XAUUSDReviewCautious

GOLD Session Review — August 19, 2026: A Digestion Call Meets a $177 Breakout

The prep mapped a co-led digestion day after yesterday's swept low reversed within its own candle -- instead gold broke and held above every mapped level, including the extreme-case-only $4,436.03 ceiling, closing $177.67 (+4.1%) higher with no confirmed catalyst behind the surge. The reversal/bounce branch and its conditional long trigger called the right side, but both the day-type call and the map's own target sequence badly undersold the move.

Prep outcomepartial
Lead scenario36% · missed
Leanneutral · incorrect
Day typerange → trend
Surprisehigh
Grade card5 of 9 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 Post-Sweep Breakout Session
Symbol
XAUUSD
Window
00:00 – 23:00 UTC
Day type called
Range (digestion)
Day type actual
Trend
Lean outcome
Incorrect (outright Neutral/Wait; session resolved sharply higher -- though the named conditional long trigger caught the correct side)
Regime
Trend -- broke and held above every mapped resistance level
Preparation
Partially accurate
Surprises
High

Grade card

5 of 9 correct
  1. Day-type callIncorrect
    Called
    Range/digestion (36%)
    Actual
    Trend -- closed near the day's high, ~1.9x D1 ATR of directional travel
  2. Lean (outright)Incorrect
    Called
    Neutral/Wait, combined weight 34%, under the 55% lean threshold
    Actual
    Session resolved decisively higher, not neutral
  3. Conditional leanCorrect · delayed confirmation
    Called
    Long on a held, displaced H1 close above $4,367.14
    Actual
    Trigger fired -- but the 13:00 UTC close ($4,367.21) cleared by only $0.07 and slipped back to $4,364.50 at 14:00 before genuine displaced confirmation arrived at 15:00 ($4,438.69 close)
  4. Lead scenario (36%, range/digestion)Incorrect
    Called
    Oscillate within $4,324.39-$4,367.14
    Actual
    Never happened -- the band was cleared decisively by 15:00 UTC
  5. Runner-up scenario (34%, reversal/bounce extension)Correct · on direction, badly undersized on magnitude
    Called
    Clear $4,367.14 -> $4,376.04 -> possible $4,396.75
    Actual
    Cleared all three and kept going to a $4,512.29 high / $4,510.28 close
  6. Key level $4,324.39 (support/pivot)Correct
    Called
    Fastest tell for trend-resumption-down
    Actual
    Held as the session low all day; never revisited
  7. Key level $4,367.14 (earliest tell, reversal branch)Correct
    Called
    Confirms reversal/bounce on a held, displaced close
    Actual
    Confirmed, with one fakeout hour first (see conditional-lean row)
  8. Key level $4,396.75 (secondary resistance target)Correct · as waypoint, undersold as a target
    Called
    "Secondary target only if $4,376.04 clears first"
    Actual
    Cleared and left far behind; not even a pause point
  9. Key level $4,436.03 ("major structural ceiling," extreme-case only)Note · Breached -- the extreme case is what happened
    Called
    "Only relevant on an extreme reversal-extension well beyond today's likely range"
    Actual
    Cleared with a $4,512.29 high, closing $74 above it

The tape

  1. Open (00:00-01:00 UTC)

    Opened $4,332.61, a marginal gap below Tuesday's $4,334.04 close, already carrying the prep's pre-session sweep-and-reverse note -- the session's defining early event (sweep to $4,324.39, reversal to $4,354.43) had happened just before the review window's first candle.

  2. Asian / London (01:00-13:00 UTC)

    Consolidated in a tight band roughly $4,326-$4,363, repeatedly testing but not clearing $4,367.14. The 08:00 UTC hour dipped to $4,334.99 -- a retest of the lower band -- before grinding back up through the London morning. This stretch behaved exactly as the range/digestion branch expected: contained, two-sided, no held break.

  3. NY open (13:00-15:00 UTC)

    The 13:00 UTC H1 candle closed at $4,367.21, a marginal, undisplaced clearance of the pivot that slipped back to $4,364.50 by 14:00 -- a fakeout by the prep's own $3-15 displacement standard, not yet a genuine break.

  4. NY overlap (15:00-16:00 UTC)

    The session's real ignition. The 15:00 UTC candle ran from a $4,361.82 low to a $4,442.09 high, closing at $4,438.69 -- an $74 hourly range that cleared $4,367.14, $4,376.04, and $4,396.75 in a single candle. The 16:00 UTC candle extended it further, to a $4,465.99 high. This is the exact window the prep's priors flag as reversal-prone (17-27% pullback-continuation) and it instead produced the day's decisive continuation leg.

  5. Late session (17:00-23:00 UTC)

    The rally continued through the 17:00 UTC bond auction (weaker demand, yield up to 5.204% from 5.163%) and the 18:00 UTC FOMC Minutes without a pause, pushing to a $4,512.29 high at 22:00 UTC before closing the day at $4,510.28 -- near the session high, $177.67 above the open.

  6. Closing posture

    Price closed near the day's high, $4,510.28 versus a $4,332.61 open, clearing every resistance level in the prep's Key Levels table, including the one explicitly reserved for an "extreme reversal-extension."

What we learned

The session unfolded well outside the expected parameters. Three material surprises, each routed per the day-type/driver-stack/tradability framework in `docs/reference/decision-frameworks/session-analysis.md` §6:

  1. Two consecutive nights of a wrong day-type call

    Yesterday's Whipsaw call graded partial with the day-type element missing; tonight's Range call missed outright as the session delivered a full trend day. Two in a row is a pattern, not noise -- route: re-examine precondition checks. Specifically, "the day after a large trend day defaults to range" needs an explicit override when the driver stack is silent (undetermined/pending on all four ranked drivers) rather than confirmed calm -- silence has now preceded the month's two largest moves in either direction, not a quiet digestion day.

  2. Silent driver stack, both times, ahead of the two biggest moves of the month

    Real yields and the dollar carried no fresh confirmed read on both the night before yesterday's $83.02 breakdown and tonight's $177.67 breakout. The stack read "undetermined/pending" both times, offering no lean toward the eventual result. Route: driver-stack read is mis-ordered/incomplete for this instrument at the current data feed -- propose a priors edit treating a fully silent stack itself as a volatility-expansion flag (widen the day-type distribution toward trend/whipsaw) rather than a neutral input that defaults toward range.

  3. The runner-up scenario called the right side but its target sequence undersold the move by roughly 3x

    The reversal/bounce branch's stated path ($4,367.14 -> $4,376.04 -> possible $4,396.75) was the correct direction and even had a working conditional trigger, but price closed $113 beyond its furthest stated target and $74 beyond a level explicitly reserved for "extreme" cases. Route: right scenario, undersized target -- tighten the target-sizing standard so that a silent driver stack widens stated targets rather than anchoring them to the nearest visible pivots, since silence has now correlated with outsized moves rather than contained ones.

Full notes

Footer

Reviewed prep: 2026-08-19-xauusd-session-preparation. Data source: MetaTrader 5 confirmed candles (D1, H1, H4) and the OpenAlpha session-review workflow.