Payrit added to coverage, and prior weeks restated

14

Venues in coverage

Payrit added, history backfilled

$11.04M

Week of Jul 27, restated

published at the time as $10.69M

+$283,429

Quidax re-attribution

81% of the restatement

Two changes to coverage this week, from the same pipeline run.

Payrit is now in coverage

Payrit joins as our 14th venue, with history backfilled to June rather than starting from today — so week-on-week comparisons across the series stay valid.

It is small. Payrit contributed $26,008 in the week of 3 August, 0.3% of everything we measure, across 56 deposits from 8 addresses. We are reporting it at that size rather than waiting for it to become material, because coverage should be a statement about what we measure, not about what is big enough to be flattering.

Prior weeks have moved

The same run applied an attribution revision across all history. The effect on the most recent published figure:

• The week of 27 July now reads $11.04M, against the $10.69M we published on 3 August.
• Quidax's record for that week now reads $5.36M, against the $5.07M we published.

Payrit's addition accounts for only $9,970 of the $348,116 difference. $283,429 — 81% of it — is a re-attribution on Quidax: deposits that belonged to Quidax and were not previously being counted as such.

So the record we announced was understated, and the correction runs in Quidax's favour.

Why this keeps happening

This is the second restatement in two issues, and it is worth being direct about why rather than letting it look like instability.

Attribution is not a lookup. An address is assigned to a venue on evidence — funding patterns, sweep behaviour, infrastructure relationships — and that evidence improves as more of a venue's on-chain footprint is mapped. When it improves, previously unattributed deposits get attributed, and history moves.

The alternative would be to freeze published figures and let the series drift away from what we actually know. We would rather restate in public and say exactly what changed.

Both figures are correct on their own basis. What matters is that they are not the same basis, and that the difference is documented here rather than discovered by someone re-running the query.