Graded in the open.
We won't sell you a cherry-picked backtest - the pools that fail vanish from public data, so a naive “our grades predicted doom” chart is just survivorship bias dressed up as proof. So we do the harder thing. Every week we grade every live pool and publish the record- recorded before the outcome is known, and anchored on-chain so it can't be quietly rewritten. Then we test the grades against that record the honest way, counting the pools that vanished as failures - including when the answer isn't flattering.
The latest grades at a glance.
How the grades have moved.
The share of pools at each grade, A through F, across every weekly snapshot. Descriptive only - this shows what we said, not yet how it played out.
Pools graded A or B, week over week.
Every snapshot, newest first.
Why this is the honest proof.
Recorded before the outcome. Every Monday we grade every live pool and write the result down - while nobody yet knows how those pools will perform.
Every pool, no cherry-picking. The failures stay in the record too, so there's no survivorship bias hiding the misses.
Append-only, anchoring soon. The record only grows; a past week can't be edited. On-chain anchoring on Base is launching soon to make that cryptographically verifiable.
The record is recorded before outcomes are known and is append-only. On-chain anchoring on Base - writing each snapshot's hash to an immutable ledger so the record is verifiable by anyone, not just trusted - is launching soon. The predictive track record matures over the coming months; grades are descriptive estimates of current pool quality, not guarantees.