Insights
Integrity
Their number · our number · our honest number — per month, with the exclusion method published
A settlement total is only as good as the assumption that the two sides of each payment are independent parties. In this market that assumption fails often enough to move the headline: addresses paying themselves, pairs cycling the same funds back and forth, and fleets of payees funded by a single wallet that pays nobody else. Those patterns settle on-chain exactly like real demand, and they are published elsewhere as real demand.
We measure them and take them out — but we publish all three figures, because an adjusted number with no visible adjustment is just a different unexplained number. Every exclusion below is a specific pattern between identified addresses in a specific week, reproducible from the record.
Not a fraud accusation. Self-trade has legitimate uses — testing, liquidity bootstrapping, internal accounting. The claim here is narrow and factual: this volume is not two independent parties transacting, so it should not be counted as market demand.
Integrity-adjusted volume· integrity_v0.1.0
What the peer index published, what we measured, and what survives after measurably synthetic volume is taken out — per month, per chain.
Our number
$45.73M
the published clean basis across 14 months
Measurably synthetic
$23.25M
50.9% of the 14 screened months
Our honest number
$22.47M
screened months only
Nov 2025 is the sharpest case on this chain: 82.2% of that month's basis is volume where the payer and the payee are not independent parties. It is published elsewhere as real demand.
| Month | Their number | Our number | Measurably synthetic | Our honest number | Share flagged |
|---|---|---|---|---|---|
| Apr 2025 | not published | $315.5K | $0.000.0% | $315.5K | |
| May 2025 | not published | $32.6K | $0.000.0% | $32.6K | |
| Sep 2025 | not published | $439.10 | $12.432.8% | $426.67 | self trade |
| Oct 2025 | not published | $3.96M | $332.5K8.4% | $3.63M | self trade |
| Nov 2025 | not published | $27.86M | $22.90M82.2% | $4.96M | self trade |
| Dec 2025 | not published | $3.27M | $16.7K0.5% | $3.25M | self trade |
| Jan 2026 | not published | $1.18M | $720.040.1% | $1.18M | self trade |
| Feb 2026 | not published | $2.07M | $704.670.0% | $2.06M | self trade |
| Mar 2026 | not published | $2.22M | $697.550.0% | $2.22M | self trade |
| Apr 2026 | not published | $1.92M | $707.450.0% | $1.92M | self trade |
| May 2026 | not published | $1.02M | $667.390.1% | $1.01M | self trade |
| Jun 2026 | not published | $774.9K | $597.880.1% | $774.3K | self trade |
| Jul 2026 | not published | $896.4K | $701.120.1% | $895.7K | self trade |
| Aug 2026 | not published | $213.5K | $38.100.0% | $213.5K | self trade |
What gets excluded, and what does not
Only volume we can convict edge by edge is subtracted: an address paying itself, two addresses paying each other back and forth, and fan-out cohorts where a single funder pays a fleet of payees that receive money from nobody else. Each of those is a specific, reproducible pattern between two identified parties in a specific week — not a guess about intent. What we deliberately do not subtract is concentration: a week where one payee takes an unusual share of transactions is a bound on how much of the market is one relationship, and subtracting it would delete most of the ecosystem along with every legitimate high-frequency integration. That bound is published separately, as a range, and never as a verdict. On this chain the convicting rules were: self-trade (payer and payee are the same party).
The adjusted figure is bounded by construction — it can never exceed the raw figure and can never go below zero — because the subtraction is capped per payee, per day, before it reaches this table.
Against the public peer index
methodology →- October 202598.6%
- January 2026127%
Under 100% is settlement they hold and we do not; over 100% is settlement we hold and they do not. Both directions are published, and the surplus decomposes back to transactions.
The peer index publishes no month-grained figure today — only a cumulative all-time snapshot — so the “their number” column stays empty rather than being back-derived from a running total.
as of 2026-08-09T08:25:56Z
Wash Bounds
A narrower concentration-analysis subset, not the headline market total · neither bracket is “true volume”
All weeks included — upper bound
Weeks where top payee exceeds threshold — wash-inflation risk
Below-threshold weeks — conservative floor
as of 8/9/2026
Check the workings
- →How we know — per rail: the source, the verification method, and what the last gate run returned, each with the date it was measured on.
- →Transparency — how much of the observed record is attributed to a named service, and how fresh each pipeline is.
- →Discovery radar — the debut weeks where this shape shows up first, before it is big enough to move a monthly total.