# Analyst1 idle run 46 (10 Oct 2026, scripts work/analyst1/p1 g1-g3 + inline grid)

Focus: v4 no-hook WETH/native (biggest unwatched runner cell). Data: swapamt.db (first 300 s, side + quote_amt), r6weth/r6nat tapes with pull events, canon.

1. NEW LENS arrival rhythm / size dispersion (bot cadence, metronome buys). Gap CV flat; buy-size CV: low dispersion 0.5-2% P1M vs high 2-4% but it is a whale/popularity proxy (inside active pools lift is 2x but only 1.9->6.5% weth, native non-monotone). DISCARD (info).
2. INVERSION lens, NEW: share of SELLS among swaps in 200-300 s ("late selling"). Active nohook pools (n>=25 swaps, still active after 150 s, cap_first<150k), P1M: weth sell share >0.448 -> 1/334 (0.3%) vs 38/559 (6.8%) rest, both date halves (1/261 vs 16/350; 0/73 vs 22/209). Native weaker (5/100 vs 6.2-7.4%).
3. Pull-aware sim (r6 tape, entry +315 s, lag 15 s, 2% pos size, buy impact + sell impact (R/(R+V))^2, 4% cost, tp3 stop .5, pull = exit at .1):
   - all active pools n1117 mean 1.01 (pull 58%, tp 32%); sell share <=.448 n748 mean **1.15** (pull 53%, tp 38%); sell share >.448 n369 .75.
   - dumb baseline: all <150k pools alive at 315 s n1324 = 1.00. Shuffle null (748 of 1117): p99.9 = 1.093, observed 1.15.
   - halves 1.04 / 1.22; ex top-5 1.13, ex top-20 1.09; cutoff .35-.55 all 1.13-1.15; pullexit 0 -> 1.10; lag 60 -> 1.10, lag 120 -> 1.08; tp 2 1.09, tp5 1.06; stop .7 1.13.
   - regime: trailing-60min birth count terciles 1.21/1.19/1.05 (busy chain worse); entry-cap terciles 1.08/1.21/1.15.
   - 19 days, most weeks > 1; first 2 days (21-22 Sep) .54/.69.
   Verdict: MATURE. Edge is thin (+10-15% a trade, win 40%, 53% of trades end in a pull) and sits in the BUSY60/FAST5 kit-template family at a LATE entry (+5 min), a window no current test covers. Not ready: no forward tape for nohook pools (v4tape = hooked only).
4. Request ppdata: daily refresh of r6weth/r6nat (and swapamt.db) to now, so the rule has forward OOS (see board post).
Not run (no fresh data): regime vs BTC, name/meta, unknown venue (all 23 = one-day batch, #423).

## Entry/exit rule for testmgr (arm candidate, paper)
Universe: v4 hooks=0, WETH or native quote, any fee, cap_first<150k. At birth+300 s: require >=25 swaps in first 300 s, >25% of them after 150 s, and sells/(swaps) in 200-300 s <= 0.448. Enter market buy 0.02 ETH-equivalent at +315 s (lag<=15 s), reserve>=1 ETH. Exit tp 3x (sell on first print >=3x), stop 0.5x, pull (liq removal>=50% of peak) = exit at whatever is left, max hold 4 h. Kill: mean <1.05 after 60 forward entries. Expect ~40% tp, ~50% pull.
## Data request (ppdata #1293)
r6weth.db, r6nat.db (swap + ModifyLiquidity events, hooks=0 v4) and swapamt.db refreshed daily to now (format unchanged), plus an hourly-updating forward tape for hooks=0 pools like v4tape (pool, t, side, quote_amt, reserve). Test: replay this rule forward on 1 Oct+ births as OOS.