cryptosnifHow the crash test works
How the crash test works. Evidence, limits and the next questions to examine.
Evidence and Editorial Integrity
We distinguish between a project's own claims, independent observations, original calculations and our editorial interpretation. Each piece of evidence has a specific scope, date and set of limitations. We identify assets by their stable identity, network and relevant contract addresses to ensure precision. Historical documents remain historical records. Reusing raw data depends on provider rights.
The Review Sequence
Our editorial process follows a strict sequence. We examine the stated promise, established facts, unknowns, potential consequences, the verdict and the evidence that could change it. Missing evidence is always visible in the report. We do not treat unsupported solidity as proof of fraud. This structured method allows readers to see exactly what was known, what was missing and how the conclusion was reached without hiding gaps in information.
Scoring Method and Calibration
A deterministic score gives the same result for the same evidence and method version. Our candidate method is version 0.1.0. Calibration, the process of testing rules for consistent assessments across project families, remains incomplete. It uses 13 criteria with a total weight of 100. Scores are calculated using integer values from 0 to 4 for each criterion, which are then normalized to a scale of 100. Unknown, contradictory or expired evidence is excluded from the calculation. This exclusion reduces the overall coverage percentage. Public project scores remain masked until the calibration phase is complete.
Eligibility and Coverage Rules
Overall eligibility requires at least 85% evidence coverage. Additionally, we require verified asset identity, control powers and supply mechanisms before assigning a score. Coverage refers to the proportion of applicable criterion weight supported by valid, current evidence. If evidence is missing or invalid, it does not count toward the score but lowers the coverage metric.
Project Families and Checks
Checks vary depending on the project family, such as proof-of-work networks, smart-contract platforms, stablecoins, governance tokens and application tokens. Non-applicable checks require explicit authorized exclusions. An unknown status cannot be relabeled as non-applicable. Rankings require comparable reviewed projects, the same method and at least 90% evidence coverage. The current pilot publishes no eligible ranking because these strict conditions are not yet met across the dataset.
Score Limits and Risk Visibility
Critical risks remain visible independently of the average score. The score assesses documented project quality. It cannot predict prices, returns or the absence of loss. Crypto assets can lose value. Current missing evidence and required further checks remain explicit in all reports. This limitation prevents the score from being misinterpreted as a guarantee of safety or financial performance.
| Check | Weight | Evidence validity |
|---|---|---|
| Exact asset identity | 3% | 30 days |
| Observable use | 15% | 30 days |
| Economic role of the token | 10% | 90 days |
| Supply mechanisms | 10% | 30 days |
| Ownership concentration | 8% | 7 days |
| Value capture | 7% | 30 days |
| Control powers | 10% | 7 days |
| Security verification | 7% | 90 days |
| Decision process | 8% | 90 days |
| Accountability | 7% | 90 days |
| Incident history | 6% | 30 days |
| Operational continuity | 4% | 30 days |
| Execution conditions | 5% | 1 days |