THE PROVENANCE LEDGER · COHORT 2026
The clock is running.
Audit us.
Anyone can publish a score. The question a serious allocator asks next is what happens when that score turns out to be wrong — and whether they will ever hear about it from us, or only from the write-down.
So here is the arrangement, set out before we have a single outcome to report. Every Clarity Score™ is sealed and time-stamped the moment it is issued, and the whole cohort gets graded in public against what actually happened. The terms are fixed now, while we still have no idea which way the results will go.
Sealed scores in the ledger
Opening soon
The live counter opens with Cohort 2026.
An empty ledger is the honest state of this page.
We could fill this space with the benchmark corpus — 110,000+ Clarity Scores on public deal data — and it would look impressive. It would also be the wrong number. Those documents were scored retrospectively, against outcomes that had already happened.
A ledger only means something if the score was recorded before anyone knew the answer. That is what Cohort 2026 is, and it is why there is nothing to show you yet.
// WHAT SEALED MEANS
Three properties, and what each one actually rests on.
The audit is anchored to SHA-256 fingerprints of the exact documents it read. Change a byte of the deck and the fingerprint stops matching — loudly, and without anyone having to notice. That is the part no policy is required to enforce, because arithmetic enforces it.
Each record carries the version of the Clarity Framework™ that produced it. Methodologies improve; a score issued under one version keeps that version's meaning rather than being quietly restated under a newer one. We do not re-score history.
A delivered score is not revised. We are labelling this as policy rather than as a control, because that is what it is today: the fingerprint proves a document was not swapped, not that a record was never touched. Independent immutability is on the roadmap and we will say so plainly when it ships.
// THE ANNUAL RECKONING
June 2027. Every Cohort 2026 score, and how it held up.
A rating agency that never publishes its hit rate is running an opinion service. The first edition of the reckoning covers every score sealed during 2026 — not a selected sample, and not only the ones that aged well.
Setting the terms in advance is the entire point. Anyone can define success after the results arrive.
- The distribution of scores issued, not a highlight reel.
- How each band aged against what happened next — survival, follow-on, down round, wind-down.
- The methodology version in force for each vintage.
- Cases where a high score preceded a bad outcome, named as such.
- Bands that do not separate. If 85+ and 50–69 age the same way, the score is not measuring anything.
- Kill shots that were not fatal, in numbers. A floor applied to a business that went on to work is a false positive and we will report the count.
- Drift between vintages that the framework version does not explain.
None of this is a prediction claim. The Clarity Score measures whether reasoning was defensible at the time a decision was made, and good reasoning still loses sometimes. What the reckoning tests is whether the number separated better reasoning from worse across a whole cohort — which is the only thing a standard can honestly be judged on.
The record is only useful if it covers your own deals.
Ledger opens 2026 · pre-launch