Pinned at the pivot

Gold opens the week sitting almost exactly on its daily pivot at 4,523, having traded a 138-point intraday range that swallowed both directional attempts. The bounce attempt has fingerprints — MACD histograms turning higher across 15m, 1h, and daily, DXY leaking against its short-term EMAs, 10y down a hard 11 bp — but daily structure is still pulling lower and managed money is sitting on a heavy long. Whoever wins the 4,527 – 4,534 cluster sets the tone into the back half of the week.

The session

Spot is changing hands at 4,523.20, down 0.75% intraday and -4.22% on the month, with the Asia/early-Europe tape carving a 4,455 – 4,593 range — roughly 3% peak-to-trough in one session. That kind of range is usually a headline-driven flush followed by a mean-reversion grab, and the print confirms it: we're back on the daily pivot (4,523.47), with the body of the session caught between weekly pivot 4,518.80 and daily R1 4,527.83.

Per Trading Economics' tape this morning, "Gold climbed toward $4,600 an ounce on Monday, recovering losses from last week as increasing optimism over a potential US-Iran agreement eased concerns about inflation and interest rate hikes." That fits the DXY behaviour — index back at 99.04, down 0.24% intraday and -0.26% on the week — and the 10y dropping 11 bp to 4.56%. The dollar/yield move is doing more of the lifting for gold's bid than gold's own demand picture.

PAXG/USDT on Binance is trading 4,557.41 — a 76 bp premium to spot. Modest, not panicky; the crypto-side proxy is willing to pay for ounces but is not throwing fists. Compared to the spikes seen earlier in the Middle East escalation cycle the premium is essentially benign.

Multi-timeframe read

The four frames split cleanly in two:

• 15m and 1h: constructive. Both above EMA20/50, RSI in the 50s (15m 56.6, 1h 50.7), MACD histograms turning up. This is what a bounce attempt looks like before it has earned a 4h confirmation.
• 4h: stuck. RSI 51.9 but price below EMA20, MACD not yet committed in the snapshot. 4h is the swing-trader's frame and the one that has to flip to make the bounce credible into Friday.
• 1d: wounded but healing. RSI 40.2 is the standard pullback-to-trend zone in a bull, and the histogram is rising off -9.44. Price is below daily EMA20 and EMA50 but still above EMA200 — the secular uptrend frame is intact, this is a pullback inside it.

Agreement: momentum is turning higher on every frame that has a printed MACD. Disagreement: shorter frames are already above their trend EMAs, daily is not — so the bounce has technically begun but has not yet reclaimed the levels institutional desks track. Until the 4h flips, this is a tactical play.

Macro frame

DXY at 99.04 is the cleanest tell. Daily-frame DXY is still constructive (above EMA20/50/200, RSI 58.4), but 1h and 4h have both rolled — RSI 37 on both, below near-term EMAs. The 30-day rolling daily-return correlation between DXY and XAU sits at -0.66, deeply inverse. So a DXY that is short-term-soft inside a still-firm daily uptrend implies a gold bounce window that is real but bounded.

US 10y at 4.56% with an 11 bp daily drop is the bigger ticket. That isn't noise — it's the rates market responding to either flight-to-quality or a re-pricing of the front end. Trading Economics flags the latter: "Fed Governor Christopher Waller signaled the central bank should no longer maintain an easing bias," with markets now pricing roughly a 55% probability of at least one 25 bp hike by October. The real-yield read is the print we don't have — TIPS isn't in this snapshot — but a nominal yield down 11 bp on a day when oil and breakevens weren't obviously crashing implies real yields softened, which is mechanically gold-supportive.

FXStreet's technical desk reads it the same way with the right caveat: gold is "holding within a downward parallel channel" and the MACD turn is "a tentative recovery rather than a clear trend shift." That matches the snapshot — the histogram is up, but the channel is not.

Two scenarios

The conviction numbers below are honest qualitative confidence, not back-tested probabilities. Don't size off the number.

Buy setup

• Trigger: Reclaim and 15m close above 4,527.83 (daily R1), with DXY remaining below its 1h EMA20.
• Invalidation: 1h close back below 4,512.27 (daily S2). That breaks the reclaim narrative and re-opens the weekly S1 air pocket.
• Target: First take at 4,539.03 (daily R3); second take at 4,572.50 (weekly R1) where the 4h EMA20 likely sits and where the prior 5-session high (4,582.60) clusters.
• Conviction: 55%.
• Rationale: MACD histograms turning up on three of four frames; daily RSI at 40 in the standard pullback-to-trend zone for a bull; DXY leaking on shorter frames; 10y down hard. The constraint is that 4h structure has not flipped, so this is a tactical reclaim, not a swing entry.

Sell setup

• Trigger: Rejection wick in the 4,534.67 – 4,539.03 cluster (daily R2/R3) on a 15m print, with DXY printing a 1h higher-low.
• Invalidation: 1h close above 4,572.50 (weekly R1). At that point the bounce becomes the trend.
• Target: First take at the daily pivot 4,523.47; second at 4,512.27 (daily S2); stretch to 4,505.43 (daily S3).
• Conviction: 45%.
• Rationale: Daily structure remains below EMA20 and EMA50; FXStreet's downward parallel channel is intact; managed money is heavily long (+159,833 net per the 2026-05-19 COT, longs 211k against shorts 51k), and any rejection at obvious resistance carries an unwind tail to chase. The trade is selling failed bounces at the cluster, not selling fresh.

Levels worth marking

From the snapshot, with confluence notes:

Resistance, near to far:

• 4,527.83 — daily R1. First reclaim level.
• 4,534.67 — daily R2.
• 4,539.03 — daily R3. Triple-stack of dailies inside a six-dollar band — expect chop.
• 4,572.50 — weekly R1. Confluence with the prior 5-session high at 4,582.60 just above. This is the level that flips the read.
• 4,624.00 / 4,677.70 — weekly R2/R3. Off the table without a fundamental catalyst.

Support, near to far:

• 4,518.80 — weekly pivot. Floor of the immediate range.
• 4,516.63 — daily S1. Sits inside the weekly pivot zone — confluence.
• 4,512.27 — daily S2.
• 4,505.43 — daily S3. Last line before the air pocket.
• 4,467.30 — weekly S1. Near the 5-session low (4,465.10) — a magnet if S3 fails.
• 4,413.60 — weekly S2. Below this is the FXStreet $4,360 channel-base scenario.

The PAXG +0.76% premium and the COT managed-money long stack both argue against an immediate flush through weekly S1; the structural break case requires a catalyst.

Calendar / catalysts

ForexFactory's calendar refused the fetch this morning (403), so this is a partial data gap and I'm limiting myself to what the qualitative sources actually confirmed:

• Fed speakers — Trading Economics has Waller's "no easing bias" line live in the tape; expect every FOMC appearance this week to be parsed for confirmation or pushback. The number desks are trading is the ~55% probability of at least one 25 bp hike by October.
• US-Iran diplomacy — the dominant headline driver. Trump's stated maintenance of the Strait of Hormuz blockade until a formal agreement is the geopolitical on/off switch. Any signal of breakdown re-bids gold; any signal of agreement leans the other way.
• Month-end flows — May 29 is the last trading day of the month. Mechanical rebalancing into a month where gold is -4.22% typically generates buying from underweighted funds; tailwind into Friday, not a trade in itself.
• PCE — the standard end-of-May US inflation print falls in this window and is the natural macro pivot for the week. Confirm the exact time on your own calendar before sizing — I couldn't pull the schedule cleanly today.

Sources cited

• https://www.fxstreet.com/markets/commodities/metals/gold — downward parallel channel, $4,650/$4,360 boundaries, "tentative recovery rather than a clear trend shift."
• https://tradingeconomics.com/commodity/gold — US-Iran agreement narrative, Waller "no easing bias" line, ~55% October hike probability, YTD +36.41% / month -2.53% context.
• Snapshot data (deterministic, fetched by code) — all prices, RSI/MACD, EMAs, pivots, CFTC COT, PAXG print.

Sources attempted but unavailable: Reuters (blocked), Kitco (header only), DailyFX (403), ForexFactory (403). Calendar is therefore a partial data gap and is flagged as such above.

(not financial advice)