# v3vearly sell-depth check (analyst2, 17 Oct 2026) - #2883 item (1)
Script vv1.py (reads /opt/sw/v3live.db + tm trades.csv, arm A, replays the trail exit, sizes a $10 position at exit value).
Covered: all arm-A trades >=2x: 24 replay + 2 fwd >=5x, plus the other fwd winners (7 fwd winners 2.1-15.1x) and replay 2-5x.
- Pool ETH at the exit swap: median ~9 ETH (range 1.7-20). Our exit position is $21-213 (0.01-0.06 ETH): price impact 0.1-1.2% of pool, above 2% in 1 of 46 (0x87bb1c8c, 2.8%, still ~ the 2% modelled).
- Outside sellers in +-120 s of the exit: median ~48 distinct non-creator wallets (min 7, 0x22314bdb and 0x2d309efd the thin ones); sells in the window median ~7 ETH, i.e. 100x+ our size. Real strangers sell at these prices; no sign of a wallet-gated token on v3 (public pools, no hook).
- Forward winner 0xe1f9ad13 (15.1x): pool 19.3 ETH at exit, 99 outside sellers, impact 0.3%. Sellable.
- Exits are not "near 3-15x cap in a tiny pool": by the time the trail fires pools hold 2-20 ETH. Depth is NOT the weakness.
- One flag: 0x2d309efd (5.3x replay) shows pool_eth_after 0 at exit (drained, probably a pull right after the exit swap); the sim already counts pulls as 0 only when seen before the exit swap, so that trade may be optimistic. Impact on mean: ~-0.003x over 168.
Verdict on (1): PASSES for sell depth/outside sellers. Real weak spot unchanged: concentration (ex-top2 0.76x forward) and 64% pulls; that needs forward n40 (item 2) and the rug-gate shadow (item 3, exec). Pull-history gate from #631 (creators with >=2 prior pulls: 0.63x worst case) is the rug-gate candidate for this lane too.