XAUUSDReviewCautious

GOLD Session Review: September 18, 2026, The 4,381 Break Failed Twice

The September 18 GOLD preparation was partially accurate. Its 38% range lead missed, but the 10% delayed upside-trap branch captured the defining sequence: acceptance above $4,381.07, reversal toward $4,341.61, and a late recovery that still closed below the cap. The next preparation should demand acceptance above both $4,381.07 and $4,402.35 before calling continuation durable.

Prep outcomepartial
Lead scenario38% · missed
Leanneutral · incorrect
Day typerange → trap
Surpriselow
Grade card8 of 14 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 Cap Break and Reversal Session
Symbol
XAUUSD
Window
00:00 to 23:59 UTC
Day type called
Range
Day type actual
Trap
Lean outcome
Incorrect
Regime
Upper-cap breakout, New York reversal, late recovery
Preparation
Partially accurate
Surprises
Low

Grade card

8 of 14 correct
  1. Day-type callIncorrect
    Called
    Range after a full-ATR advance, with no scheduled USD catalyst
    Actual
    Price accepted above the upper edge, reversed through the breakout, then failed a second late test
  2. LeanIncorrect
    Called
    Neutral / Wait
    Actual
    Price closed $34.83 above the open despite the intraday reversal
  3. Conditional leanPartial
    Called
    Long after a held H1 close above $4,366.16, short after a held loss of $4,317.22
    Actual
    The long trigger fired by 09:00 UTC and extended to $4,399.39, but the move failed before $4,402.35; the short trigger never fired
  4. Lead scenarioIncorrect
    Called
    Post-trend range digestion at 38%
    Actual
    The range trigger did not form; the 10% delayed upside trap fired
  5. Scenario mapCorrect
    Called
    Four branches covering digestion, continuation, bearish failure, and a delayed upside trap
    Actual
    The delayed upside-trap sequence matched: multiple H1 closes above $4,381.07, then closes below $4,381.07 and $4,366.16
  6. Driver stackPartial
    Called
    Price resilience challenged the clean range call, while the empty calendar favored digestion
    Actual
    Price resilience carried the early advance, but the lack of a fresh catalyst left the breakout unable to hold
  7. $4,434.55 structural resistanceCorrect · not tested
    Called
    Outer continuation target after acceptance above the nearer caps
    Actual
    Not tested
  8. $4,402.35 resistanceCorrect · within noise
    Called
    First extension target, with rejection able to start the trap path
    Actual
    The high stopped at $4,399.39, $2.96 below the level, before the reversal
  9. $4,381.07 bullish decision edgeCorrect
    Called
    A held break favored continuation; a later failure activated the trap branch
    Actual
    Three H1 closes held above it before price reversed below it, then a late retest also failed by the close
  10. $4,366.16 early trend switchPartial
    Called
    Acceptance switched the day from Range toward Trend
    Actual
    Early acceptance led to the session high, but the level failed during the New York reversal and was reclaimed late
  11. $4,341.61 balance markerCorrect · within noise
    Called
    Rotation through it supported digestion; a close below warned that continuation was failing
    Actual
    The reversal reached $4,342.52, within $0.91, and bounced without an H1 close below it
  12. $4,317.22 bearish decision edgeCorrect · not tested
    Called
    A displaced loss and failed retest activated bearish repair
    Actual
    Not tested; the session low held $17.13 above it
  13. $4,290.63 supportCorrect · not tested
    Called
    First target after a confirmed $4,317.22 break
    Actual
    Not tested because the bearish trigger never fired
  14. $4,261.33 to $4,257.36 support zoneCorrect · not tested
    Called
    Final mapped downside target
    Actual
    Not tested because the bearish trigger never fired

The tape

  1. Open, first 60 minutes

    Gold opened at $4,342.89 and stayed close to the prior $4,341.61 balance marker. The broader overnight phase held $4,317.22 comfortably, printed the day's low at $4,334.35 during the 06:00 hour, and recovered to $4,358.83 before London. The bearish branch never armed.

  2. Mid-session

    London supplied the first clean expansion. The 07:00 H1 close reached $4,366.56, the 08:00 close reached $4,377.18, and the 09:00 close at $4,393.39 confirmed acceptance above $4,381.07. Gold reached $4,399.39 during the next hour, stopping just short of the $4,402.35 extension level. The break held for three hourly closes, then the 12:00 candle closed back below $4,381.07 at $4,376.49. With no medium or high-impact USD event on the calendar, the turn came from failure at the mapped resistance rather than a scheduled catalyst.

  3. Late / close

    Price stabilized around $4,381 through the first part of New York, then the mapped 15:00 to 16:00 UTC reversal window took control. The 15:00 candle fell to $4,353.59, and the 16:00 candle reached $4,347.13 before closing at $4,349.21. That completed the delayed trap trigger by carrying price below both $4,381.07 and $4,366.16. A late rebound reached $4,396.73 and briefly restored closes above $4,381.07, but the second attempt also failed. Gold settled at $4,377.72, above the $4,366.16 switch and below the $4,381.07 cap.

What we learned

The session did not produce an unmapped path. The low surprise comes from the exact delayed-trap sequence already sitting in the scenario map, even though its 10% weight understated the chance that repeated acceptance above an obvious cap would later fail.

  1. Day-type precondition checks

    Raise the trap branch when a quiet post-trend session approaches an obvious, recently tested cap. The empty calendar supported digestion, but it also removed the catalyst needed to sustain a clean break. This is a day-type selection miss, so the next preparation should re-examine the preconditions before defaulting to Range.

  2. Day-type precondition checks

    Treat two or more H1 closes above the decision edge followed by a close back inside as a live reclassification from Trend to Trap. The preparation described that sequence, but its 10% weight left it too far behind the co-leads.

  3. Driver-stack order

    No instrument-priors edit is justified from this session alone. The prep correctly elevated price response over an assumed macro sign, and that read explained the early advance. The next version should state more clearly that absent real-yield and dollar confirmation reduces the durability of price-only acceptance.

  4. Tradability standard

    Keep the held-close trigger. It activated before the move from $4,377.18 to $4,399.39 and did not rely on a level touch. Once the 12:00 H1 close returned below $4,381.07, the continuation case had to be reset rather than carried into New York. Reviewed prep: 2026-09-18-xauusd-session-preparation