The original Ashgang51 complaint states no loss amount. Earlier investigation
wording referring to a larger complaint amount or larger claimed loss was
unsupported and is withdrawn. The separately reported June campaign estimate
has not been independently bound to this particular cited transfer.

The8.690148pUSD amount is the actual cited transfer, not an established total
incident loss. That8.690148pUSD already existed in the source Safe at the
parent block before five market sales generated another0.958pUSD. The later
transfer cannot be uniquely assigned to those sale proceeds.

One09:46:37UTC exchange co-settlement included the source selling10YES shares
for0.53pUSD and the previously Car-funded notworthit wallet selling10NO shares
for9.47pUSD. Those complementary positions were burned to release10units of
collateral, which the exchange allocated to the sellers. This was not a
direct source-wallet payment to notworthit and does not establish theft,
coordination or shared control. Five earlier executions of the identical
maker order were observed, first at07:26:29UTC; its signed timestamp is not
an independently measured creation or actual signing time.

Earlier Car-authorized private funding53886pUSD and owner-authorized returns
48454pUSD establish a financial relationship with notworthit, not ownership
or theft intent. Direct private cash/native joins across the specified Car
and June endpoints were zero. The ordinary market co-settlement positive
is retained separately; a bounded private-pair negative does not erase it.

Previously published proof archives remain unchanged. This note and the
current primary-source context supersede inconsistent earlier narrative
wording without changing raw transaction records. No causal/control/private
payment/coordination relationship establishing Car involvement in a drain
has been proved by this evidence.

Main curated files:
  positions/POSITIONS-CURATED.json
  positions/INCIDENT-ORDERS-VERIFIED.json
  context/INDEPENDENT-COSETTLEMENT-AUDIT.json
  context/CURATED-PUBLIC-CONTEXT.json
  links/DIRECT-PAIR-CURATED.json
  root/PUBLICATION-CORRECTION.json
  root/TRACKED-TWO-CHAIN-STATE-VERIFIED.json

Run all portable checks without changing the evidence files:
  python3 REPLAY_ALL.py /tmp/positions-offline-replay.json

REPLAY_ALL verifies file hashes, JSON/gzipJSON and Python syntax, then copies
the evidence into a temporary directory. Five verifier processes run with
sockets and DNS disabled. Outputs stay in that temporary copy. The replay
tests retained byte/event/signature/state consistency, not chain consensus,
provider completeness or the identity of private people. Full social-post
bodies, HTML, screenshots, full market rules and other local research are
excluded; source URLs, metadata and original-response hashes are retained.
