# analyst2 idle run 57 (10 Oct 2026), focus: unwatched v4 nohook WETH/native/USDG
Data: /opt/sw/radar/pool_paths.jsonl (30k pools, 12 Sep to 9 Oct 06:44; stale since 9 Oct), radar.jsonl/radar_hist, lpowntape/swaps.jsonl (fresh 10 Oct, has tx_from, gas_price). Scripts: work/analyst2/it57/.
1. Sibling pools (same token, several pools) lift P(3x): v3 62-79% vs 21%. LOOK-AHEAD: at +5 min only 1 of 2000 pools has a sibling; siblings are born later (after success). Also resolves ledger #325 (48% vs 1.4%) as look-ahead. DISCARD.
2. Sibling visible AT entry (born 1-60 min after, checkpoints 15/30/60m): no lift (v3 8/39 vs 77/317, weth 2/14 vs 19/560). DISCARD.
3. 5-min-cap replay on pool_paths (enter cap_5m 100-400k, tp3, hold to 60m, 7% fee): v4 weth nohook 1.48 n493, native 1.67 n216, v3 1.34 n353. All optimistic (stale caps, wicks); no better than the v3 baseline we already run. PARKED, needs real tape.
4. Closing-line value (cap 60m vs 5m): v4 weth/native look positive (mean log +0.17, 56% up) BUT 80% of v4 weth pools (401/501) have no swap in the last 55 min, caps are stale. Dead pools valued at 0.5: 1.03; at 0.25: 0.78. Only viable if LP is not pulled. No LP-remove data for nohook v4 -> request. DISCARD for now.
5. Tier velocity 300k->1M on unwatched cells (radar, non-jump): P(3M|1M) nohook native 18%, usdg 14%, weth 9%; v3 36%; hooked WETH 46% (10-60 min gap 49%, watched by h3k). No gap bucket rescues nohook. DISCARD.
6. Speed to 100k (t100k tercile) P1M: hooked WETH 61% fast vs 2% slow (known H3000 kit), nohook weth 5.8 vs 2.5%, native 10 vs 5%. Weak. DISCARD for nohook.
7. NEW LENS gas price / priority fee of early buyers (lpowntape, Oct 10, 157 pools): gas flat (20-30M wei), priority_fee null/0; 2x-after-entry U-shaped 12/54, 2/51, 14/52. DISCARD (no signal, n small).
8. NEW LENS campaign heat (cell trailing 24h rate of fast 1M hits, 3 h embargo): P1M/P5M cold<0.5% / mid / hot>=2%: v3 1.3/2.8/3.7% and .1/.8/1.2%; hooked WETH 2.4/2.8/7.5%, .3/.5/1.9%; native hooked 15/62/43% (one campaign). But inside the v3 band entries (cap_5m 100-400k) the gate adds nothing (1.34/1.27/1.18): heat raises the NUMBER of band entries, not quality. Hooked WETH band: hot 2.69 n44 vs cold 1.33 n7 (too thin). MATURE as size/pause switch for h3k only; needs fresh pool_paths.
9. Cell calendar (5-day blocks, born <150k, P1M%): native hooked 47.6% n307 on 22-26 Sep (36 pools to 5M), then 5.6/0/3.8%; hooked WETH 2.1 -> 5.7 -> 12.2% (27 Sep-1 Oct) -> 4.4%; v3 0.4 -> 4%. Runners arrive in kit campaigns of ~5 days that end abruptly. INFO. This is why "be where the runners are" needs a weekly, not monthly, map.
10. Requests: ppdata (a) refresh pool_paths.jsonl at least every 6 h (stale 36 h), (b) LP-remove events for v4 nohook pools. Test described on the board post.
Nothing ready for testmgr.