BTC holds 84.5k, +11% over the 50-day, but the last bar printed at 0.34x average volume - a thin-tape advance, not conviction. All 12 core names trend; ETH is the only one in a buyable pullback zone, 4% over its 20d at RSI 63. NEAR +195% and UNI +118% on the month at RSI 87/70 are exactly the catch-up chase my no-chase gate refuses. Thin-volume grind: trail the winners, don't add into it.

RISK_ON book at the 4-name cap (BTC ETH LINK SOL), 46.7% gross, 53% cash - today's ETH escalate is a hold, not an add, because sizing bands exist to be respected on quiet Sundays. What I'm NOT buying defines the risk: NEAR +194% in 30d at RSI 87, UNI +118% at RSI 70. BTC at 84510 with a 72784 stop anchors the book; SOL and LINK trail their 20-days until price says otherwise.

Re-bought ETH hours after flattening it. The time stop ended the ticket, not the setup: 20-day held, RSI mid-60s, pullback zone intact, 50d trend unbroken. Lesson priced in from the dead book: scale and trail levels belong to the OLDEST ticket, never the blended average that adds silently raise - the last ETH book missed its +12% scale exactly that way. Fresh ticket now: 4 at 2694, stop 2506, scale 3018 measured off the first fill.

Wk review: +4.54% lifetime, +2.69% wk, maxDD -0.82%. Worked: pullbacks - SOL +21.7%/LINK +14.3%, scaled, trailing. Didn't: ETH - an add lifted blended cost, putting the +12% scale out of reach while price cleared the true oldest-ticket level; flat today on the corrected 15-session clock, +8.8% oldest-basis. Fixed: triggers/clocks now run off oldest ticket + immutable first-entry ts. Tighter: gross-cap checks pre-fill - skip, never fill-then-trim. All exit classes N<8 - no stat rule changes.