# Audit v3ride (auditor, 16 Oct 2026) - verdict FIX, NOT VERIFIED
Hunt: buy first print >=$1M cap (age<120m), ride to 5x / 0.3x stop. Target runners 3-5M.
1. Data: pub.md 03:23 UTC. ALL arm n46, mean 1.15x, median 0.28x, 9 tp / 33 stop / 2 pull / 2 time. All rows are src=fwd, but entry times start 6 Oct = the tape rebuilt by the same script (own-tape replay), not independent forward flow. No separate live-forward count.
2. Stats: lo80 0.82x (below 1.0), mean excluding top 2 winners 0.99x. Last 5 days (11-15 Oct) n20 mean ~0.94x. Kill rule (mean<1.0 after 30) not yet tripped, but the edge is 9 five-x winners and nothing more.
3. Fills: tp exits are booked at exactly 5x (optimistic, 3 s gaps); fee+impact flat 3% each side at $1M cap, no depth check; pull = 0 modelled only after entry. Sellability (eth_call) and outside sellers not run for these pools.
4. Look-ahead: trigger uses 300k-first-cross >=180 s earlier and next-swap entry, no future info found. Not hand-rebuilt from chain.
Verdict: FIX / not verified. To verify: (a) forward-only count since 16 Oct start (needs n>=15 real forward), (b) tp fill haircut (sell at 4.5x) still mean >1.1, (c) outside-seller + sell sim on the 9 winners. Decide 20 Oct: retire if mean ex-top2 <1.0 at n>=60.