SP500ReviewCautious

SP500 Session Review — August 20, 2026

The Coil Broke Before the Print, and the Session Never Looked Back

SP500's overnight 7,713-7,728 coil gave way by 10:00-11:00 UTC, roughly two hours ahead of the 12:30 UTC Philly Fed / Jobless Claims cluster the day-type call was built around, and a strong Philly Fed beat (47.4 vs 13.7 forecast) failed to arrest the slide. The 14:30 UTC cash open confirmed the breakdown with real displacement, driving price through 7,698.68 and 7,680.12 into an 88-point, ~121%-of-ATR trend day that closed at 7,641.40 — the lowest-weighted (28%) scenario branch is what fired, not the 40%-weighted range/coil lead.

Prep outcomepartial

Lead scenario40% · missed

Leanneutral · incorrect

Day typeevent-suspended → trend

Surprisemoderate

Grade card6 of 10 correct

Every prep call is graded like this. See the full track record →

Session chart

SP500 — session vs prep levelsH1

Shaded band: the reviewed session. Price lines: the preparation's key levels — see which held and which broke.

Graded against live session data from Cortiq

Every call in this review is checked against the session data captured in Cortiq — the AI trading workspace that powers our daily preparation, review, and calibration loop.

See how Cortiq works

Session: SP500 Pre-Print Breakdown Into a Trend Day
Symbol: SP500
Window: 00:00 – 21:00 UTC
Day type called: Event-suspended
Day type actual: Trend
Lean outcome: Neutral-incorrect (conditional short leg: Correct)
Regime: Breakdown trend day, ~121% of ATR(14), closing near the session low
Preparation: Partially accurate
Surprises: Moderate

Grade card

6 of 10 correct

Day-type callIncorrect
Called
Event-suspended — compression holds until the 12:30 UTC data cluster, resolved by the 14:30 UTC cash open
Actual
Compression broke by 10:00-11:00 UTC, roughly two hours before the print; the session extended into a full trend day, closing near the low
LeanNote · Neutral-incorrect
Called
Neutral / Wait
Actual
Session resolved firmly bearish, closing ~74 points below the open and holding the move through the close
Conditional leanCorrect
Called
Short on a confirmed, displaced H1 close below 7,716.94, surviving the 19:00-21:00 power hour
Actual
Triggered by 11:00 UTC, reconfirmed with real displacement after the cash open, and held through the power hour to the close
Lead scenario (Range/coil hold, 40%)Incorrect
Called
Two-way chop inside 7,716.94-7,742.74 through and past the data cluster
Actual
Broken decisively before the data cluster even landed; no coil held
Scenario Map weightingIncorrect · map quality
Called
Breakdown carried the lowest weight of the three branches (28%, an 8-point trail behind the 40% lead)
Actual
Breakdown is exactly the branch that fired
Key level 7,742.74 (resistance)Correct
Called
Expected to cap any upside test
Actual
Day high 7,727.65, set in the overnight book; never approached
Key level 7,727.65 (near resistance)Correct
Called
Immediate overhead pivot inside the coil
Actual
Matched the overnight session high exactly
Key level 7,716.94 (breakdown trigger)Correct
Called
A confirmed, displaced loss opens the retest of 7,698.68
Actual
Lost by 10:00-11:00 UTC; 7,698.68 was tested within the hour
Key level 7,698.68 (support)Correct
Called
Earliest tell the breakdown branch is extending
Actual
Swept intraday (11:00 UTC low 7,697.45), then closed through decisively after the cash open
Key level 7,680.12 (critical support)Correct
Called
A confirmed, displaced loss reopens the 7,602-7,609 shelf
Actual
Lost with real displacement from 15:00 UTC onward, held broken through the close

The tape

Open (00:00-01:00 UTC)
The session opened at 7,715.02, immediately inside the overnight coil the prep had already flagged as compressed — the first hour traded a tight 7,713-7,725 band, consistent with the "dead overnight book" baseline.
Mid-session
The coil held through 09:00 UTC, capped almost exactly at the prep's own 7,727.65 overhead pivot. Then, in the 07:00-09:00 EU-cash-open window the prep itself had flagged as "the first real liquidity check," price cracked lower: the 10:00 UTC hour closed at 7,711.95, just below the 7,716.94 checkpoint, and the 11:00 UTC hour extended to a 7,697.45 low before closing at 7,700.45 — a confirmed break of both 7,716.94 and a test of 7,698.68, roughly two hours before the 12:30 UTC Philly Fed / Jobless Claims cluster. The data itself printed a sharp Philly Fed beat (47.4 vs. 13.7 forecast, above even the 41.4 prior) and an in-line claims number; price bounced modestly into 13:00 UTC (close 7,712.45) but did not reclaim the coil. The 14:30 UTC cash open then confirmed the breakdown outright: the 14:00 UTC hour closed at 7,686.20 off a 7,685.07 low, and the 15:00 UTC hour swept 7,680.12 with a 7,662.20 low, closing at 7,678.70 — both key levels lost with real displacement, in the order the prep's invalidation ladder described.
Late / close
The 16:00-19:00 UTC stretch chopped in a 7,672-7,702 band without reclaiming 7,680.12 on a closing basis, then the 19:00-21:00 UTC power hour extended the move rather than pinning it — 20:00 UTC closed at 7,665.45, 21:00 UTC at 7,652.20 — before drifting to a 7,641.40 close into the maintenance window. The day's 7,639.65-7,727.65 range (88.0 points, ~121% of the 72.58-point ATR) finished well beyond the event-suspended compression premise, closing near the session low and just above the deeper 7,602-7,609 shelf without testing it.

What we learned

No material surprise in the level cascade — the prep's own ladder (7,716.94 → 7,698.68 → 7,680.12 → 7,602-7,609) called the path in the order it happened. The surprises were in the timing and the reaction to data:

The overnight coil broke before its own catalyst
The compression the day-type call was built around gave way in the 10:00-11:00 UTC window, roughly two hours ahead of the 12:30 UTC data cluster the map treated as the decision point. The prep's Session Map had already named the 07:00-09:00 UTC EU cash open as "the first real liquidity check on whether the overnight coil holds," so this wasn't unforeseeable — it just wasn't scored as its own branch. Routing: chronically wrong day-type calls → re-examine precondition checks. Tighten the event-suspended precondition to require the coil surviving through the 07:00-09:00 UTC EU cash open, not just the 00:00-07:00 UTC dead book, before calling a day event-suspended.
A clear data beat did not buy the level back
Philly Fed printed well above both its forecast and its prior — an unambiguous beat — yet price never reclaimed 7,716.94 and extended lower through and after the 14:30 UTC cash open. The driver stack correctly flagged the rates read as "two-sided, pending" but had no branch for a beat that failed to support price. Routing: right scenario, untradable trigger → the breakdown branch had the correct trigger and target ladder but was under-weighted; when the Session Map itself names an early-break risk window, that risk should lift the corresponding scenario's probability weight rather than living only in prose.
The overnight-recovery read in the driver stack overstated its own confidence
"Prior-day structure holding its overnight recovery" was graded the one driver with a clear, constructive read and cited as support for the coil lead — but the recovery it was reading came from the same thin, pre-cash-open liquidity the shared priors already treat as non-tradeable ("session ignition, not overnight"). Routing: driver-stack mis-ordering → discount overnight-book structural signals in the alignment verdict until they're confirmed by the EU cash open, consistent with the existing "session ignition, not overnight" prior. Session-analysis grading check: pass — all three misses routed to their section 6 category (precondition check, scenario re-weighting, driver-stack discount) rather than left as generic notes. --- Reviewed prep: 2026-08-20-sp500-session-preparation