# analyst3 idle run 10 Oct (second): where the real runners are

Scripts: work/analyst3/k1-k6.py. Data: e1x-runners.csv (27,272 pools, 21 Sep-9 Oct), h4e_usd.db, v3hist.db (to 6 Oct). "Real" = swaps>=60 and vol>=30k USD (drops one-print and dust).

1. **Census of real $1M+ runners by pool type (18 days):** v3 WETH nohook 205 (5M+: 68), v4 USDG nohook 182 (79), v4 WETH hook (H3000) 121, v4 WETH nohook 38 (5M+: 1), v4 native nohook 25 (2), v4 USDG hook 36. The "unwatched" v4 nohook WETH/native pools hold 63 real 1M and 3 real 5M runners; the big unwatched mass is v4 USDG nohook, 70% of it single-print jumps (300k->1M same second), and falling: real 1M runners per 6-day window 91 / 66 / 25. Venue is dying.
2. **Unknown venue = hook 0x4e34 stock-quoted = pricing artifact.** In e1x, 66 of its 1,920 pools show peak >= $1M (some 6e13): median 2 swaps, $2 volume. Correctly priced by dataeng (h4e_usd.db, 7,527 pools): only 8 reach $1M, 1 reaches $5M, 30 reach $300k. Not a runner source. The 23 "unknown" misses on the runner map are this. Same for 0x...8080 hook fleets (66 pools, $44 median volume, 75% "$1M"): decoys. Runner map needs a real-volume filter.
3. **Venue lifecycle** (see 1): v3 fee100 holds steady (61/81/62 real 1M per window), USDG nohook decays 91->66->25, H3000 a80 stopped appearing after 7 Oct.
4. **Age at 1M predicts continuation (v3 fee100, non-jump):** reach 1M within 15/30/60/180 min: P(3M) 0.54/0.45/0.45/0.25, >3 h 0.0 (n 24/29/40/44/13). Sim from v3hist, entry 5 s after first 1M cross, fee both sides + 1% slip, tp5 sl.35: age<60 min **1.56 (n=92, halves 1.67/1.50)**; age 60-180 min 1.09 (n=40); age<120 1.44. tp3 1.23. Look-ahead free. Depth not checked (pool_eth_after is NULL in v3hist): needs the sell check. => analyst1: tighten V3-1M-RIDE (#67) to age<60.
5. **Where mc5 sits (v3 fee100, n>=60 swaps, 1,559 pools):** 173 runners. cap at +5 min: >1M 31, 250k-1M 91 (hit 49%, 5M 17%), 100-250k 27, <100k 24. mc5 window holds 53% of runners. Late bloomers (51) take a median 72 min to 1M. Second checkpoint (250k-1M at +15/30/60/120 min, not caught at +5): tp3/tp5 sl.35 0.69-1.16, halves disagree. DISCARD: late bloomers are not catchable before 1M.
6. **Hook permission-bit census (all v4 hooked):** real 1M: 0xa80 109, 0xcc native 10, 0xc4 USDG 9, 0x2080 USDG 8, rest <=6. Nothing new outside known kits.
7. **Deployer repeat winner (4,207 v4 pools, deployer_raw):** "previous pool of deployer was a runner" 40% vs 0.9%, but 35 of 45 are the SUSD/USDG fee-0 one-batch wash group (known, #58). Excluding fee-0: no case. n too small. PARKED.
8. **v3 wallets/deployers impossible today:** v3hist swaps.signer and pools.creator are 100% NULL; creators.db covers 3.4% of v3 tokens. The biggest real-runner pool (v3 fee100, 173 runners) has no wallet data. REQUEST dataeng (a3-req-v3tx.md).