The score

A number from 0 to 100 attached to every graduated token, and the breakdown that produced it is on the token page. The number alone is the least interesting part.

What it is

A sum of signals. Each signal is one measurable fact about the launch, worth some points in one direction or the other, with a sentence saying what it found.

The score exists to make dozens of separate readings comparable at a glance. The reason it always ships with its breakdown is that the breakdown is the actual product: "6 of 6 bundlers are repeat players across 410 prior launches" tells you something. "31" does not.

What the signals look at

Grouped by what they are asking. Exact weights are not published — see what this handbook does not contain.

The launch itself : How much of the supply the launch group took, whether they are still holding it, whether the float that remains is tight or already distributed, and how quickly the curve filled.

The people : Whether the wallets involved have appeared together before, whether the deployer has launched many tokens, whether any of them are flagged, and how the crews attached to this launch have performed historically.

The market : Liquidity, trading volume against pool size, the balance of buys to sells, and the market cap relative to where a completed curve should land.

The token : Mint and freeze authorities, whether metadata is still mutable, holder count and concentration, and whether the ticker duplicates something recent.

Signals worth zero

Several signals are recorded and deliberately not scored. They appear in the breakdown with no points and a note.

This is not an oversight. A signal that measured backwards from intuition, or measured nothing at all, is more useful recorded honestly than quietly deleted — it stops the same idea being re-added next quarter, and it keeps the observation available if the population changes.

If a signal in the breakdown carries no points, that is a statement: we looked, and it does not predict what you would expect.

Hard blocks

Some findings are not points at all. They refuse the launch outright, whatever it scores:

• The launch could not be reconstructed — no score can be earned on evidence that does not exist.
• The market cap is past the ceiling — the move being priced has already happened.
• No market cap could be read, so the ceiling cannot be checked.
• The crew has a long record of never producing anything.

A block is always shown with its reason. A refused token is not a low score; it is a different outcome.

The bar moves by cohort

A given score does not mean the same thing in every cohort, so the bar a token has to clear is set per cohort rather than globally. A number that is emphatic for a fast launch is unremarkable for an instant one.