Methodology
Version 1.1.0 · the editorial wall between money and ratings is the product.
Proof-of-usage verification
Crypto uniquely lets us prove a reviewer actually used the thing they're reviewing. We offer tiered, cryptographic verification:
- Verified Holder — you connect a wallet and sign a message; we confirm on-chain that the address holds (or held) the token. The signature is free and authorizes no transaction.
- Verified User — for exchanges and companies, prove account activity with a read-only API key. We verify, then discard the secret immediately and store only a hashed reference.
- Verified On-chain — for protocols and dapps, we confirm your address interacted with the contracts.
Unverified reviews are allowed, clearly marked, rate-limited harder, and weighted far lower.
Reputation-weighted ratings
The headline score is not a naive average — naive averages are trivial to game. Each review's weight rises with the reviewer's verification level, reputation, and account history, and new accounts are dampened during a cool-down. We always show both the raw mean and the weighted score so nothing is hidden.
Reputation itself moves only through the events below. Every change is written to an append-only log, so any reviewer’s score can be reconstructed from its history.
| Event | Change |
|---|---|
| A review of yours is published | +2 |
| That review carried proof of usage | +5 |
| Someone marks your review helpful (capped at 20 per review) | +0.5 |
| Someone marks your review unhelpful | -0.25 |
| A review of yours is removed for a policy violation | -10 |
| You flag a review and moderation agrees | +3 |
| You flag a review and moderation disagrees | -1 |
The editorial wall
Paid status, claimed status, and featured placement never influence the organic rating or sort order. Our ranking function takes zero input from any billing or claim data — this is enforced in code and tested. Featured items, if any, appear only in clearly labeled slots.
Trust Score v1.1.0
The CryptoChit Trust Score combines objective signals with verified-review sentiment into a single transparent number from 0 to 100. These are the exact weights the scoring code uses — this table is generated from the same constants, so it cannot drift from the formula.
| Component | Weight | What it measures |
|---|---|---|
| Verified review share | 26% | Proportion of reviews backed by cryptographic proof of usage. |
| Reputation-weighted rating | 22% | The weighted rating, normalised to a 0-1 scale. |
| Review volume | 13% | How many published reviews exist, saturating at 50. |
| Recency | 9% | Share of reviews written in the last 180 days. |
| Entity age | 9% | Months since the contract deployed or the listing was created, saturating at 24. |
| Market depth | 9% | Log-scaled market capitalisation. Tokens only. |
| Security posture | 12% | Scanned contract risk: honeypot behaviour, mint authority, pausable transfers, reclaimable ownership. Absent when a scan has not succeeded. |
We show a Trust Score only when an entity has at least 3 published reviews and enough signals to be meaningful. Below that we say we cannot score it rather than publishing a confident-looking number built on two reviews. Components that cannot exist for an entity type — market depth for a company, for instance — have their weight redistributed across the rest, so nothing is penalised for a signal it could never have.
Contract security flags come from GoPlus, a third-party scanner. When a scan has not succeeded, the security component is absent rather than assumed clean — a missing scan lowers our confidence in the score, it never counts as a clean bill of health. Audit status stays operator-curated: there is no reliable programmatic audit registry, and guessing would be worse than saying we don’t know.
We will never ask you for an exchange API key. Verified User works through read-only OAuth: you authorise a read-only scope at your exchange, we confirm account activity once, and we discard the token immediately — it is never stored. If a site claiming to be CryptoChit asks you to paste an API key, it is not us.
Changes to this formula
Re-weighting moves every score on the site, so each change is published here rather than shipped quietly.
| Version | Date | What changed |
|---|---|---|
| v1.1.0 | 2026-09-02 | Added a security-posture component from scanned contract flags, and began preferring on-chain contract deployment date for entity age. The other six weights were reduced proportionally to make room. Scores moved as a result. |
| v1.0.0 | 2026-09-01 | First published Trust Score. |