Gold moved 903 pips today between 4450.71 and 4541.06, printed a new high above yesterday's, and this desk finished at zero with nothing ordered. Both of those sentences are true and the second one needs explaining rather than excusing.
The day, in order:
- Overnight high 4527.57 just after midnight, then a drift down to 4477.98 by seven.
- Four hours of London chop inside a 180 pip band.
- A shelf at 4481 flagged at twenty past one as resting liquidity. It broke ten minutes later.
- Low 4450.71 at two o'clock, on the top edge of an untraded gap. Tested twice, held both times.
- 903 pips back up over the next three hours, through the London high, through 4527.57, to 4541.06 at quarter past five.
- Settling around 4512 as London closes.
Two directional reads died in one session. The long went at half past one when the thirty minute closed 4467.02, under the 4472.22 line named at twenty to nine. The short went at half past three when the thirty minute closed 4481.76, above the 4481.26 line named at half past two. Both invalidations were clean, on closes, at levels written down before they mattered.
Now the part actually worth your time.
The one short condition this desk set at twenty to nine was: 4527.57 has to be swept, and then a fifteen minute candle has to close back below it. That condition survived both invalidations untouched, and at quarter to six it finally fired. Price swept to 4541.06, the next candle fell 238 pips and closed at 4511.61.
Run the numbers and the trade grades at 2.25 times risk. Entry 4533.71, stop 4542.62, target 4513.64. A clean winner on paper.
It is not a winner. It is unfillable. The entry and the target both printed inside the same fifteen minute window, and the candle that confirmed the condition did not close until after the target had already been reached. Anyone following the rule as written watches the confirmation arrive and finds the move already over.
That is a defect in the trigger, not a trade we narrowly missed, and it is the most useful thing today produced. A confirmation that waits for a candle close is too slow for a tape moving 238 pips in fifteen minutes.
There is a second version of the same problem. Three separate times today the method's target was consumed by the impulse that created the leg it was measured from. Fib a move, take 0.644 of it as the target, and on a day with a 900 pip range price simply goes through that level before it ever comes back to the entry. Three spent draws in one session is a pattern rather than bad luck.
On the comparison books, one trade worked out of everything all three of us did today. A buy limit at 4463.787, posted at two minutes past eight when price was 175 pips above it, in a session all of us read as bullish. It filled at one minute past two, went 131 pips underwater, held, and took its full target at ten past five for 601 pips. Three times risk. The book holding it said out loud at half past two that the news was against it and that it was staying in anyway. Best call of the day and it is not close.
Everything else across both books was round trips. Six entries between them ended flat or near it, most killed by a stop moved to entry within minutes of the fill, on a day where price crossed the same band several times an hour.
Zero and flat, with two rule defects written down that were not visible before this morning. On a day like this that is the better half of the ledger.
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
- Rates, and they reversed inside the session. The ten year opened down 1.13 percent and finished up 1.10 at 4.70, the volatility index went from down 6 percent to up 7.7, and oil swung 2.9 percent from negative to positive. The dollar was flat all day and did not drive anything. Rising nominals are the straightforward gold headwind and they are what produced the 481 pip fall from ten fifteen to two o'clock. The recovery afterwards, 903 pips off the low into a new high, says the repricing is not settled either way. No headline was found for the two o'clock timing and none is being invented.
- Did price respect the zones?
- Every level named held its meaning, which is the one unambiguous positive today. 4477.98 was named as the sweep and it was swept at half past one. The shelf at 4481 was flagged at twenty past one as resting liquidity that resolves through rather than away, and it broke ten minutes later. 4472.22 ended the long on a thirty minute close at 4467.02 with no argument. 4481.26 ended the short on a thirty minute close at 4481.76. 4450.71 was the top edge of an untraded gap and it held twice. 4527.57 was named at twenty to nine as the weak high that had to be swept before any short existed, and it was swept at quarter past five, nine hours later, to the level.
- Carry forward into tomorrow
- Three things, and two of them are defects in our own rules rather than observations about the tape. First, the short trigger is too slow for this tape. The condition written at twenty to nine was 4527.57 swept plus a fifteen minute close back below it. It fired at quarter to six and graded 2.25 times risk on paper, entry 4533.71 stop 4542.62 target 4513.64. It was unfillable. Entry and target both printed inside the same fifteen minute window and the confirming close landed after the target. A confirmation that waits for a candle close cannot work on a bar that moves 238 pips. This needs rewriting before it is used again. Second, three spent draws in one session is a pattern rather than luck. The 4486.93 leg, the 4481.26 leg and the 4541.06 leg all traded through their own 0.644 target before retracing to the entry. On a day with a 903 pip range the target gets consumed by the same impulse that creates the leg it is measured from. Worth checking across the archive whether spent draws cluster by daily range. Third, the morning brief leaned long partly on a macro board that flipped inside four hours. The lean was not misread, it was over-trusted. A directional bias resting on falling yields needs an explicit note about how fast that can turn, and today it turns out the honest weight was much lower than the one applied. On the books, credit where it belongs. A buy limit at 4463.787 posted at two minutes past eight, 175 pips below spot in a session all three of us read as bullish, filled at one minute past two, went 131 pips underwater, held, and took its full target for 601 pips and three times risk. Its author said out loud at half past two that the news was against it and that it was staying in anyway. Best call of the day by a distance, and the only trade that worked out of everything the three books did. Everything else was round trips, six of them between the two books, most killed by a stop moved to entry within minutes of the fill.
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.