed 733.24 above the 731 long-level — direction right — but the oversold-reclaim mechanism only half-delivered as A/D rolled under) and the FBR·S (IWM closed 296.69 below the 299.49 short-level and tagged target — direction right — but the rsp>spy gate blocked a clean fire). The structural leak remains LRT·S (3 of the 6; n=17, Brier 0.370 — the worst multi-n pattern): right on direction, an entry-location problem.
LensToday’s slate scored a run-Brier of 0.132 — three low-conviction reads (LRB·L 0.38/0.45, FBR·S 0.22) that correctly did not cleanly fire as the bounce round-tripped — pulling the rolling-50 down 0.272 → 0.261, the lowest in the record. The structural signal: both MIXED outcomes are right-direction / wrong-mechanism (the level call was right, the trade gated out), the same signature LRT·S carries. New data point: LRB·L is now 1/3 (33%) — the bottom-reclaim long has fired once in three, an early sign it needs the breadth confirmation to actually hold, not just the level to be defended. MS·S (3/3) and GFD·L (75% at 0.41 prob) stay the edges priced below their hit rate; MS·L (31% at 0.49) the overconfident chase; SBD remains 0-for-10 (0/6 L, 0/4 S) — low Brier masking a pattern that simply has not worked.

03Tape & Rate Backdrop

FADED BOUNCE — BROADENING INTACT, CAP-WEIGHT ROUND-TRIPPED Realized regime (lookback, not a forward call): the morning RANGE-BOUND STABILIZATION / bounce-attempt call and the midday TREND-UP upgrade both delivered for ~4 hours — SPY ran +0.72% to 739.95, IWM to a fresh record — then the close gave the entire move back. Cap-weight SPY (-0.05%) / QQQ (-0.42%) round-tripped while equal-weight RSP (+0.71%) and IWM (+0.46%) held green: the broadening held at the median-stock level, the index headline faded to flat into Micron + Thursday’s PCE.
IndexCloseDay %RangeVWAP
SPY733.24-0.05%730.84–739.95734.32
QQQ710.62-0.42%704.45–719.93712.34
IWM296.69+0.46%294.69–299.69297.00
RSP (equal-wt)210.38+0.71%209.25–211.55210.46
LensA bounce that round-tripped. The morning read was right for half a session; then the catalyst calendar took over — tech stayed heavy into Micron and the tape de-risked into PCE. The tell is the split close: equal-weight and small-caps green, cap-weight flat-to-red. SPY’s 733.24 — above the 731 50-day, below the 736.92 VWAP — is the single number that made the wider level-hold long work and the tighter VWAP-continuation fail.

04Cross-Asset

AssetCloseDay %Read
UUP (USD)28.53+0.28%Dollar firm — mild bid
USO (WTI)106.29-4.47%Crude crushed to a 3-mo low — US-Iran ceasefire, supply not demand
GLD (gold)365.92-3.02%Gold dumped — haven unwind, no flight-to-safety
TLT (20y+)87.38+1.37%Duration bid — yields eased, disinflation tailwind
HYG (HY credit)79.85-0.03%Credit dead flat — no stress
LensA clean disinflation / haven-unwind story, not an equity-stress story: oil and gold dumped, miners and crypto with them, while HY credit stayed dead flat (-0.03%) and bonds caught a duration bid (TLT +1.37%). Lower oil + lower yields = a disinflation tailwind that helped the broad/equal-weight bounce early — but not enough to offset the tech/Micron drag into the close.

05Sector & Breadth (realized)

XLIIndu+1.16
XLYDisc+1.15
XLUUtil+1.04
XLPStpl+0.86
XLVHlth+0.77
XLBMatl+0.57
XLRERE-0.29
XLFFin-0.30
XLKTech-0.62
XLCComm-0.68
XLEEngy-1.63
LensA genuinely split breadth picture, and reading it right is the whole game. Within the S&P breadth was positive (RSP beat SPY by 76bp, 6/11 SPDRs green, cyclicals AND defensives bid); across the full liquid tape A/D was NEGATIVE at 0.73, because the commodity/resource complex was crushed on the ceasefire. That 0.73 is exactly the figure that tripped the curd LRB’s breadth_rolls_over_adrn<1 kill — the reason that setup is MIXED, not a clean FIRE.

06Single-Name Movers (illustrative)

Confirmed from the Massive close (snapshot / grouped). Closing-bell newsletters (Stocktwits Daily Rip / Axios Closer) not pulled this run (headless egress) — granularity limited to ETF/sector level; illustrative of the tape only, never trade plans.

LensWith newsletters unpulled, the confirmed read is sector-level — and it tells the day cleanly: energy and the commodity complex led the downside on the Iran ceasefire, semis sat flat into Micron, and the green was spread across cyclicals and defensives. The one binary that matters now is Micron FQ3 AMC tonight — the catalyst the open weekly semis-long is explicitly waiting on.

07Morning & Midday Lens vs Reality

The Early Bird Curd called RANGE-BOUND STABILIZATION / bounce-attempt (event-gated, conviction medium) — after the two-day Korea-led AI/memory margin call, US futures steadied premarket, and the curd bet SPY holds the 50-day ~731 and the broad market re-rates as the tech rotation continues. The Midday Frappe upgraded it to TREND-UP / broadening risk-on (SPY +0.72%, IWM at new highs, RSP beating SPY, 9/11 SPDRs green).

Reality: the bounce was real but did not hold. SPY ran to 739.95 by midday, then faded the entire move to close 733.24 (-0.05%) — back above the 731 50-day (curd level held) but below the 736.92 VWAP (midday continuation level lost). The broadening held at the index level (RSP +0.71%, IWM +0.46%, 6/11 SPDRs green) but cap-weight round-tripped as tech stayed heavy into Micron and the tape de-risked into Thu’s PCE.

LensThe morning lens was directionally right for half a session, then the catalyst calendar took over. The curd’s “hold the 50-day, broad re-rate” thesis worked at the level and in the equal-weight tape; the midday TREND-UP upgrade got the broadening right but over-extrapolated the intraday continuation, which the late de-risk unwound. Nothing resolved — Micron tonight and PCE Thursday are still the two binaries, and the two open weekly Sundae setups (semis-long, level-rejection short) both hinge on them.

08Execution Debrief — realized execution on closed setups, NOT a forward trade plan

SetupProfileEntryStopTargetRealized R
FBR·S md-260624-1209-FBR-IWM-Sday (5m ATR)299.40300.49297.76+2.49R
LRB·L mm-260624-LRB-SPY-Lswing (1d ATR)731.00713.39766.23+0.13R*
LRB·L md-260624-1209-LRB-SPY-Lday (5m ATR)736.90735.28739.33-2.26R

Realized R computed entry→close. Aggregate: mean +0.12R, 2 of 3 positive, 1 stopped out (the midday LRB — close 733.24 below the 735.28 stop). The FBR short genuinely tagged below its 297.76 target (IWM 296.69) — a fillable +2.49R. * the swing curd LRB carried an EXEC-FLAG (rpu 17.61 below the swing 1.5×ATR floor ~17.95) — a slightly tight stop at emit; +0.13R is the token hold of the 50-day. The midday LRB −2.26R is the close-based metric; live, the 735.28 stop caps the loss at −1R.

LensThe night’s harvestable trade was the one the setup labeled “NOT armed, watch” — the FBR·S returned +2.49R once IWM failed its 299.49 breakout and SPY lost VWAP (both arm conditions), tagging below the 297.76 target. The midday LRB continuation long is the cautionary one: entered at the 736.92 VWAP, it lost the level and the 735.28 stop on the late fade (−1R live / −2.26R close) — the spy_loses_vwap kill was the signal to stand down, and honoring it capped the damage. The swing curd LRB held the 50-day for a token +0.13R. Net +0.12R mean — the gate’s kill discipline turning a faded tape into a small net positive.