Gold coils at the daily pivot while the dollar tests its 4h ceiling

Gold is doing the most boring thing a $4,700 gold can do — parked almost exactly on its daily pivot while DXY quietly stages its first credible breakout attempt of the month. The setup is asymmetric: bulls need the dollar to fail at 4h resistance, bears need today's US retail sales to confirm what hot PPI started. Daily MACD is still rising and managed money is still long, but the 4h has already broken below its EMA20 and yields jumped 11bp overnight. Until 4,720 prints clean or 4,677 fails, the right trade is fading the edges of a 50-handle box and respecting whichever side breaks first.

The session

Spot opens the European morning at 4,709.70, flat on the intraday tape, up 0.21% on the week, and still down 1.88% on the month after the post-CPI/PPI air-pocket the prior two sessions. The recent multi-session range — 4,645.20 to 4,783.40 — has compressed into a 50-handle box today; Investing.com's tape narrative ("Gold coils in $4,675–$4,725 range with ADX 11.57") matches what the snapshot shows numerically — directional strength is exhausted.

What actually moved overnight is the dollar. DXY printed +0.46% to 98.49 and the 10y reopened +11bp at 4.48%. Gold not breaking lower against that is the only constructive thing bulls have today; the discount on PAXG (-0.47% vs spot, per the Binance tape) tells you the crypto-side bid is sleepy, but the COMEX bid is absorbing the yield pop. That's not strength — it's stalemate. Yesterday's TradingEconomics-flagged narrative is unchanged: April CPI at 3.8% (highest since May 2023) and PPI accelerating to its fastest since 2022 have pulled Fed cuts out of the curve and surfaced talk of another hike before year-end. Gold has, so far, refused to follow yields to the downside. The question is whether retail sales today gives the dollar the final shove.

Multi-timeframe read

TF	Bias	Read
15m	Mildly long, exhausted	RSI 55, above all EMAs, MACD hist 1.45 but rolling. A finished impulse, not a reversal.
1h	Constructive long	RSI 53.6, above EMAs, MACD hist 1.49 and turning up. This is the cleanest bull frame.
4h	Bear-flag posture	RSI 48.8, below EMA20. The 4h is where short-side conviction lives.
1d	Long under-the-hood	RSI 51.2, MACD hist +13.34 and rising, above EMA200 — but still below EMA50.

The picture: the daily wants higher, the 1h is happy to take it there, and the 4h is the gate. Until 4h reclaims its EMA20 (which lines up almost exactly with R1 4,720.43), every rally is a 4h-shorter's gift. Below 4,677.23, the 1h flips and the daily MACD stalls fast. Where the timeframes agree: the trend is undecided and consolidation is the prevailing read. Where they diverge: 1h and daily lean long; 4h alone leans short. That divergence is what the box is.

Macro frame

DXY is the prime mover and the snapshot tells you exactly how loaded it is. 1h RSI 54.3 and 4h RSI 66.8 — the 4h is approaching overbought without yet being there, daily MACD hist is +0.04 and turning. Crucially, on the daily, DXY is above its EMA20 but still below its 50 and 200; this is a reflexive bounce inside a broader downtrend, not a regime change. The 30-day daily-return correlation between DXY and XAU sits at -0.68 — strong enough to dominate when DXY actually trends, weak enough that gold can decouple in 4–6 hour windows.

US 10y at 4.48% up 11bp matters more than the spot DXY print. The cleanest read I can give without TIPS data (snapshot does not include real yields — data gap) is that the nominal pop is being driven by hot prints, not growth optimism, so real yields are likely up commensurately. That is the textbook gold headwind. India's import-tariff hike (6% → 15%, per Trading Economics' news flow) is a slow demand-side bleed, not a flow-of-the-day driver, but worth holding in mind for any extended weakness. No G10 central-bank decisions today.

Two scenarios

These are qualitative conviction reads on a single-session horizon, not back-tested probabilities. They sum to less than 100% because consolidation inside the box is the third (unstated) outcome.

Buy setup

• Trigger: Reclaim of daily pivot 4,699.97 followed by a 1h close above R1 4,720.43 — ideally with DXY rejecting at or under the 98.80 zone
• Invalidation: 1h close back below 4,677.23 (daily S1), or DXY 1h close above 98.80
• Target: First 4,743.17 (R2), then 4,763.63 (R3) — the prior session high cluster
• Conviction: 40%
• Rationale: Daily MACD is rising into a positive print, the daily is above EMA200, and managed-money net length (+163,303 per the 2026-05-05 COT) is supportive but not blow-off. If DXY's 4h RSI rolls from 66.8 without breaking 98.80, the 4h reclaims its EMA20, and the box resolves up. The trade is only valid if DXY confirms — gold leading the dollar lower is the rarer setup.

Sell setup

• Trigger: 1h close below S1 4,677.23 with DXY printing above 98.65
• Invalidation: 1h close above R1 4,720.43, or DXY rolling under 98.20
• Target: 4,656.77 (daily S2) into 4,656.43 (weekly pivot) — that is the key confluence on the chart — then stretch 4,634.03 (daily S3)
• Conviction: 50%
• Rationale: 4h already below EMA20, yields up 11bp, Fed cuts priced out, dollar 4h heating. If retail sales today print hot and DXY clears 98.80, the path to weekly pivot opens quickly and there is no meaningful technical support between S1 and the S2/weekly pivot confluence ~20 handles below. The COT positioning skew also means there are stale longs to flush.

The remaining ~10% lives in the "do nothing, stay in the box" outcome — and frankly that is the most honest read of the ADX 11.57 tape until a US data print breaks the tie.

Levels worth marking

Top-down, with confluences:

• 4,763.63 — daily R3, aligns with the prior multi-session high zone (4,765.20)
• 4,743.17 — daily R2; typical institutional fade level on first touch
• 4,720.43 — daily R1; clusters with where the 4h EMA20 sits given the snapshot's "above EMA20" flip-point. The bull trigger.
• 4,699.97 — daily pivot. Magnet. Spot 10 handles above.
• 4,687.74 — PAXG mark (cross-check, not a level)
• 4,677.23 — daily S1. The bear trigger.
• 4,656.43 / 4,656.77 — weekly pivot and daily S2 in tight confluence. The key downside level of the week. Loss of this opens room.
• 4,634.03 — daily S3, also near the lower end of the recent 5-session band (4,645.20 was last week's session floor)
• 4,576.67 — weekly S1; tail target if 4,634 fails
• 4,512.70 — last week's low, the actual capitulation level

Calendar / catalysts

• Today (Thu 14 May): US April Retail Sales is the marquee print (flagged by FXStreet's note). Initial jobless claims same window. Both are dollar-positive on upside surprises given the current rate-path debate.
• Friday (15 May): Typical end-of-week — UoM consumer sentiment / inflation expectations usually here; the inflation-expectations subcomponent is the one that moves gold given the CPI/PPI backdrop.
• Heads-up: Trump–Xi meeting traffic is on the wires (per Investing.com news flow). That's tail risk in both directions for gold — escalation lifts it, breakthrough kills it.
• COT refresh: Friday afternoon (CFTC). Worth watching whether managed-money long got trimmed on the PPI move.

Data gap: I could not load ForexFactory's calendar directly (403) — the calendar list above is reconstructed from FXStreet's reference to retail sales plus the usual Thursday/Friday US slate, not from a fresh primary fetch. Verify before sizing.

Sources cited

• fxstreet.com/markets/commodities/metals/gold — technical structure, retail-sales flag
• tradingeconomics.com/commodity/gold — CPI/PPI/Fed-path narrative, India tariff, YoY context
• investing.com/commodities/gold — multi-timeframe signal cross-check, ADX/box framing, Trump–Xi headline
• (Attempted but failed: reuters.com, dailyfx.com, marketwatch.com, forexfactory.com — degraded gracefully)

(not financial advice)