XAUUSDReviewCautious

GOLD Session Review — September 11, 2026: Hot CPI Produces a Two-Sided Whipsaw

Gold closed $31.52 above its open after a 1.06×-ATR session broke both decision edges but accepted neither, making the Neutral/Wait posture useful while invalidating the Event-suspended day-type call and the 41% bearish lead. The key carry-forward is to treat gold's refusal to follow a hot inflation print as fresh driver evidence and to map post-event whipsaw independently of whether the release is inline.

Prep outcomepartial
Lead scenario41% · missed
Leanneutral · correct
Day typeevent-suspended → whipsaw
Surprisehigh
Grade card8 of 17 correct
Session chart
XAUUSD — session vs prep levelsH1
Shaded band: the reviewed session. Price lines: the preparation's key levels — see which held and which broke.
Powered by Cortiq

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
GOLD CPI Repricing Whipsaw Session
Symbol
XAUUSD
Window
00:00 – 23:00 UTC
Day type called
Event-suspended
Day type actual
Whipsaw
Lean outcome
Correct
Regime
Delayed post-CPI two-sided whipsaw
Preparation
Partially accurate
Surprises
High

Grade card

8 of 17 correct
  1. Day-type callIncorrect
    Called
    Event-suspended around the 12:30 UTC CPI release
    Actual
    Pre-release balance gave way to a 1.06×-ATR session that broke both decision edges and held neither into the close
  2. LeanCorrect · no durable resolution
    Called
    Neutral/Wait because the 41% downside lead was below the threshold for an outright directional lean
    Actual
    Gold closed $31.52 above the open but near the midpoint of the day's range and back inside the decision band
  3. Conditional lean — short below $4,313.59Correct · no held trigger
    Called
    Short only after a held H1 close $3–$15 below the edge and a failed reclaim
    Actual
    The 06:00 H1 close cleared the level by $9.80, but the next hour reclaimed it; the later low at $4,290.63 also reversed inside the same H1 candle
  4. Conditional lean — long above $4,365.40Partial · trigger worked briefly, no durable hold
    Called
    Long only after an H1 reclaim above $4,365.40 with $4,341.16 holding on retest
    Actual
    The 15:00 H1 close reclaimed the edge, price extended to $4,402.35, and $4,341.16 held; the move then failed back below $4,365.40 before the close
  5. Lead scenario — CPI-reinforced bearish continuation (41%)Incorrect
    Called
    Hot CPI followed by a held $3–$15 H1 break below $4,313.59, opening $4,282.44
    Actual
    CPI printed hot, but no post-release H1 close held below $4,313.59; gold instead reversed from $4,290.63 to above $4,400
  6. Recovery scenario (37%)Partial · price path appeared, catalyst and durability did not
    Called
    Soft CPI followed by a held reclaim of $4,365.40 and a retest hold above $4,341.16
    Actual
    The price sequence appeared after a hot CPI print, reached $4,402.35, then failed back inside the band
  7. Whipsaw scenario (22%)Incorrect · broad shape right, defining conditions wrong
    Called
    Inline or mixed CPI, followed by a failed break of either edge and no acceptance by 18:00 UTC
    Actual
    CPI was materially hot; both edges broke, the upside remained accepted past 18:00, and the return inside the band came later
  8. Scenario-map coverageIncorrect · no branch fired as written
    Called
    One directional branch tied to hot CPI, one recovery branch tied to soft CPI, and whipsaw tied to an inline print
    Actual
    A hot print produced the unmapped combination of an initial drift lower, a delayed upside reversal, and a late return to balance
  9. Driver stackPartial · event dominance right, response mapping wrong
    Called
    Bearish D1/H4 structure gave downside a modest lead, while CPI repricing was expected to decide whether yields and the dollar reinforced or reversed it
    Actual
    The hot CPI outcome did not produce sustained gold weakness; gold's refusal of the expected inverse response became the session's decisive information
  10. $4,510.61 structural resistanceCorrect · not in play
    Called
    Outer bearish invalidation, not a base-case intraday target
    Actual
    Not tested; the session high was $4,402.35
  11. $4,490.65 resistanceCorrect · not in play
    Called
    Acceptance above would materially weaken the bearish sequence
    Actual
    Not tested; price remained $88.30 below it
  12. $4,434.55–$4,442.80 resistance / liquidityPartial · recovery activated but fell short
    Called
    Final upside sweep zone after recovery through $4,380.98
    Actual
    The recovery reached $4,402.35 before reversing, leaving the zone untested
  13. $4,380.98 regime resistanceCorrect · temporary repair, then rejection
    Called
    Recovery checkpoint; rejection would preserve lower-half control
    Actual
    Price accepted above it for two H1 closes, reached $4,402.35, then rejected and closed back below
  14. $4,365.40 bullish decision edgePartial · genuine repair attempt, no closing acceptance
    Called
    A held reclaim would begin structural repair; failure would keep lower control active
    Actual
    Five H1 closes held above after the 15:00 reclaim, but the level failed at 20:00 and the D1 close returned below it
  15. $4,341.16 near-term pivotCorrect · accurately marked the control pivot
    Called
    Acceptance above supports repair; rejection favors another floor test
    Actual
    London reclaimed it, the post-CPI selloff briefly lost it, and the session ultimately closed $8.32 above it
  16. $4,313.59 bearish decision edgeCorrect · confirmation filter rejected both false breaks
    Called
    Downside sweep target; continuation required $3–$15 displacement and a failed reclaim
    Actual
    The first qualifying close was immediately reclaimed, while the later deeper sweep reversed violently; neither became bearish acceptance
  17. $4,282.44 structural supportCorrect · not reached
    Called
    First downside objective after a confirmed bearish break
    Actual
    Not tested; the session low stopped $8.19 above it

The tape

  1. Open / Asian hours (00:00–07:00 UTC)

    Gold opened at $4,317.96 and spent the first five hours rotating around the $4,313.59 downside edge. The 06:00 hour then closed at $4,303.79, a $9.80 displaced break inside the prep's preferred confirmation band, but the move was not held: the next hour reclaimed $4,313.59 and closed at $4,331.69. That immediate failure kept the bearish condition from qualifying.

  2. London open and pre-CPI positioning (07:00–12:00 UTC)

    The reclaim accelerated through $4,341.16, reaching $4,360.64 during the 09:00 hour before settling into balance. From 10:00 through noon, price remained inside the $4,313.59–$4,365.40 decision band. The preparation's instruction to wait through the 12:00–13:00 CPI blackout was aligned with the tape.

  3. CPI and the first response (12:30–15:00 UTC)

    Headline CPI printed at 0.4% month over month against a 0.0% forecast, core CPI at 0.3% against 0.2%, and headline CPI at 3.4% year over year against 2.7%. Despite the hot cluster, the release hour closed almost unchanged at $4,346.15. Gold then drifted lower through the 13:00 and 14:00 hours, but still produced no held close below $4,313.59; the lead scenario never confirmed.

  4. Delayed reversal and close (15:00–23:00 UTC)

    The decisive volatility arrived later. During the 15:00 hour, gold plunged to the session low at $4,290.63, reversed through both decision edges, and closed at $4,388.43. The 16:00 hour extended to $4,402.35, after which price retraced steadily, lost $4,365.40 at 20:00, and closed at $4,349.48. The final close was $31.52 above the $4,317.96 open, $58.85 above the low, and almost exactly at the midpoint of the $111.72 range. At roughly 1.06× the prep's $105.80 D1 ATR, this was a normal-to-expanded event-day range, not a compressed full session.

Full notes

What We Learned

The major surprise was not volatility but the sign and timing of the response. CPI printed materially hotter than forecast, yet the mapped bearish continuation never confirmed; instead, gold reversed nearly $100 from the 15:00-hour low, held above the bullish edge for several hours, and then returned to balance. The prep anticipated first-impulse failure and delayed resolution in general, but it made whipsaw conditional on an inline print and recovery conditional on a soft print, leaving the actual hot-print whipsaw outside the map.

  1. Day-type miss → re-examine precondition checks. “Event-suspended” correctly described the hours before CPI but not the full trading day. Future event-day preps should separate the pre-release state from the expected post-release day type and assign explicit full-session weights to Trend, Trap, and Whipsaw outcomes.

  2. Driver stack mis-ordered for this instrument → propose a priors edit. Once hot CPI failed to hold gold below $4,313.59 and the market reclaimed $4,341.16, observed price refusal should have outranked the assumed real-yield and dollar response. Add a gold-specific re-score rule: failure to follow the expected macro sign after one H1 close is fresh driver evidence, not merely a delayed version of the original thesis.

  3. Day-type miss → re-examine precondition checks. The whipsaw branch should be defined by price behavior, not reserved for inline or mixed data. A hot or soft print can still produce two-sided failure; the next map should allow Whipsaw whenever both decision edges break and neither retains H1 acceptance.

  4. Right price branch, untradable trigger → tighten the tradability standard. The recovery sequence cleared its first checkpoint inside the same H1 candle that created the reclaim, while the confirming retest arrived only after much of the move had already occurred. Future recovery branches should require the retest before counting the path as actionable and should not validate a trigger retrospectively after its first objective has already traded.


Reviewed prep: 2026-09-11-xauusd-session-preparation