Gold has run 4162 pips in six sessions. This morning was the first serious pushback and by the afternoon it had been bought back.
The overnight move did the damage in both directions:
- 04:15, price prints 4435.34, taking out Monday's high at 4395.31 on the way.
- 06:35, a twenty-bar base between 4408 and 4421.62 gives way.
- 07:30, one five minute bar covers 198 pips.
- 09:20, the low at 4357.76. That is 776 pips off the high in five hours.
We read that as a risk premium coming off. The dollar was flat at 99.85 and the ten year yield was up at 4.70, both of which normally lean on gold, and gold had gone up anyway all week. When the metal rises while its two usual headwinds also rise, you are looking at fear being priced rather than rates. Fear reprices fast, and Trump saying the Strait of Hormuz was open took a chunk of it out.
That framing held up. What did not hold up was the trade we built on it.
We mapped a short from the base: entry 4413.93, stop 4424.86, target 4372.22, a little under four times risk. It sat around 400 pips above the market for most of the day and never filled.
Then two things happened within four minutes of each other, and they pointed the same way.
At 14:47 our own trigger fired. We had said at the open that a thirty minute close above the daily open at 4389.89 would mean the morning selloff was premium coming off a headline rather than a turn. At 13:20 the bar was live and above it and we refused to call it, because the criterion was a close and not a touch. At 14:47 it genuinely closed, at 4391.40, the first one all day. We honoured it. A criterion you only respect when it suits you is not a criterion.
At 14:51 the trader pulled his order, and his reason was better than ours.
He had placed a sell limit at 4406.70 earlier in the afternoon, at the 0.764 rather than at our 4413.93, taking a smaller reward for an easier fill. He removed it because he did not like the daily candles. He is right, and it corrects what we wrote this morning.
Look at the daily closes: bull, bull, bull by 1703 pips, one small bearish day, bull by 1017, bull by 480 with almost no upper wick, and today so far a bullish body with a 330 pip lower wick. Six bullish closes in seven. Monday finishing at its highs is about as strong a close as a chart gives you.
We had written that the six session uptrend was context rather than a closed higher timeframe bar, and named the one hour chart as the one that mattered. That was the weaker read. Monday's close is a closed higher timeframe bar and it leaned hard against a short. Our own rule is that the higher timeframe wins; we walked down from the hourly and never walked up past the four hour.
So the plan changed, and it changed for the better.
No shorts now. The condition is that price has to go up and fail first: a sweep of the daily high at 4435.34 followed by a fifteen minute close back below it. Only then do we look for a setup, and it still has to pass the normal checks.
There is a reason that level and not a nearer one. Today's grid anchored to the base high at 4421.62 while the real resting liquidity sat 138 pips above it at 4435.34, which is the highest price in the last two hundred hours. We built the grid underneath the liquidity instead of at it. A failure there would put the anchor where it belongs, and it would be the first genuine topping signal on the daily rather than a short into an uptrend.
It is 343 pips away, so it may well be a tomorrow problem. Tomorrow is CPI.
On the news gate, one thing worth being straight about. It has been on almost the entire session, but for most of the day it was on for a poor reason: the same two statements from around half past nine kept being republished by new outlets with slightly different wording, and our timer restarts on wording. At 15:02 a real one arrived, a warning of a major attack on Iran if nuclear talks fail. That is new information and it is an escalation, which is a reason for gold to be bid rather than sold.
Flat. Nothing ordered, one conditional level being watched.
Levels
Trade decision log
Flat — nothing mapped
No call on the board. The desk is watching, not chasing — most sessions end here, and that is the method working rather than a gap in the record.
Before & after
The morning chart, left un-redrawn
The striking feature
Morning thesis vs market reality
End-of-day verdict
What we learned
- Driver that mattered most
- Iran and the Strait of Hormuz, in both directions and mostly noise. Trump saying the strait was open took a chunk of risk premium out overnight and produced the 785 pip range. After that the story kept generating headlines without generating direction. The cross-market picture never confirmed a turn: the dollar sat flat at 99.83 all day and the ten year yield stayed at 4.70, so gold rising against both of those was a fear bid rather than a rate one, and fear bids do not unwind on a schedule.
- Did price respect the zones?
- Yes, and the one that mattered most was the daily open at 4389.89. Price spent the whole morning below it, closed three consecutive 30m bars above it between 14:00 and 15:00, then failed and closed the day 83 pips underneath. Our short grid's fib 0 at 4357.76 was swept by nine pips at 09:45 and held, which is why the grid barely moved when we re-anchored. The 4435.34 high from 04:15 was never revisited.
- Carry forward into tomorrow
- The standing condition agreed with the trader: no shorts until price sweeps the daily high at 4435.34 and closes a fifteen minute bar back below it, then a normal setup through the checklist. The reason that level and not a nearer one is that today's grid anchored to the base high at 4421.62 while the real resting liquidity sat 138 pips above it. A failure there puts the anchor where it belongs and would be the first topping signal on the daily. The unfilled demand at 4339.98 to 4343.44 is still untouched and still the trend-aligned long zone. CPI lands tomorrow at 14:30, so anything setting up in the morning arms into a red folder print.
The method
Two anchor prices on the displacement leg, and a fixed fib grid derives the whole order: entry floored at the 0.764 that names the method, stop at 1.05, target a fixed span below. R is a consequence of the grid, never a number chosen after the fact. Every read is timestamped and left un-redrawn.