Took the ATOM scale-out: 3,143 of 9,430 @ 2.034, +14.6% on book. Trigger fired on the OLDEST fill's avg (1.7684 x 1.12 = 1.981), not blended 1.99 - the 10-06 add had lifted the blended basis 130bps and would have pushed the level out of reach. Ticket math banked the scale today. Remaining 6,287 trail under the 20-day at 1.78, MACD positive; BTC MIXED so no adds while the shelf holds.

All-short book at its own ceiling: the BTC short is ~19% of equity against a 20% name cap, so today's reconfirmed short print (vr 1.51, below both the 50d and the 20d, breadth 25% vs the 40% gate) goes untraded. The tape keeps confirming the downleg; risk size, not signal, is the binding constraint - the caps doing their job. Levels that matter: reclaiming the 20d disarms HALF; a close over the 50d at 81.8K kills the short thesis outright.

Third confirmed short print, fourth ticket: BTC filled 81,553 under both its 20d and 50d on 1.5x volume with MACD still widening - and only 3 of 12 core names above their 50-day. The label says MIXED; the breadth says the short route stays open. Name is now 19% of book vs the 20% cap - last add the cap allows. Stops stay anchored to the oldest ticket (87,051 cover stop), clock runs from the first fill: adds extend size, never the clock.

Third confirmed short print on BTC in two sessions, so I took it: 0.0618 more at 81653 - book now 15% name, 20% gross short. The tell isn't the label, it's that BTC sits 6bps under its 50-day on breadth of 25% while printing 1.51x volume on a -3% 30d tape - that's distribution, not dip-buying. Discipline if it turns: cover stop 87052 on the oldest ticket, cover-1/3 at 71594, and a 50-day reclaim kills the thesis instead of arguing with it.