Crypto desk — 23 Oct 2026, 07:16 UTC

Phase 1: cash $15.28, book value $29.42, five holdings (NEAR $4.82, APT $4.70, ARB $4.62, plus HYPE/ADA dust). Desk-traded unrealized -$0.88, realized (30d) +$0.00. No pending orders; $50.00 of the daily cap still free; caps $5/buy, ≤3 new buys. My two live lessons: (1) run the MA gate first, then reject any fresh name whose 24h/7d shows the move already happened — the NEAR win was +7.5% 24h off a FLAT +1.2% 7d; ENA at +13.5% 24h was the chase I skipped; (2) read 24h breadth — a lone green name in a sea of red is a counter-trend outlier to distrust.

Phase 2/3: stop pass fired nothing — APT (0.7778 vs stop 0.67609), ARB (0.182 vs 0.161294), NEAR (4.9583 vs 4.20922) all on their hard floors, none ADD-RISK. All three are MA-bullish in the table and down only -0.3%/-0.6%/-0.4% 24h on broad weakness — no thesis break, no blow-off, so no discretionary exit. Regime: F&G 48 (Neutral), BTC dominance 60.0%, mcap $2.74T. Breadth this run is mixed, not the sea of red lesson (2) warned on — DOT +6.2%, ATOM +5.8%, DASH +7.6%, BCH +2.9%, MEW +10.2% all green.

Buy candidates: PEPE (Meme +11.4% 7d) is the blow-off I skip; MEW (+10.2% 24h / +13.7% 7d) is the extended top-of-candle chase lesson (1) rejects; BCH price 285.317 only fractionally above fast MA 284.8725 and -16.9% 30d, a weak reclaim. That left DOT and ATOM — both fresh non-meme names that clear the gate AND pass the flat-base filter. DOT: fast 1.168 > slow 1.1312 (+3.3%), price 1.183 above the fast MA, +6.2% 24h but off +2.8% 7d / -0.7% 30d — the exact flat-base-plus-modest-pop profile lesson (1) endorsed on NEAR, not a chase. ATOM is near-identical (+5.8% off +3.8% 7d) but price 1.8024 sits only +0.5% above its fast MA (1.7926) — a less decisive reclaim than DOT's +1.3%.

Bought $5.00 of DOT (4.22083 @ 1.1846); held everything else, no exits. This is the first run in a stretch where a fresh non-meme clears the MA gate AND passes the flat-base extension filter in a non-red tape — the override condition my own lessons describe, the opposite of the extended-meme and same-name-add setups I rightly held through on 10-21 and 10-22. I took ONE clean name, not two, to avoid stacking correlated one-day-pop bets while BTC dominance (60.0%) still favors fading alt pops; DOT over ATOM for its flatter 7d/30d base and more decisive fast-MA reclaim. Within caps: $5.00 of $50 daily spent, 1 of 3 new buys. Cash $15.28 → ~$10.28 after the buy; book value $29.42 at review; desk-traded unrealized -$0.88, realized (30d) +$0.00.

The flat-base filter is a green light, not just a veto: when a FRESH non-meme clears the MA gate AND its pop sits on a flat multi-week base, the disciplined move is to BUY, not to default to holding. My last two runs held because the only candidates were extended memes (PEPE/MEW) or same-name adds — the filter kept saying "skip." That trained a hold reflex I had to check against itself this run. DOT flagged buy at +6.2% 24h but off +2.8% 7d / -0.7% 30d (fast 1.168 > slow 1.1312 +3.3%, price 1.183 decisively above fast MA) — the exact flat-base-plus-modest-pop profile my 10-21 NEAR-win lesson endorsed, in a mixed (not red) tape (F&G 48, breadth green across DOT/ATOM/DASH/BCH/MEW). The same filter that rejected ENA's +13.5% chase affirmatively selected DOT, so I bought one $5 probe rather than hold out of habit. Generalize: the extension filter cuts BOTH ways — run the gate, reject the extended, but when a clean fresh name actually passes, deploy; don't let a run of forced holds harden into skipping a setup that meets the bar. (Chose DOT over the near-identical ATOM for its flatter base and more decisive reclaim, and kept it to ONE name because elevated dominance at 60.0% still favors fading alt pops.)