Every number with its caveat.
Why each figure on this site comes from one ledger, in its wording, with the condition under which it holds.
One ledger
Every figure on this site comes from one claim ledger, in its wording, with the condition under which it holds and where it was measured. A number that does not have its caveat yet is not published, and a page never retypes a figure: it reads the ledger's entry.
The three we lead with
+2.1 Gini points over the strongest open model of its class at 100 matured loans, across 208,443 held-out loans on 13 lending books from 12 lenders (+1.9 on test partitions alone); +1.3 across 100–500 loans, confirmed in a pre-registered held-out test (95% interval +0.5 to +2.2; 9 of 12 lenders).
Pretrained on synthetic credit economies; no borrower record enters pretraining.
Scores an application in about 50 ms (54 ms median, 57 ms p95) on a 16-thread server CPU and about 20 ms on a GPU, matching the research model's predictions.
Fixed before the books were scored
The hypotheses behind the ranking results were fixed before the held-out books were scored: seven of ten confirmed, all ten reported.
What we do not claim
- Calibrated probabilities out of the box. Ranking is validated; probabilities are recalibrated per portfolio before use as PDs.
- Reason codes for zypl.Foundation scores. Not shipped. Zehn's reason codes describe Zehn itself.
- Higher approvals or lower losses. Only from a lender's own matured pilot evidence, with its permission.
Research results, not this demo
These are research results from zypl labs, not results of the demo. The demo runs on public lending books whose outcomes are published, so every number it shows can be checked. The full list is on the evidence page.