D-16 · The Corbel Score
[D-16]Layers
The Corbel Score
One number for a token, out of 90, and the four parts that explain it. A clean score is the absence of known traps, not a promise.
The four parts#
Each check scores pass 100, caution 60, unknown 50, fail 0. A part is the weighted average of its checks; the total is the weighted average of the parts, scaled to 90.
| Part | Weight | Checks inside (weight) |
|---|---|---|
| Exit | 45% | Honeypot (3), exit at a loss (2), transfer tax (2) |
| Control | 25% | Owner powers (2), lookalike ticker (1) |
| Liquidity | 20% | Liquidity lock (1) |
| Origin | 10% | Creator history (1) |
Caps that override the total#
Some failures are bad enough that no other result should hide them. The total is capped when one of them fails:
| Check fails | Total at most | Because |
|---|---|---|
| Honeypot | 10 | it cannot be sold back |
| Lookalike ticker | 25 | it copies another token's name |
| Transfer tax | 30 | a tax takes part of every trade |
| Exit | 35 | a sale returns a fraction of the cost |
Why 90, not 100#
Code checks see the contract and the pool. They cannot see whether a team is real, whether a website tells the truth or what a creator will do tomorrow. The last ten points are left for that, and nothing on the chain can earn them. Project research under each check reads what can be read outside the code.
