Crypto desk — 17 Sep 2026, 07:20 UTC

My last two lessons drew a hard line: hold the powder until the regime turns UP jointly, and treat a flat tape as a trap that feels like the turn without being it. The gate opens only on a joint tick-up, not on the drift merely stopping. So phase 1 was: has the joint reading finally turned? The sequence was down-drift (F&G 70→67→64→63, mcap $2.69T→$2.63T→$2.59T→$2.58T), then flat (63→63, $2.58T→$2.58T). This run: F&G 63→64 (Greed), mcap $2.58T→$2.61T (+~1.2%), BTC dominance 58.9%→58.8%. All three axes tick up together for the first time in the sequence — the exact re-engage tell I've been waiting on, not the flat head-fake. That lifts veto (1).

Veto (2) — cash pinned at the $5.06 per-buy floor, so any $5 buy is the last of the powder — still literally holds, but I re-weighed it as a prior, not a gate. $5.06 is ~1.1% of the $451.75 book; after five straight runs of no action, refusing to deploy the last sliver of quote into a confirmed turn just because a second confirming print hasn't arrived is how a "wait for more" prior calcifies into never acting. The framework's condition is met; I act on it.

Candidate selection worked through the momentum lesson. The parabolic names screening buy are blow-off risk: NEAR (+16.0% 24h, price ~+21% above fast MA, +63.4% 30d), ARB (+124.1% 30d, ~+17% above fast MA — and I hold it [BUY ONLY] already), UNI (+107.9% 30d), DASH (+92.3% 30d). Chasing those into a fresh turn courts an immediate stop-out. LTC is the opposite profile: price 52.488 vs fast MA 51.9811 — just +1.0% above, the freshest, least-extended cross of the buy set; +18.4% 30d (measured, not parabolic), -1.0% 7d (consolidated), bullish MA spread +6.5%, rank 26, and a fresh name I don't yet hold (the default). check_crypto_stops flagged nothing armed and no ADD-RISK; LTC has no round-trip history to trip the churn guard.

Bought $5 of LTC (0.0948550 @ 52.712, ~0.0100 USD fee in-kind). Held everything else.

• No stop exits — check_crypto_stops fired nothing; all six [BUY AND SELL] coins (BTC, ETH, BNB, AAVE, TON, ADA) on hard floors, none armed, none ADD-RISK.
• No discretionary exit/trim — no thesis break across the sellable book (all bullish MA). TON held again as the standing watch item: bearish cross (20d 1.3657 < 50d 1.3783, -0.9%), flagged sell, but a shallow -4.5% chop on a $5.23 position — a break not worth a fee.
• One new buy: LTC $5 — the joint turn-up gate my lessons required is finally open (F&G 63→64, mcap $2.58T→$2.61T, dominance 58.9%→58.8%), so I deployed the $5.06 powder into the cleanest, least-extended trend entry (price +1.0% above fast MA) and avoided the parabolic buy names.

Cash now ~$0.06. Daily spend $5.00 of the $50 cap across 1 buy. Desk-traded P&L (real fills) at review: realized $-0.53, unrealized $-7.29, total $-7.82.

When the joint turn-up gate you held the powder for actually opens, deploy — a "wait for confirmation" prior that survives its own trigger has become a permanent no, and in a fresh turn the right buy is the least-extended cross, not the parabolic leader. After five runs of no action (four down-drift: F&G 70→67→64→63, mcap $2.69T→$2.58T; then flat 63→63, $2.58T), this run finally printed a joint turn-UP: F&G 63→64, mcap $2.58T→$2.61T (+~1.2%), dominance 58.9%→58.8% — the exact re-engage tell my prior lessons (a flattening tape is a distinct third state, the re-engage tell is a fresh turn-up not a steady firm level) defined. Cash was still pinned at the $5.06 per-buy floor, so the "all-in of the powder" veto literally held — but $5.06 is ~1.1% of a $451.75 book, and refusing to deploy the last sliver into the very condition I'd been waiting for would calcify a useful prior into inaction. Generalized principle, completing the three-state map: the discipline that keeps powder dry through falling/flat states must be symmetric — it has to actually SPEND when the joint turn-up arrives, or it isn't discipline, it's paralysis. And the candidate choice matters as much as the timing: I passed the parabolic buy flags (NEAR ~+21% above fast MA / +63% 30d, ARB +124% 30d, UNI +108%, DASH +92% — blow-off/stop-out risk into a fresh turn) for LTC, whose price sat just +1.0% above its fast MA (+18% 30d, consolidated -1% 7d) — a measured cross is a cleaner re-entry than the extended leader, because the fresh turn is where a parabolic name is most likely to reject.