The Board
One file, start to finish.
This is how a grade is made. Every step reproducible from public transactions. Two real wallets from the desk's records — one clear, one ambiguous — graded by the exact same chain.
Worked Example 1 — The Clear Case
The HALO dev
CXoUJHtBGnuNR6zJ3zm8nTnNzCwXxFViXh29d1kTJQmK
1
Transactions — what the chain shows
500+ launches recorded20 launches in the last 24h16 graduated17 peak-and-collapse endings9 successes (ATH ≥ $100K)
The 500 count is a lower bound — the pulled history window truncates, and the file says so rather than bluffing a lifetime number.
2
Attribution — observed vs inferred
OBSERVEDEvery launch above names this wallet as its creator in the on-chain record. The counts are arithmetic on public transactions — anyone can recount them.
INFERREDThat one person or one team controls this wallet. The record attributes launches to the wallet. It does not establish who holds the keys, and the grade never claims it does.
3
Classifications — mechanical labels only
17 prior coins carry the peak-and-collapse ending under the desk's published proxy: ATH ≥ $20K, then a fall below 10% of ATH and under $5K. That is a measurement of how those coins ended — printed as measured, not as a claim about why.
The cadence is also a measurement: 20 launches in 24 hours, on top of 500+ recorded. Volume at that rate is itself record data.
4
Grade
PROBATION — ADVERSE RECORD
The key exists — everyone launches. The grade sits at Probation and the adverse record prints on the file in the open: 500+ launches · 17 peak-and-collapse endings · 20 more in the last 24 hours. The coin's story — a genuine news narrative — changes nothing. The grade reads the wallet, not the pitch.
5
Privileges — what the grade sets
Launch caps: smallest — size capped
Placement: no front page
Pairing: SOL only
The way back is also mechanical: launch, don't produce exit events, build endings that aren't collapses — the record updates and the grade can climb. Nothing about this grade is permanent except the record underneath it.
Worked Example 2 — The Ambiguous Case
The Zoo dev
7xXKejGTNdyz6iAisVzSt1oXPWgrASsbHzsqj9tNX2jw
Resolved from the Zoo mint GwpXch98ft9MRQByRpjG36SV79mSKMt93L2oo3Pmpump.
1
Transactions — what the chain shows
1 launch recorded1 graduated0 peak-and-collapse endings0 successes
That one coin graduated — ATH ≈ $67.5K — then fell to ≈ $9.7K within hours the same night. One data point. It proves neither skill nor intent, and the file says exactly that.
2
Attribution — observed vs inferred
OBSERVEDThe launch, the graduation, and the price path above.
UNREADABLEThe funding trail behind this wallet. The coin's birth sits under ≥42,000 signatures — beyond the trace budget — so where this wallet's money came from could not be read. The method stops there and labels it, instead of guessing.
3
Classifications — mechanical labels only
One GRADUATED ending. No exit event is recorded. No liquidity withdrawal is recorded. No adverse classification applies — and none is invented to fill the gap.
The provenance step returns UNREADABLE, which is a statement about our tooling, not about this dev.
4
Grade
PROBATION
An unreadable trail is a starting point, not a verdict. No adverse finding is printed, because there is no adverse evidence. The key starts where every unknown key starts, and builds its record in the open from here.
5
Privileges — what the grade sets
Launch caps: smallest while the record is thin
Placement: no front page yet
Pairing: SOL only
Same starting line as the clear case — different file. This key's record is one graduation and a question the chain hasn't answered yet. What it does next decides the grade, not what we couldn't read.
LAUNCHKEY — WORKED EXAMPLE · METHODOLOGY v1