Core PCE printed 0.1%. The dollar cracked, gold ripped through 2,360, and the week-long short-gold trade unwound in twenty minutes.
The risk we flagged is exactly what happened. Core PCE came in at 0.1% month-over-month, a tenth below consensus, and the reaction was the violent repricing that crowded positioning always produces. Front-end yields dropped eight basis points in the first half hour, the dollar index gave back the entire week's gain, and gold tore through the 2,360 ceiling that had capped it for five sessions, printing 2,388 before lunch.
The morning's neutral calls on both equity indices were the right posture: standing aside meant you weren't long the most rate-sensitive tech names into a print that, had it gone the other way, would have hammered them. As it happened the cool number lit a broad rally, the S&P proxy broke its range to the upside and, notably, breadth was wide for the first time all week. When the catalyst finally arrives, it tends to move everything, not just the leaders.
Core PCE at 0.1% versus 0.2% expected. A single tenth of a percent, but into one-sided positioning it was enough to unwind a week's worth of dollar strength and short-gold conviction inside one session. The smallest miss can produce the largest move when everyone is leaning the same way.
Being directionally right and being positioned right are not the same thing. The bearish-gold lean was correct for four sessions and then wrong in twenty minutes, and the reason it was so painful for those who held it is that everyone was on that side. The lesson isn't 'don't trade trends.' It's that a crowded trade carries a tail risk priced into nobody's stop. When the whole tape is leaning one way into a binary event, the right size is smaller than your conviction wants it to be.