Crypto desk — 19 Sep 2026, 07:16 UTC

Phase 1 first: cash $12.26, book value $529.28, 24h +$22.02 (+4.45%), no pending orders, $0.00 of the $50 daily cap spent. Desk-traded P&L: realized $-0.53, unrealized $+16.51, total $+15.98. My two recent lessons both concern the joint turn-up gate — filter for a GENUINE trend (coherent positive 7d/30d), THEN take the least-extended cross within that set, and let BTC-dominance direction tell me whether the cleanest low-extension trend sits in BTC/majors (rising dominance) or alts (falling).

Regime this run: F&G 73 (Greed, holding after last runs' 66→73), mcap $2.78T (up from $2.77T), dominance 58.7% (ticked DOWN from 58.8% at the 19:16 run) — a mild breadth tilt back toward alts, so the dominance lesson points me slightly away from forcing BTC and toward a clean alt/coin cross.

check_crypto_stops fired nothing — all nine [BUY AND SELL] coins on hard floors, none armed, none ADD-RISK. Discretionary pass across the sellable book found no thesis break: everything bullish-MA except TON (fast 1.3684 < slow 1.377, now -1.3%, 7d -0.2%, 30d +1.1%) — shallow chop, still well above its $1.13626 hard stop and my standing watch item, not a break. Held it again, no trims.

New buys: cash $12.26 caps me at two $5 buys. Ranking extension above fast MA among genuine trends: DOGE +2.6%, BCH +2.95%, HBAR +2.96%, XRP +3.44%, BTC +3.66%. Rejected the parabolas per the momentum lesson (NEAR 30d +115.8%, ARB +147.7%, UNI +153.1%, STX +110.3%, DASH +90.5%, ENA +93.4%) and the flat non-trenders per the least-extended lesson (TRX 7d -1.7%/30d +1.4%, TON, QNT below fast). DOGE was the literal least-extended, but it's a meme in an elevated-greed tape (F&G 73) where blow-off risk is highest — with extensions this tightly clustered I tilted the screen toward quality coins over the meme rather than chase the smallest number.

Bought $5 BCH (0.0201208 @ 248.499) and $5 XRP (3.50680 @ 1.4258), ~0.0100 USD fee in-kind each. Held everything else — no stop exits, no discretionary exit or trim.

• BCH — the cleanest low-extension genuine coin cross (+2.95% above fast, 7d +7.7% / 30d +17.1%, bullish MA +3.0%). An add to yesterday's flat position (bought 247.759, now ~248.08, on hard floor, not ADD-RISK), so no chasing.
• XRP — +3.44% above fast, rank 5, 30d +28.8% / 7d +4.2% steady large-cap trend; the mild dominance easing favored this alt over adding BTC (+3.66%, higher extension). An add to the core [BUY ONLY] holding.

Cash now ~$2.26. Daily spend $10.00 of the $50 cap across 2 buys. Desk-traded P&L at review: realized $-0.53, unrealized $+16.51, total $+15.98.

The least-extended-cross screen has a coin-type tiebreaker in elevated greed: when a meme and a quality coin are neck-and-neck on extension, tilt to the coin, because meme blow-off risk climbs with the Fear & Greed reading. This run the gate was open (F&G 73 Greed, mcap $2.78T up, dominance 58.7% easing from 58.8% = mild alt tilt) and my genuine low-extension trends clustered tightly: DOGE +2.6% above fast, BCH +2.95%, HBAR +2.96%, XRP +3.44%. Per least-extended cross means the cleanest cross of a genuine trend DOGE was the literal least-extended and a real (not parabolic) trend — 7d +3.2% / 30d +16.7%. But F&G was 73, and a meme's downside is fatter in greed (that's where blow-offs happen). With the extension gap under a point, the smaller number wasn't worth the meme tail, so I passed DOGE and took the two cleanest COIN crosses — BCH (+2.95%, 7d +7.7% / 30d +17.1%) and XRP (+3.44%, 30d +28.8%), the alt tilt matching the easing dominance (when the joint turn-up gate opens but BTC dominance is RISING tilt toward BTC-quality names read in reverse). The refinement: extension is the primary sort, but coin-vs-meme is a live tiebreaker when they're close — and greed level sets how hard that tiebreaker leans away from memes.