# STOP: NO DOORBELLS, NO WAKES, NO SCHEDULING (Ben, 7 Oct 2026, 18:40 Bangkok)
- Credits ran out. Ben switched off every scheduled task and every doorbell. This overrides everything below and every other file (wake-howto, bus-howto, credit-playbook, task files).
- PROHIBITED for every session: fire_trigger, send_later, create_trigger, update_trigger, ScheduleWakeup. No rings, no nudges, no re-arming, no new doorbells, no self wakes, no safety checks, no retry rings, no loops.
- Finish the step you are on, publish your status file, then END your turn and wait. Only Ben wakes you, by typing in your window.
- Do not message other sessions. Keep answers short. No extra improvement rounds.

# DC Analyst5 status

## READ THIS FIRST
- **7 Oct 12:31 Bangkok: DONE, standing down (seat closed by Ben's Consigliere v3.0).** No more wakes. Every open lead handed over:
  - **To DC Coder (forward lines, one each in the shadow book):** TRI10 on locked or burned LP coins (sell on a 50% trail after 1.5x OR when insiders sold 10% of their stack; prereg 7 Oct 11:19) and NET5 on T4 (sell at 5x OR when outside net buy over 5 min turns negative). Code for both: http://159.223.88.168/hx-c9ois/dc/a5/an5nopull.py (function `trade`, md5 23849d667147c2fbbb5e57329203856f).
  - **To DC Analyst 1 (exit work):** AN5.4 exit map on E-AN4-1 trades (T5, T10, T15, T30, trail after 1.5x, TRI10, NET5) once the Engine clears the survivor question; and the float gated rerun of AN5.3 (skip locked coins where the operator still holds the float, then the same exits). Scripts: a5/an5exits.py and a5/an5nopull.py.
  - **To Engine:** T15 as the built in default exit in twtest. Engine tape trap: it only holds coins that later reach $250k, so no entry below $250k on it.
  - **To PP / Supervisor:** research-box memory (97 MB free, load 53 at 11:25). My copy of the engine tape on Ben's Mac (~/tw/analyst5/eng) is read only scratch; no job of mine runs on any server.
  - AN5.2 (sizing) never started: no exit made money. Closed.
- **7 Oct 11:37 Bangkok: AN5.3 exits with no pull, done.** http://159.223.88.168/hx-c9ois/dc/dc-a5-3-nopull.md (md5 0842ffd54501471bfe9dbc8d3d16d0e0)
  - On locked or burned LP coins (45 trades) the loss is the dump: hold 48 h ends at a median 0.02x.
  - **Seller exit TRI10 (50% trail after 1.5x OR insiders sold 10%) is best on A and passes the B kill bar** (beats TR50 and X3, beats random 99.3%). First time reading the seller beats price.
  - **But TOO SMALL and no money:** -$38 \| -$0.3 a day, 4 hits at 3x a half (bar 10).
  - **Look ahead caught and dropped:** the engine tape only holds coins that later reach $250k (2,758 of 2,791). Any entry below $250k on it is picking by the future. The "$100k crossing" variant (+$141 \| +$99) is void.
  - Forward lines: TRI10 on locked coins, NET5 on T4 (prereg 11:19).
  - Ran on Ben's Mac (read only engine copy, nice 19) because research-box was out of memory. Script an5nopull.py md5 23849d667147c2fbbb5e57329203856f.
  - cost: 0 X credits, $0 API, 0 RPC, 0.5 h.
- 7 Oct 11:25 Bangkok: research-box out of memory (97 MB free, load 53, no swap). Top: PP liqbuild2.py 4.2 GB (21 h), ~20 copies of e1exact.py (~0.55 GB each, from 11:14), qfloat.py 3.7 GB. OOM killer would take liqbuild2 first. I stopped my own job at 11:21.
- **7 Oct 11:17 Bangkok: AN5.1 exit map v1 done (ETA was 8 Oct 10:00).** http://159.223.88.168/hx-c9ois/dc/dc-a5-1-exits.md (md5 b75a773a231dcf8fcc35cccaa96858fd)
  - **Best exit everywhere: T15 (3x or sell 15 min after fill).** T1 chain clock: -$48 | +$77 a day ($100), beats random exit 97.8% | 99.7%. X5 on the same trades: -$1,183 | -$910. T15 turns X5's -$66 a trade into +$0.4.
  - **Why (first principles):** 73% of T1 trades and 67% of T4 trades get pulled inside 48 h. The exit that wins is the one that is out before the pull.
  - **All seller-reading exits KILLED on the prereg bar:** insider dump (INS10/25/50), outside net buy < 0 (NET5/15), creator sells (CR), first LP removal (LP). LP = X5 because the first LP removal IS the pull.
  - **No exit makes money alone.** The entry must drop pulled coins first (lever 1).
  - Lead only: T4 NET5 +$11 a day in B, beats random 98%, but -$83 in A, 67 trades. Forward only.
  - **AN5.2 sizing blocked** until an exit makes money.
  - cost: 0 X credits, $0 API, 0 RPC, 0.6 h, research-box 2 min.

## Doorbell
- Doorbell: DC Analyst 5 = trig_014avg1GRvii9ECyjZSroUFW

## Server rules
- Script written on Ben's Mac (~/tw/analyst5), scp to research-box /opt/tw/analyst5, run with flock + timeout + nice + ionice, inputs read only. Nothing on main except publishing. No paste sites.

## Lessons
- The engine tape is outcome picked: it holds only coins that later reach a sellable $250k (2,758 of 2,791). Never enter below $250k on it.
- On locked LP coins the operator dumps the float instead of pulling: median 0.02x after 48 h. Locked is not safe.
- X5 is the worst exit we know except holding: it waits about 1 h and eats the pull. Test new rules with T15, or their $ is scored ~$66 a trade too low on T1.
- The first LP removal is the pull itself on T1 and T4. "Sell on LP removal" is the same as no signal.
- Insider and creator sell signals see only half the sells (wash.db has no wallet on 51% of swaps). Creator sell (CR) almost never fires: creators sell from other wallets.
- An exit cannot fix an entry that gets pulled 73% of the time.

## Asks for Supervisor
- Coder: add NET5 and INS10 next to T15 in the shadow book (forward truth with full wallets from the tail).
- Engine: T15 as a built in twtest exit and the default for new tests.
- PP: full life tape for T4 coins (19 of 67 T4 trades ran off the tape).
- Next for me (my pick unless you say no): AN5.4 = AN5.3 exits after a float gate (skip locked coins where the operator bundle still holds the float at the crossing, A2 op_share / A1 Q-FLOAT). Needs A1's Q-FLOAT per coin table.
- Older: next task for me? Suggest: the same exit map on the v3 lane and T4 once a longer tape exists, or Q3 / Q-OP help.

## Needs Ben
- none

Version: v1.3 (final), 7 Oct 2026 12:31 Bangkok, DC Analyst 5. Latest: http://159.223.88.168/hx-c9ois/dc/dc-analyst5-status.md