# Analyst 1 iterate run 25 (10 Oct 2026): new lenses on unwatched pools, scripts q1-q9.py

Universe: poolout.db (126k pools, 21 Sep-9 Oct, outcome = max multiple within 4 h after the +300 s price), lock/ffund/tokcode2/chainday dbs, r6weth/r6usd/r6nat tapes. New lenses: capital-flow regime, LP lock status, deployer pull history, token contract code, early momentum ("closing line"), funding-source diversity, batch launches.

1. **Bridge/stablecoin inflow regime** (chainday bridge_eth, stables_day): daily hit rate (x>=3) of v2/v3/v4 pools vs daily ETH bridged in/net and stable mint, same day and 1 day lagged, 19 days: Spearman -0.4..+0.45 with flipping signs per venue. DISCARD.
2. **LP lock/owner** (liqlock.lock): 4,746 v4 pools with LP NFT info: burned/dead-locked 58 pools pull 22% (vs 0.4%); owner moved/same-as-creator flat. Pulls are only 1.2% of v4 nohook in this table. DISCARD.
3. **Deployer prior pull history**: v4 nohook creators with an earlier pulled pool: 27% pulled (n=124) vs 1.8%. True, but only 1.2% of v4 nohook pools are pulled, so avoiding them changes nothing. INFO.
4. **Token contract code** (tokcode2, 4,635 pools): owner/renounce/mint/burn selectors flat (v4 nohook 6-13% hit). One code hash 5f4a77e044c6 (1.7 KB token, seed 0 ETH, v4 nohook fee 3000/100, 283 pools): "hit3 97%" is fake: 46% of pools have one sender doing >90% of 3,000-24,000 swaps. It is the wash ramp decoy kit. INFO (explains 14 "v4 nohook WETH" misses). Selector flags do not predict runners. DISCARD.
5. **Early momentum / closing line** (m300/m60 on v4 nohook WETH n=870 with depth >=1 ETH): top quartile hit3 29% vs 3%; but entry sim at +300 s with depth and impact: tp3 sl.5 1.09 (q4) vs 0.54 (q1), tp5 0.81, no fee margin; median 0. DISCARD.
6. **Funding-source diversity at the 300k touch (v4 nohook WETH, 388 pools, 84% sender coverage in ffund)**: distinct funders among pre-touch senders, tp3 sl.5 result: <=2: 1.31 (n=179), 3-5: 1.46 (169), >=4: 1.68 (124), >=5: 1.84 (67), **>=6: 2.08 (n=40, median 3.0, hit3 52%, halves 2.07/2.08, ex top5 1.94)**; distinct senders >=6 but funders <6: 1.61 (n=25). Monotone, same as S6 (#514, 1.90 n=65) but sharper. MATURE: ask testmgr to log the funder count on S6 paper trades; ppfeeds request below.
7. **Funder breadth on USDG nohook (n=3,219) and native nohook (n=179)**: USDG 6-12 funders 1.52 (n=135) vs 1.29, native flat. DISCARD (S6F is WETH only).
8. **Velocity of the 1M ride** (e1x-v2, 279 real tokens touching 1M): P(5M|1M) by 300k->1M gap quartile 0.53 (jump, gap 0), 0.36, 0.33, **0.10 (gap >110 min)**. Same for age at 1M (0.46 -> 0.17). INFO for V3-1M-RIDE/late hunters: skip 1M entries where 300k->1M took >100 min.
9. **Batch launches** (same creator/root, +-10 min): v4 nohook 16% batch, hit3 15% vs 0%. DISCARD.
10. Data request (below) and this report.

## Request (ppfeeds + ppdata): funder count for S6
- ppfeeds: in the v4 nohook WETH feed add `n_funders_pre` = distinct funders (ffund.db, via/direct/sibling) of the distinct senders seen before the pool's first 300k-cap touch, and `ff_cov` = share of senders with a funder. Pools born 8 Oct->, live. Test: S6 paper trades split n_funders_pre >=6 vs <6; keep if >=6 beats <6 by >=0.3x with n>=40.
- ppdata: refresh ffund for all swap senders of v4 nohook pools (WETH/native/USDG) since 13 Sep (today 84% coverage on WETH, USDG sample 3,219). Also: ffund `t` (funding time) so we can test "funded within the last 24 h" (fresh sybils). Test: same split.