SP500ReviewCautious

SP500 Session Review — August 26, 2026

An In-Line Print Ends the 7,680 Coil, But Only After Three Round Trips

SP500 finally closed above the five-times-failed 7,680.12 pivot at a confirmed 7,686.77, but only after an in-line Core PCE/GDP print triggered a 16:00 UTC breakout that round-tripped back below the level three separate times before holding into the bell. The prep's 40%-weighted breakdown lead never engaged — price never came within 8 points of 7,648.90 — while its 33%-weighted reclaim branch correctly named the print condition that fired, just not cleanly enough to call the day-type resolution 'event-suspended' rather than a genuine whipsaw.

Prep outcomepartial

Lead scenario40% · missed

Leanneutral · incorrect

Day typeevent-suspended → whipsaw

Surprisemoderate

Grade card3 of 8 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 Core PCE/GDP Decision Session Ahead of Jackson Hole
Symbol: SP500
Window: 00:00 – 23:00 UTC (index session clock; 14:30 UTC cash open)
Day type called: Event-suspended
Day type actual: Whipsaw (pre-print compression held; post-print resolution round-tripped the pivot three times before holding)
Lean outcome: Incorrect
Regime: In-line print, five-times-failed pivot finally clears on a closing basis after repeated round trips
Preparation: Partially accurate
Surprises: Moderate

Grade card

3 of 8 correct

Day-type callPartial
Called
Event-suspended
Actual
Pre-print compression held exactly as expected, but the post-print resolution whipsawed across the pivot three times before settling
LeanIncorrect
Called
Neutral / Wait
Actual
Session closed above the pivot for the first time in six sessions — a real, if modest, directional resolution
Conditional leanPartial
Called
Long on a confirmed, displaced H1 close above 7,680.12 surviving the cash open and power hour, following a soft/in-line print
Actual
Print landed in-line, trigger activated; break came at 16:00 UTC but was invalidated at 17:00, 18:00, and 19:00 UTC before finally holding into the 23:00 UTC close
Lead scenarioIncorrect
Called
Breakdown continuation on a hot print (40%)
Actual
Print landed in-line, not hot; price never closed within 8 points of 7,648.90
Key level 7,680.12 (operative pivot)Partial
Called
Any test needs a confirmed, displaced close surviving the power hour
Actual
Broken at 16:00 UTC, round-tripped below three times, held only by the 23:00 UTC close and the 7,686.77 daily close
Key level 7,657.15 (fresh intraday support)Correct
Called
A loss pre-print would be an early breakdown tell
Actual
Touched exactly at 04:00 UTC (session low) and again at 19:00 UTC (7,657.45) — never closed below
Key level 7,648.90 / 7,639.65 (contested support band)Correct
Called
Today's operative breakdown test
Actual
Never approached — session low 7,657.15 stayed clear by 8+ points, the first session in a week without a sweep of this band
Key levels 7,699.32 / 7,696.21 / 7,716.94 (reclaim extension targets)Correct
Called
Ceiling any reclaim attempt must clear for a real extension
Actual
Session high 7,692.77 fell short of all three — the reclaim held the pivot but never extended

The tape

Open (first 30-60 minutes)
The session opened at 7,679.15, right at the pivot, and immediately faded through the dead overnight book toward a 04:00 UTC low of 7,657.15 — an exact touch of the fresh intraday support named in the prep, with no close beneath it. Price recovered to 7,675.9 by the 06:00 UTC close, a fifth approach to the pivot region without a confirmed break in either direction, exactly as the prep's overnight read anticipated.
Mid-session
The EU cash open (07:00 UTC) produced a modest dip to 7,669.52 before the market ground sideways in a tight 7,666.89-7,677.45 band into the 12:30 UTC print. The release itself was a genuinely in-line set — Core PCE m/m 0.2% (vs. 0.2% forecast), Core PCE y/y flat at 3.3%, GDP q/q 1.5% (vs. 1.5%), GDP Sales 2.2% (vs. 2.2%) — with only Durable Goods surprising to the upside (1.1% vs. 0.8%). The immediate reaction was muted (the 12:00-13:00 UTC candle held a 6.6-point range), consistent with an in-line print not delivering a decisive repricing. The 14:30 UTC US cash open produced no immediate resolution either; price actually faded to a fresh session-adjacent low of 7,662.95 at 15:00 UTC before reversing sharply. At 16:00 UTC, price broke 7,680.12 with real displacement, printing a session high of 7,687.70 and closing the hour at that high — the first confirmed break of the pivot in six sessions. It did not hold: the 17:00 and 18:00 UTC closes (7,676.95 and 7,671.70) both fell back below 7,680.12, and the 19:00 UTC power-hour open extended the round trip to a fresh low of 7,657.45, closing the hour at 7,665.64 — deep below the pivot with no reclaim in sight.
Late / close
The reversal came late. The 20:00 and 21:00 UTC closes (7,675.20 and 7,683.45) reclaimed the pivot for a third time, dipped once more to 7,679.03 at 22:00 UTC, then pushed to a fresh session high of 7,692.77 in the final hour, closing at 7,687.59. The confirmed daily close of 7,686.77 settled 6.65 points above the pivot and 7.62 points above the day's open — a real but narrow directional resolution (about 14% of the 55.78-point ATR) that took three failed holds to arrive at.

What we learned

No surprise in the print itself — Core PCE and GDP landed exactly on consensus, and the prep correctly flagged that an in-line read would favor the reclaim branch over the breakdown branch. The surprises were in how messily that reclaim actually resolved:

The reclaim branch's own stated invalidation fired three times before its own scenario won
(route: tighten the tradability standard). The prep's rule for calling the reclaim confirmed was "a confirmed H1 close back below 7,680.12 after the break" as the invalidation — that exact condition triggered at 17:00, 18:00, and 19:00 UTC, and only the 23:00 UTC close and the daily close ultimately vindicated the branch. A single power-hour-close check is not enough to call a multi-touch pivot confirmed; the next prep should require a closing-basis hold through the full session close, not just a single qualifying hour, before treating a break of a five-times-tested level as decided.
The print-conditional branch's own trigger logic worked; the static lead weighting did not
(route: driver stack). This is the second straight graded session where the branch that fired was not the 40%-weighted lead — the reclaim branch's stated condition ("soft/in-line print, confirmed break above 7,680.12") activated exactly as designed once the print landed in-line. The fix is not the trigger language, which worked, but the practice of weighting a structurally-favored lead (a level's failure record) above a print-conditional branch when the structural case itself is genuinely two-sided — a level that has both failed as resistance and held as support in the same week is not strong enough evidence to justify a 7-point lead margin over the branch that explicitly depends on the incoming data.
Day-type call undersold how much chop followed the print
(route: re-examine precondition checks). The pre-print compression held exactly as the "event-suspended" call anticipated, but the label implies the print itself settles the range in one move. Instead, the session took three additional round trips after the release to resolve. When the print in hand is genuinely in-line across every headline component (no hot or soft surprise on the core measures), the day-type default should anticipate a post-print whipsaw around the operative pivot rather than a single clean break, and the Session Map's no-trade language should extend past the immediate post-print window into the full 16:00-21:00 UTC stretch this instrument used to settle today.
Full-day range undershot the tier-1-day ATR prior
(route: re-examine precondition checks, log as an instance only). The confirmed 35.62-point day range was roughly 64% of the 55.78-point ATR(14), below the "80-120%+ of ATR" pattern the priors carry forward for tier-1 event days. A cleanly in-line print across every component may compress realized range more than a surprise print does — worth watching for a repeat before touching the priors file, since this is a single data point. --- Reviewed prep: 2026-08-26-sp500-session-preparation