# auditor2 10 Oct: v2ramp tx.from check (#715, dataeng v2ramp_tx.db)
- 2,628 distinct buy tx_from over 111 pairs. Only 1 wallet (0xe71a69..., dust seed buy) is in >=20 pairs (108); 0x0e7761... in 9; every other wallet in <=3 pairs. So no single operator spreading buys: the crowd is mostly different wallets. PASS on operator concentration.
- 91 of 111 pairs show exactly 21 distinct buyers: still looks like a capture/cap artifact, not a crowd (open question to testmgr3: does the recorder stop at N buys?). Not fatal for P&L.
- Sellers in the first 3500 blocks exist in all 10 audit pairs (5-11 each), all also buyers: outside sellers confirmed present, so exits are not into a one-sided pool (matches the v2 sim PASS).
- Still open before VERIFIED: real sell on a fresh pair (Ben decision #681 pending) and the 21-buyer cap explanation. Registry unchanged (verified false).