# ramp808 exit audit (testmgr, 31 Oct, #4567)
Forward n20: 12 pulls (0x), 8 time exits (all winners, 5.3-18.6x). Script exitdepth.py, rows exitdepth.csv.
- Hand fills (3 of 3): entry and exit txs of the first 3 winners found on chain, status 1, 4 logs each, blocks match the swap index (entry 100905011 -> exit 100909523, etc.). Pool/price rows reproduce from the index.
- Exit depth: pool ETH after the exit swap is only 0.012-0.04 ETH (one 0.12 ETH outlier: 0.029). A $10 position (~0.0025 ETH bought) is worth 0.014-0.12 ETH at exit = 0.2-1.7x the whole pool. Last-swap price is NOT fillable at that size: 0 of 8 pass depth >= 3x position value.
- Constant-product haircut (value*D/(D+value)): forward mean 3.95x -> 1.48x (n20, pulls = 0). Still >1.05 but this is a crude estimate; real impact in concentrated v3 liquidity may be worse or better.
- Outside sellers near exit: all 8 winners had 1-4 sells by 1-4 distinct wallets in the last 120 s before exit, 0 buys. Exit happens into a sell-off: the last price is a falling price.
- Verdict: sellability NOT proven; haircut mean 1.48x keeps it alive. Not VERIFIED. Needs an archive-state eth_call sell at the exit block (not just latest) for the 8 winners; sellprobe works at latest only, pools are older than 1 h. Decide-by 3 Nov stays; recommend: if archive sim not available by then, retire.