Gold's first real flush: 164k of spec longs hit the exit at 4,600

A 5%+ intraday range that started near 4,830 and ended through the daily S3 pivot is what a +164,006 managed-money long book looks like when it gets washed. Short timeframes are properly oversold, daily MACD is still pointing down, and the week's risk stack — FOMC Wednesday, advance GDP and core PCE Thursday — is entirely ahead of us. Weekly S2 at 4,576.77 is the only level that matters into Powell. This is a positioning unwind dressed up as a macro event; treat it as such.

The session

Gold prints 4,591.10 at 13:30 UTC, down 4.67% on the day. The intraday range is 4,567.60 to 4,828.80 — about $261, roughly 5.4% peak-to-trough. That eclipses the prior five-session range of 4,567.60 to 4,754.20 in both directions: today's high broke up and today's low broke down. Price is now below the daily S3 pivot at 4,627.80 and sitting on the weekly S2 at 4,576.77.

The wire framing (per FXStreet's gold wrap) credits dollar strength on safe-haven flows tied to the second round of US–Iran nuclear talks and unresolved Hormuz Strait friction. That fits the tape but doesn't explain its scale. DXY at 98.77 is +0.64% intraday and US 10y is +7bp to 4.37% — neither is a "5% gold day" macro catalyst on its own. The size of the move is positional. The 21 April COT had managed money at +164,006 net long (212,893 gross long), commercials at -202,940 net short. That is a crowded book heading into a known event-risk week, and today is what it looks like when the first tranche of weak hands gets stopped out.

PAXG on Binance prints 4,576.09, -2.33% on the 24h, at a -0.33% discount to spot. Tokenised gold underperforming and trading at a slight discount is consistent with crypto-side risk-off running a hair ahead of the spot tape — not a panic signal, but not a bid either.

Multi-timeframe read

The four timeframes agree on direction and disagree on momentum, which is the configuration you want at a flush low — not the configuration of an ongoing trend leg.

• 15m: RSI 35.9, MACD hist +0.12 and turning up. Below all three EMAs. Short cycle is trying to base.
• 1h: RSI 28.0 — properly oversold. MACD histogram still negative at -7.80 but inflecting up. Below EMA20/EMA50.
• 4h: RSI 32.6, MACD hist -12.20 and flat. Below EMA20. Stretch from EMA20 is what you'd expect after a one-shot drop, not a grinding downtrend.
• 1d: RSI 40.7, MACD hist -6.04 and still pointing down. Below EMA20 and EMA50 but above EMA200. The longer trend is intact; this is a counter-trend washout that has cleared a meaningful chunk of speculative excess.

Where they agree: trend is bearish on every timeframe shorter than the daily, and price is below the EMA20 on all four. Where they diverge: 15m and 1h MACD histograms are inflecting up while the daily MACD is still extending down. That gap is the seam. If 15m/1h base while the daily catches down, you get a reflex bounce into 4,650–4,700. If the daily keeps extending and pulls the lower TFs back into trend, the next leg targets weekly S3 at 4,496.

Investing.com's technicals summary corroborates the picture: Strong Sell on 30m, hourly, 5h and daily; Sell weekly; Strong Buy monthly. The structural bull is intact above; the cyclical bull is broken below.

Macro frame

The 30-day daily-return correlation between DXY and XAU is -0.01 — effectively zero. Gold is not trading off the dollar right now; it is trading off positioning and rate expectations. That matters for how you read the macro tape.

• DXY 98.77 is firm but not breaking out. 1d RSI 49.9 and a flat daily MACD histogram say the index is consolidating, not trending. 1h RSI 62.2 with the histogram rolling means short-term DXY is itself overbought into the FOMC.
• US 10y at 4.37%, +7bp is the bigger tell. Yields rising into a Fed meeting means the bond market is not pricing a dovish surprise. FXStreet flags CME FedWatch at roughly 35% probability of any Fed cut by year-end. Real yields are gold's structural problem; until that probability moves materially, every rally on the daily fights the rates tape.
• Non-US central banks: the BoJ held at 0.75% overnight in line, per the TradingEconomics calendar. ECB and BoE are both on Thursday's calendar with consensus holds. No rate-differential surprise expected to widen DXY against its peers — which argues against DXY breaking out, and by extension against the gold flush extending on macro grounds. The risk into the rest of the week is positional and event-driven, not structural.

Two scenarios

These are qualitative reads, not back-tested probabilities — the conviction figures reflect honest confidence, nothing more.

Buy setup

• Trigger: Reclaim of 4,627.80 (daily S3) on a 1h close with the 1h MACD histogram fully positive. Cleaner version: 1h close back above daily S2 at 4,651.60.
• Invalidation: 1h close below 4,576.77 (weekly S2). Below that there is no structural support in this snapshot until weekly S3 at 4,496.03.
• Target: First take-off at daily P 4,687.30. Runner into the 4h 200-SMA confluence at 4,723 (per FXStreet) which lines up with daily R2 at 4,723.00 — clean overlap.
• Conviction: 55%
• Rationale: Lower-timeframe momentum is already inflecting (1h RSI 28, histogram turning), price is sitting on weekly S2, daily trend remains above the 200 EMA. This is a counter-trend mean reversion against a well-defined invalidation — the trade that pays when the spec long flush has done its job and the macro book hasn't actually broken yet.

Sell setup

• Trigger: 4h close below 4,576.77 (weekly S2) with the 1h MACD histogram failing to cross zero on any bounce attempt.
• Invalidation: 1h close back above 4,651.60 (daily S2) — that says the flush is being absorbed.
• Target: First take-off into 4,500 round / weekly S3 4,496.03. Stretch only on a confirmed daily close below; there is no cleaner support level in this snapshot beyond weekly S3.
• Conviction: 35%
• Rationale: Daily MACD is still pointing down, COT positioning is extreme on the long side and these unwinds rarely finish in one session, and Wednesday's FOMC is binary risk into a market that is not pricing dovishness. Lower conviction because you are selling into already-oversold short-cycle indicators on the back of a $260 range day.

Levels worth marking

• 4,828.80 — today's high; only relevant on a violent V-bounce.
• 4,803.03 — weekly R1.
• 4,734.90 / 4,723.00 — daily R3 / R2 cluster, plus the 4h 200-SMA per FXStreet at ~4,723. Heavy confluence; likely cap on any reflex rally.
• 4,699.20 — daily R1.
• 4,687.30 — daily P; first real upside magnet for a recovery.
• 4,663.50 / 4,651.60 — daily S1 / S2 — reclaim zone for the bull case.
• 4,627.80 — daily S3. Price is below this.
• 4,591.10 — last.
• 4,576.77 — weekly S2. The level that matters into the FOMC.
• 4,567.60 — today's low.
• 4,496.03 — weekly S3 / bear target if S2 fails.

Calendar / catalysts

The next 48 hours carry an unusually concentrated risk stack (per TradingEconomics calendar):

• Tue 28 Apr (today, later in the US session): CB Consumer Confidence. Second-tier; consensus 88.9. Not the main event but tape-relevant.
• Wed 29 Apr ~18:00 UTC: FOMC rate decision and Powell presser. Consensus hold at 3.75%. The bar is the dot plot and Powell's tone. Anything that lifts the year-end cut probability materially above 35% is gold-positive; anything that pushes it down fuels further unwinds.
• Thu 30 Apr 12:30 UTC: US Q1 advance GDP, core PCE YoY (consensus around 3.0%), initial jobless claims. Core PCE is the real-yield input gold cares about most.
• Thu 30 Apr 11:00 / 12:15 UTC: BoE then ECB rate decisions. Both expected to hold; relevant for DXY cross-currents rather than gold directly.
• BoJ already held at 0.75% overnight — neutral.

A FOMC + advance GDP + core PCE + ECB + BoE stack inside 48 hours is precisely why a +164k spec long book was vulnerable today. The week's binary risk is in front of us, not behind.

Sources cited

• https://www.fxstreet.com/markets/commodities/metals/gold (technical levels, Iran / FOMC framing, FedWatch ~35%, 4h 200-SMA at 4,723)
• https://www.investing.com/commodities/gold (multi-TF technical summary, prior close 4,693.70, MontesDeOca VC PMI mean reference)
• https://tradingeconomics.com/calendar (BoJ, FOMC, GDP, PCE, BoE, ECB schedule and consensus)
• (Attempted: Reuters commodities, DailyFX, Kitco — Reuters and DailyFX returned blocks; Kitco home page returned no article body.)

(not financial advice)