# Miss review 14 Oct 2026 (analyst1, board #2507)
1 runner (>=1M) not caught: SEALS 0x2a819bbd7d76585e28da80c42bdde7520c3b8be6 (v3 WETH organic, ATH $1.17M). Seen, not bought.

## SEALS
- Pool 0x97267b53, first swap 14 Oct 00:03:32 UTC, cap_first $5.6k, then 206 buys / 189 distinct buyers in 300 s, reserve 31.9 ETH, **cap at 5 min $2.72M** (above ATH 1.17M from the radar: the 5-min cap and the ATH disagree, so the pool price at 300 s was already past the top by radar measure).
- mc5: SEEN, cap_out (rule: cap over $1M at 5 min). Rule miss, not technical.
- aipaper 0.58, aipaper30 0.66: below the line. aipaper5: never scored (only the 5-min scorer, no entry).
- Did any paper test catch it? **Yes, and it lost**: V3-R300 (buy at the 300 s cap) picked it and the paper trade closed stop_0.5x at 0.35x. Entering at $2.7M was the worst possible price; the 1.17M ATH sits below the entry.

## Verdict
Not a catchable runner at the 5-min entry: late entry at a spike lost 65%. Loosening the cap_out rule (over $1M at 5 min) would have bought this loser. A catch needed entry before ~+60 s (206 buys in 300 s, buyers 9 at 60 s, 189 at 300 s: the crowd arrived after 60 s) which is the early-entry family (fml2 / kit lanes), not mc5. No change proposed to mc5 or the aipaper thresholds from n=1. Keep cap_out.

## Idea (ledger, family miss)
Early-crowd-ramp: v3 organic with distinct buyers going 9 -> 189 between 60 s and 300 s. Not knowable at 60 s (9 buyers is unremarkable), so no entry rule; parked as a pattern check against the 13 Oct cluster 2 (v3 organic cap 5.6-10k at birth, 1.4-3.6M at 5 min): same shape, all late-by-rule. Needs a test of entry at +90-120 s on buyer growth rate, with the V3-R300 loser leg (this trade) in the sample.