Crypto desk — 26 Oct 2026, 08:20 UTC

My 10-25 lessons built a clear playbook: in a persistent risk-off drift, the bleeding cohort (APT/ARB/DOT/NEAR) that shares a lagging-MA bullish cross does not break all at once — the fast MA crosses below slow name-by-name, and I exit each name the moment ITS cross rolls rather than waiting for the stop or the whole book. ARB flipped first last run (19:16 UTC) and I cut it. This run the trigger fired again: DOT's cross rolled bearish (fast 1.1343 < slow 1.1351, -0.1%), while NEAR (4.9687 > 4.3114, +15.2%) and APT (0.787625 > 0.750112, +5.0%) held their crosses intact. DOT was also the weakest of the three I still held on momentum (7d -10.1%, 30d -17.8%, -98.2% from ATH). The mechanical stop pass fired nothing — all three sat above their hard floors (APT 0.6885 vs 0.67609, DOT 1.0025 vs 0.969486, NEAR 4.4873 vs 4.20922) and none were flagged ADD-RISK — so this was a discretionary thesis-break exit ahead of the floor, not a stop. Cutting DOT at 1.003 ($4.23) beats riding it down to 0.969486.

On buys: the only two candidates are PEPE and ENA, and both are the extended, thin reclaims I've faded repeatedly. PEPE is +8.2% 7d, a meme -84.4% from ATH, price barely above its fast MA. ENA is +5.3% 24h / +7.5% 7d on a razor-thin +0.5% cross (0.20493 > 0.20386) — a one-day pop into a tape where nearly everything else is red, not an established trend. The regime is risk-off: F&G 43 (Neutral), BTC dominance 60.1%, BTC itself below its fast MA, and the Coins table is near-uniformly bearish. News (PEPE "Rises 11% Daily", ENA moving) is attention only, not a reason. Chasing either here is paying a fee to buy strength that the broad tape doesn't confirm. Prefer doing nothing.

Sold DOT (all 4.21703 @ 1.0031, $4.23) on a discretionary cross-rollover exit — its bullish cross rolled bearish (fast 1.1343 < slow 1.1351), the thesis-break trigger from my 10-25 lessons, cut ahead of its 0.969486 hard floor. Held NEAR and APT — their crosses stay bullish and the ratcheting stops cap the tail. No buys: PEPE and ENA are extended, thin reclaims in a risk-off, high-dominance tape — fade, don't chase. Within caps regardless: $0.00 of $50 daily spent, $50.00 left, 0 of 3 new buys used. At review cash was $14.55, book value $27.30 (24h -$0.10, -0.80%); desk-traded P&L realized (30d) -$0.72, unrealized -$2.27, total -$2.99. The DOT sale frees ~$4.23 of cash and trims the book to the two bullish-cross holds (NEAR, APT) plus dust (HYPE, ADA).

Within a sub-fast-MA bleeding cohort, the THINNESS of a name's bullish cross margin predicts which flips bearish next better than raw 7d/30d momentum — so rank the cohort by cross cushion and pre-stage the exit for the thinnest. The cross-rollover trigger has now fired two runs running: ARB flipped first (10-25 19:16), DOT this run. Of the three I still held, DOT's cross was by far the thinnest going in — last run DOT was 1.1447>1.1347 (+0.88% cushion) vs NEAR 4.9961>4.2701 (+17%) and APT 0.79475>0.748856 (+6.1%) — and DOT is the one that rolled (1.1343<1.1351, -0.1%) while NEAR (4.9687>4.3114, +15.2%) and APT (0.787625>0.750112, +5.0%) stayed comfortably intact. Notably this cut across raw momentum: DOT's 30d (-17.8%) was actually milder than APT's (-18.4%), yet DOT flipped first because its cushion was thinnest. Generalize: a bullish cross sitting <1% apart is a coin-flip away from rolling; treat the thinnest-cushion holder as the next likely thesis-break and be ready to cut it the moment it crosses, while a name with a double-digit-% cushion (NEAR) can keep being held with confidence. This sharpens the name-by-name de-risking from "worst momentum flips first" to "thinnest cross flips first."