Methodology · V1

How we measure exchange transparency

The VIF Score is a 0–100 measure of how independently verifiable a centralized exchange is. It combines six evidence-based components using a published, deterministic rubric.

Glass figure holding balanced scales, representing transparent and impartial evaluation

VIF Score

Calculation

Each component is calculated on a 0–100 scale and clamped to that range. The overall score is the weighted sum, rounded to one decimal place.

VIF Score = 0.25A + 0.25L + 0.15F + 0.15P + 0.10M + 0.10I
Asset Verification25%
Liability Verification25%
Proof Freshness15%
Public Verifiability15%
Methodology Transparency10%
Incident Disclosure10%

Grade bands

A+

90–100

A

85–89.9

A−

80–84.9

B+

75–79.9

B

70–74.9

B−

65–69.9

C+

60–64.9

C

55–59.9

C−

50–54.9

D+

45–49.9

D

40–44.9

F

Below 40

VIF Score determines an exchange's position in the transparency rankings.

Component rubrics

The rules below reflect the implemented v1 policy. Bonuses and penalties are applied to each component before the result is clamped to 0–100.

01 · 25%

Asset Verification

What evidence supports the exchange’s reserve-asset claims?

  • Audited financial statements, base 95: the relevant entity’s customer assets and broader asset position are formally audited. The base rises to 100 when the exchange also publishes a Merkle/zk cryptographic Proof of Reserves the public can self-verify.
  • Cryptographic asset proof plus an asset attestation, base 80.
  • Cryptographic asset proof plus agreed-upon procedures, base 70.
  • Cryptographic asset proof without independent audit, base 65.
  • Standalone third-party asset attestation, base 60.
  • Standalone asset agreed-upon procedures, base 45.
  • No meaningful asset verification, base 20.
  • −15 when confidence in the recorded facts is estimated.

The financial-statement tier requires audit scope covering the relevant operating entity and customer assets; reaching the full 100 additionally requires a user-verifiable cryptographic proof. Public wallet verification and freshness are evaluated separately.

02 · 25%

Liability Verification

Can customer balances owed by the exchange be checked?

  • Audited financial statements, base 95: the relevant entity’s customer obligations and broader liabilities are formally audited. The base rises to 100 when the exchange also publishes a Merkle/zk cryptographic liability proof the public can self-verify.
  • Global cryptographic liability proof plus a liability attestation, base 70.
  • Global cryptographic liability proof plus agreed-upon procedures, base 67.
  • Global cryptographic liability proof without independent audit, base 65.
  • Standalone third-party liability attestation, base 55.
  • Standalone liability agreed-upon procedures, base 45.
  • Exchange disclosure without independent verification, base 30.
  • Base 20 when total customer liabilities are not demonstrably verified.
  • −15 when confidence in the recorded facts is estimated.

Assurance raises a component only when its documented scope covers that component. Agreed-upon procedures report findings but do not provide an assurance conclusion. Public reproducibility and user inclusion are evaluated separately.

03 · 15%

Proof Freshness

How frequently is proof published, and how old is the latest snapshot?

  • Score the demonstrated cadence ceiling: continuous 100; daily 95; weekly 85; monthly 75; quarterly 55; yearly 35; ad hoc or unknown 25.
  • Score absolute snapshot age: ≤1 day 100; ≤7 days 90; ≤31 days 75; ≤92 days 55; ≤366 days 35; older 15.
  • Use the lower of the cadence ceiling and absolute-age score.
  • No dated proof: 0.

Continuous requires continuously updated asset and liability evidence. A live reserve-wallet balance alone does not qualify, and claimed cadence should be supported by publication history.

04 · 15%

Public Verifiability

How much of the evidence can outsiders independently reproduce?

  • Base: 25, reduced to 10 when all evaluated evidence is opaque.
  • Liability verification: LPOR +25; global ZK proof +18; Merkle-sum +10; plain Merkle inclusion, auditor-only, or none +0.
  • User inclusion/root check: full +25; partial +12; none +0.
  • Reserve-wallet disclosure: full +25; partial +12; none +0.

05 · 10%

Methodology Transparency

Does the exchange explain how its proof was produced?

  • Base: 60.
  • +15 when a public proof URL is available.
  • +10 when at least two citable sources are recorded.
  • +10 when a cryptographic method such as Merkle or zk is named.
  • −10 when confidence in the recorded facts is estimated.

06 · 10%

Incident Disclosure

What relevant incidents are present in the current record?

  • Base: 85.
  • Initial penalty by severity: high 15; medium 8; low 4.
  • Penalty retained by age: under 1 year 100%; 1–3 years 75%; 3–5 years 50%; over 5 years 25%.
  • Unresolved incidents retain 100% of their penalty. Monitoring incidents retain at least 50%.
  • The retained penalty is reduced by 20% when the exchange responded.
  • The component score has a floor of 40.

A response does not by itself establish that disclosure was timely or adequate. Historical incidents retain a minimum penalty and are never erased.

Evidence and confidence

Verifelse records source URLs and point-in-time Proof-of-Reserves facts for each exchange. The rubric distinguishes complete public verification from partial disclosure and unavailable evidence.

Full
The evaluated artifact can be independently checked using the published evidence.
Partial
Some evidence is public, but independent reproduction is incomplete.
None
The relevant artifact is not publicly available or cannot be independently checked.

Confirmed

Recorded facts are corroborated by primary or official sources.

Estimated

Facts are inferred or only partially sourced. Relevant component penalties are applied.

Interpretation and limitations

Not a solvency guarantee

Proof of Reserves may omit liabilities, encumbrances, governance risks, or future liquidity conditions.

Point-in-time evidence

Published proofs and disclosures describe a particular snapshot and may become stale.

Evidence availability

A low score may reflect missing public evidence; it does not prove undisclosed misconduct.

Association, not causation

A positive Transparency Delta does not establish that transparency caused market outperformance.

Versioning

VIF methodology v1

Initial deterministic public scoring policy. Future revisions should document changes to component rules, weights, or evidence requirements.

Corrections and challenges

Better evidence can change the record

Exchanges, researchers, and users may challenge underlying facts or evaluations. Updates are made when the submitted evidence warrants a correction.

Challenge an evaluation →