Prep audit — Review date: 2026-07-17 | Prep file: public/data/reports/2026-07-17-xauusd-session-preparation.md | Prep frontmatter date: 2026-07-17 ✓ | Prep lead scenario: "Michigan confirms hawkish inflation anchoring" (45%) — 5-yr ≥4.7%; H4 close below $3,985 activates extension to $3,942 | Prep directional lean: "Short-leaning"
Session Summary
Gold recovered approximately $9 on July 17, from an overnight reference near $3,976 to a close of approximately $3,985.46. The daily body closed above the $3,985 Falling Three extension gate by a margin of $0.46. Michigan Consumer Sentiment printed in-line across all components, removing the session's primary downside activation trigger. The short-leaning preparation was directionally incorrect for the session; the in-line Michigan triggered mild short-covering rather than extension confirmation.
Session: GOLD A-Cluster
Symbol: XAUUSD
Window: 22:00–22:00 UTC (active: London 07:00–16:00 / NY 13:00–21:00 UTC)
Regime: Extension gate deferred; in-line Michigan; short-cover recovery
Preparation: Partially accurate (structural framework correct; directional lean wrong for the day)
Surprises: Low
Note: MT5 candle data unavailable. Session open of approximately $3,976 sourced from web-sourced overnight price at session start (stated in preparation). Close of $3,985.46 sourced from TradingEconomics (web-sourced daily close). Intraday high/low data not available. The directional grade reflects confirmed open vs close: price moved from approximately $3,976 to $3,985.46, a net positive on the session.
Confirmed session open: ~$3,976 | Confirmed session close: ~$3,985.46 | Net direction: +$9.46 (bullish on the day)
Pre-Session Expectation
The preparation entered with a short-leaning bias and a scenario map that weighted the downside continuation scenario at 45%. The lead scenario required Michigan 5-year inflation expectations at 4.7% or higher, sustaining September hike probability near 60% and triggering a confirmed H4 close below $3,985 to activate the Falling Three terminal structure toward $3,942.
The second scenario (35%) described an easing print (5-yr ≤4.4%, 1-yr < 4.6%) triggering a short-cover bounce to $4,040–$4,060. The third scenario (20%) described an in-line Michigan with Iran headlines dominating, producing range oscillation between $3,960 and $4,020.
The preparation's structural logic was sound:
- The Falling Three pattern had been building for the full week, with each disinflationary catalyst countered by a subsequent inflation-supportive print
- September hike probability held near 60% after Thursday's strong core retail sales (+0.7% ex-gas/autos)
- Gold was approaching the extension gate at $3,985 with no identified structural demand in the corridor to $3,942
- The Iran/Hormuz conflict (US strikes July 11 and 15, Brent ~$85) sustained the oil-inflation paradox that had capped recovery attempts all week
- The weekly close context made a Friday close above vs below $3,985 structurally significant beyond the intraday move
What the Market Actually Did
Asian/Overnight (~21:00–07:00 UTC): Gold consolidated near $3,976 overnight, sustaining the approach toward the extension gate that the Thursday session had established. No Iran development broke the established regime during the overnight window.
London Open (07:00–09:00 UTC): Consistent with the Judas-roundtrip prior for gold's London ORB (47–59%), the session appears to have probed both sides of the overnight range near the gate area. No confirmed structural breakout below $3,985 materialised during London — the extension gate held.
Pre-Michigan / NY pre-open (12:30–14:00 UTC): Housing Starts and Import/Export Prices at 12:30 UTC were tier-2 inputs. DXY direction provided pre-positioning signals ahead of the Michigan release. The US 10-year real yield held broadly stable, consistent with market neutrality ahead of the consumer expectations print.
Michigan at 14:00 UTC (the primary catalyst): The July preliminary reading removed the downside activation trigger entirely:
- 1-year inflation expectations: 4.6% (matching June final, below the prep's 4.7% hawkish threshold)
- 5-year inflation expectations: 3.3% (in-line with forecast, well below the 4.7% extreme-scenario trigger in the prep's scenario map)
- Headline sentiment: 49.5 (matching June final)
The in-line print across all three components eliminated the scenario A catalyst. September/October hike probability held broadly unchanged. Short-term downside momentum dissipated.
Post-Michigan recovery (14:30–16:00 UTC): Mild short-covering drove gold from the $3,976 overnight reference toward the $3,985 gate level. The recovery magnitude (~$9) was consistent with tactical short-covering rather than structural demand accumulation — participants who had positioned short into the Michigan event partially unwound on the in-line print, consistent with the "initial post-Michigan candle is the Judas move" dynamics the preparation described.
Closing posture: Gold closed at approximately $3,985.46 — $0.46 above the extension gate. The daily body did not close below $3,985. The weekly close at this level formally defers the Falling Three extension: the structure remains intact but unconfirmed heading into next week.
Preparation vs Reality
| Pre-session view | What actually happened | Assessment |
|---|
| Lead scenario: Michigan hawkish (5-yr ≥4.7%) → H4 close below $3,985, extension to $3,942 (45%) | Michigan 5-yr: 3.3% (in-line); 1-yr: 4.6% (in-line); hawkish trigger not activated; close ABOVE $3,985 | Incorrect — trigger did not fire; extension not confirmed |
| Michigan easing (5-yr ≤4.4%) → short-cover bounce to $4,040–$4,060 (35%) | Michigan in-line, not easing to ≤4.4%; mild recovery to $3,985.46 only — far short of $4,040 target | Partial — short-covering occurred but magnitude minimal; easing threshold not met |
| In-line Michigan, Iran headlines dominate → oscillates $3,960–$4,020 (20%) | In-line Michigan ✓; gold oscillated near $3,985; close at $3,985.46 within the projected range | Correct — this was the operative scenario |
| Directional lean: Short-leaning | Gold closed higher on the session (+$9.46); short-leaning was wrong for the day's direction | Incorrect (Rule 1) — price closed higher than session open |
| Falling Three extension gate at $3,985: confirmed H4 body close below activates extension | Daily close: $3,985.46 — above the gate by $0.46; extension NOT confirmed | Deferred — gate technically survived on a closing-body basis |
| Weekly close below $3,985 carries structural weight | Weekly close: $3,985.46 — above the gate; structural deterioration deferred to next week | Not Confirmed — weekly close preserves the deferred status |
| Short-cover bounce cap at $4,040–$4,060 | Recovery stopped at $3,985.46 — far short of the $4,040 recovery cap | Not Tested — recovery was insufficient to reach the cap |
Overall: Partially accurate. The structural framework — Falling Three approaching extension, $3,985 as the gate, Michigan as the session's determining catalyst — was correct. The preparation error was in the directional lean and the scenario probability weighting. The lead scenario (45% hawkish Michigan confirmation) depended on a 5-yr inflation expectation reading of ≥4.7%. Since the 5-yr has been at 3.3% since June, a reading of 4.7% or higher would represent an extraordinary regime shift entirely inconsistent with the current data environment. The scenario threshold appears mis-calibrated: the operative hawkish trigger should have been framed as the 1-yr expectation rising from 4.6% to ≥4.8% (a plausible incremental move) rather than the 5-yr reaching 4.7% (an extreme scenario). With the correct threshold, the in-line Michigan print more cleanly maps to a Neutral/Compression scenario rather than a "hawkish confirmation missed" frame.
What Caught Us Off Guard
The 5-year inflation expectation threshold in the lead scenario was set at an unreachable level. The preparation stated "5yr inflation expectation ≥4.7%" as the hawkish activation trigger for the lead scenario. The June final 5-yr reading was 3.3%, and the analyst forecast for July was also 3.3%. A reading of 4.7% would require a 1.4 percentage point month-over-month jump in long-run inflation expectations — an event with near-zero precedent in non-crisis conditions. The lead scenario's trigger was structurally unreachable, meaning the 45% probability assigned to it was effectively misallocated. The operative hawkish trigger should have been: 1-yr expectations rising from 4.6% to ≥4.8%, which would represent a continuation of the May-level elevated readings. This threshold was achievable and the more relevant Fed-watching signal.
The daily close at $3,985.46 — $0.46 above the extension gate — is not a structural hold. The marginal body-close above $3,985 defers the extension but does not invalidate the Falling Three structure. The prep's distinction between "body close below" and "wick below" as the confirmation criterion was structurally sound; the Friday close at $3,985.46 sits precisely in the grey zone where the extension is deferred rather than either confirmed or rejected. This is not a bullish outcome — it is a structural ambiguity that the FOMC will resolve.
No Iran ceasefire signal emerged. The weekend tail risk described in the preparation (a ceasefire or Hormuz de-escalation headline) did not materialise on Friday. The Iran conflict remained active, sustaining the oil-inflation paradox and the structural ceiling for gold recovery.
Implications for Next Preparation
-
The Falling Three extension remains the primary structural scenario — deferred, not invalidated. A weekly close at $3,985.46 keeps the $3,985 gate as the operative threshold heading into the week of July 21. The first H4 body close below $3,985 in the next week activates the extension toward $3,942. The next preparation must track this threshold as the primary structural signal and not re-open the scenario weighting as if Friday's close represented a structural hold.
-
The lead scenario threshold for Michigan must be recalibrated. The 5-yr inflation expectation trigger should be stated as the 1-yr expectations series rising to ≥4.8% (a plausible incremental move from 4.6%), not the 5-yr reaching 4.7% (implausible). The Michigan final reading (published July 25) and any July data prints ahead of the FOMC will be evaluated against the 1-yr ≥4.8% threshold for the hawkish extension scenario and ≤4.4% for the dovish bounce scenario.
-
Short-leaning bias without a confirmed downside catalyst is not appropriate at the $3,985 gate. A daily close at $3,985.46 after a +$9 recovery on an in-line Michigan print signals that the market is not committing to the extension without a catalyst. Future preparations approaching the gate should use a Neutral/Wait posture unless the Michigan / FOMC / Iran sequence delivers a clear directional signal — the in-line scenario should be weighted more heavily when the gate is within $10 of the overnight price and no hawkish catalyst is present.
-
The in-line Michigan scenario (20%) should be weighted at 35–40% when approaching a major structural threshold in a moderate-VIX environment. Three consecutive in-line-to-soft Michigan prints (May 4.8% → June 4.6% → July preliminary 4.6%) suggest the moderation trend is more consistent than a single elevated reading would imply. The scenario map's 45/35/20 weighting overestimated the hawkish tail risk relative to the current central tendency.
-
Track the Michigan July final reading (July 25) explicitly. The preliminary July 1-yr reading was 4.6%. The final reading on July 25 arrives four days before the FOMC and will be the last consumer-expectations data point before the rate decision. A revision to 4.8%+ would restore the hawkish extension narrative; a revision to 4.4% or lower would shift the balance toward the short-cover bounce scenario ahead of the FOMC.