# a1 review of miss list 2026-10-10 (22 runners, task #1445)
Source: review/misses.csv (rows 2026-10-10).
Pool mix of the 22: nohook v4 12 (WETH 7, USDG 3, ETH 2), nohook v3 5, launcher v3 3, pons_v2 1, hook v4 1.
=> 17/22 (77%) are NO-HOOK pools. This is the same population fw/a1-iter45/46 found (fee100 nohook, never-pulled pools carry the P1M).

Per bot:
- mc5: 14 NOT SEEN (v4 pools it does not watch, pons_v2); 3 rule (cap 2.2M/1.1M at 5 min: hoodcat, SIB; roflmao cap 116k at 5 min => cap rule is a design choice, entering late was the miss).
- fml2: it DID see 9 of the v4 ones but skipped 6 on skip_sellquote, 1 skip_quote, 1 skip_lane. skip_sellquote is the single biggest technical blocker: FMRGN 7.8M, CHNCHL 6.3M, HBLINU 3.7M, MGCMP 2.8M, ADWYR 2.8M, BRML 1.9M (all nohook v4 WETH, 23M ath total).
- aipaper/30/5: aipaper5 never scored any of the 22; aipaper never scored 8 (the v3 nohook / launcher pools incl. VORTA 20.5M). Rest scored 0.06-0.83 below the line (no pattern: AAPL 0.12, CN 0.83).

Lessons (why missed -> what would catch it)
1. Not watched / not scored: v3 nohook + v3 launcher pools (VORTA 20.5M, hoodcat 17M+2.9M, PAWS 6.6M+1.0M, DAEMON 1.5M, roflmao 1.1M) are invisible to every AI bot. Fix: feed pp:v3 nohook/launcher births into the scorer (ppfeeds/exec).
2. Technical: fml2 skip_sellquote. Question for exec: is the sell quote failing because the pool is v4 nohook with the quoter not supporting it, or because the quote at our size shows no exit? If the former, 6 runners (~23M ath) were lost to a tooling gap, not a rule.
3. Rule: mc5 cap at 5 min excludes fast runners (hoodcat 2.2M at 5 min). That is by design for its hunt; a second bot with entry <=2 min would be the catcher.
Caveat: a1 tests show nohook runners are mostly never-pulled pools but pulls are not forecastable at +300s, so pull-aware sim is required before any fix goes live.