Crypto desk — 2 Nov 2026, 08:16 UTC

Phase 1 carried two QNT lessons forward, and they converge on one operative line: the exit trigger is price-vs-fast-MA (plus cross width and the floor) — NOT 7d deceleration, which I've now watched burn from +13.6% → +9.7% → +1.7% → +2.0% without ever producing a roll-over. Both lessons also independently kill the QNT add case: a decelerating 7d re-flagging buy is the weakest possible entry, pure local-top chasing on an already-concentrated name.

Regime first (Phase 2): Fear & Greed 36 (Fear), BTC dominance 60.0%, mcap $2.62T — a risk-off tape, and the Binance-hack news flow (BTC ETF -$643M October outflows, ETH -16% in October) is pure attention, not a reason either way. The MA screen is near-unanimously bearish: of 50 coins, only three escape sell — NEAR (bullish cross but price 4.1137 below its 4.6928 fast MA, so a hold, not a fresh entry), PEPE (buy), and QNT (buy). That breadth alone argues for defense.

QNT (my one real holding at $4.71): check_crypto_stops shows price 197.47 vs the $171.692 hard floor (-5.7%, peak +0.2%) — not armed, not ADD-RISK. The structure is intact: price 197.3 above the 195.205 fast MA, wide bullish cross +6.2% (195.205 vs 183.8204), 7d back to +2.0%. By my own rule, nothing in that is a roll-over (price hasn't lost the fast MA, the cross hasn't narrowed), so there's no discretionary exit and no reason to churn a healthy hold. The QNT add I pass on principle: concentrated, local-high, exactly the weakest entry my lesson flagged.

PEPE is the only fresh buy: price 0.00000483 is ~+8.3% above the 0.00000446232 fast MA on a thin +4.4% cross, 7d +8.8%, 30d +13.4% but 24h -2.4% — an extended meme already rolling over intraday in a Fear/high-dominance tape. That's chasing a pump, not riding a trend; I've passed it on the same read the last two runs and the setup has only gotten more extended. Pass.

No trades. check_crypto_stops fired no exits (QNT price 197.47 vs the $171.692 hard floor, -5.7%, peak +0.2%, not ADD-RISK). Held QNT — structurally intact (price 197.3 above the 195.205 fast MA, +6.2% cross, 7d +2.0%); no fast-MA break = no roll-over = no sell, and no churn. Passed the QNT add — won't deepen a concentrated hold at a local high on a decelerated 7d. Passed PEPE — an extended meme (~+8.3% above its fast MA, thin +4.4% cross, 24h -2.4%) in a Fear(36)/dominance 60.0% risk-off tape; chasing, not a trend. HYPE/ADA are dust ($0.00), nothing to do. No new buys. Within caps: $0.00 of $50 daily spent, $50.00 left, 0 of 3 new buys used.

A lone buy flag against an almost fully bearish screen is a divergence to distrust, not an opportunity to seize. This run the MA screen was near-unanimous: 47 of 50 coins flagged sell, only QNT (a hold I already own) and PEPE surfaced buy, in a Fear(36)/BTC-dominance-60.0% risk-off tape. PEPE has now been the sole fresh buy three runs running, and across those runs the setup DECAYED rather than built — price sits ~+8.3% above its 0.00000446232 fast MA on a thin +4.4% cross, but 24h flipped to -2.4% even as 7d (+8.8%) / 30d (+13.4%) stayed green, i.e. the extension is cracking at the near edge. Generalize: when a single extended meme is the only green name in an otherwise red, risk-off breadth, that isolation is evidence the move is a late solo pump diverging from the tape, which typically resolves DOWN toward the tape — not the start of a trend. Breadth is a regime filter: one buy flag in fifty is a reason to sit out and protect capital, and the right default when a name keeps re-flagging buy while its 24h deteriorates is to keep passing, not to finally cave. (Next run: did PEPE lose its fast MA / roll over toward the red tape, confirming the lone-buy divergence resolved down — or did breadth broaden, more names turning buy, making a green flag mean something again?)