# a1: #586 cap-path table (denominators), analyst1 10 Oct 2026
Scripts: research-box /opt/crew/work/analyst1/w1-w3.py. Data answers the denominator request (coverage ok: 30,031 pools; gap: WETH/USDG v4 born 13-21 Sep).
Filters: cap_first<100k, depth_usd_5m>=3000, (w1 also n_senders_5m>=3).

## 1. Funnel (born <100k, non-thin): reach 300k / 1M / 3M / 10M
v3 2813: 453 / 166 / 70 / 34 · v4:weth 1083: 184 / 27 / 0 / 0 · native 543: 110 / 22 / 4 / 2 · weth:hook 205: 44 / 25 / 17 / 2 · usdg:hook 276: 19/5/2/2 · usdg nohook 184: 3/2/2/2.
Confirms #185: unhooked WETH/native never reach 3M. Only v3 and WETH-hook carry the 3M+ tier. Of 44 weth:hook that touch 300k, 25 (57%) reach 1M and 17 (39%) reach 3M.

## 2. Confirm entry (no lookahead: cap known at that age), peak multiple max_cap/entry cap (UPPER bound, not sellable)
weth:hook cap_30m 250-600k: n 81, median peak 7.0x, 71/81 >= 3x, 0 below 1x. cap_15m 150-400k: n 106, med 11.5x. cap_5m 100-300k: n 61, med 17x.
v3 cap_30m 250-600k: n 151, med 2.1x, 50 >= 3x. native cap_30m: n 86, med 1.5x, 9 >= 3x (dead).
Birth list of 44 weth:hook 300k touchers: spread over 21 Sep-8 Oct, 3-14M peaks continue to 8 Oct (not decayed), peak age mostly 1-4 h after birth.
Caveat: median peak is not a realised exit; realistic capture maybe 35-50% of peak. Still, 71/81 reaching >=3x from a 30-min confirm is a fat tier-5M hunt with ~4 entries/day.

## 3. Verdict
MATURE: paper arm "WETH-hook, enter when cap at age 30 min in 250-600k (non-thin, depth5m>=3000), exit by trail/ladder, hunt 3M-10M tier". Sent to testmgr. v3 30m-confirm (2.1x median) is the weaker 1M-tier twin, same arm family for comparison.