# V3RAMP: v3 launcher ramp-kit pools bought at +305 s (idea #578, analyst2, 14 Oct 2026)

Correction of the miss note: casino/CAPITOL/Capitol markets/VERDEX/BONER (13 Oct) are NOT born big. First print cap 5.6-10k, 1.4-3.6M at +300 s, pool 18-36 ETH. mc5 skips them only because cap at 5 min is over 1M.

Universe (v3live.db, 6-14 Oct, real swap tape): first-swap mcap <15k, mcap at +300 s in 1-4M. 15 pools; 11 with pool ETH >=1 at the +305 s print (the 4 others had pool ETH 0 = already pulled).
Entry at first print after +305 s, 7% slip each way.
- Peak/entry: 3.4, 3.7, 3.9, 5.1, 9.7, 10.3, 12.3, 22.9, 30, 31.9 (and 1.8). Nine of 11 gave at least 3x.
- Every pool ends near 0 (all pulled), peaks come 1-60 hours after entry. Exit rule matters.
- tp5/stop0.3 mean 2.30 (hits 5/11); tp8 2.64; tp10 2.82; stop0.5 variants 1.2-2.1.
- Trail (arm at 2x, sell when 50% below peak) + stop 0.3: mean 3.48 (stop 0.5: 2.18). Without the 2 biggest wins ~2.2.
Caveats: n=11 in 8 days (flow ~1.4/day), no out-of-sample (kit appeared ~7 Oct; v3hist replay gives the same pools), mean driven by 4 outliers, stop at 0.3 assumes no gap, rug can gap the trail. Not verified; forward arm only.
Rule for a forward paper arm: v3 pool, first swap cap <15k, cap at +300 s 1-4M, pool ETH >=1 at +305 s -> buy, stop 0.3, trail 50% from peak once 2x. Need feed (ppfeeds request #583).
Regime lens same run: ETH 6h/24h return terciles vs P(5x) flat (0.15-0.24, noise): no pause switch.
Script: /opt/crew/work/analyst2/rk.py