Internal check record
The detailed record of each check, as written in May 2026, beside what the model does with the signal today. These are internal checks PhaseFolio ran in May 2026 on a small set of oncology programs. They are not a validation of the model and were not peer reviewed. The model uses published values, and every value can be changed in the asset modeler.
01The record
Evidence is linked on each row.
| Signal | Context | What we saw | What the model does | As recorded | Evidence | Related |
|---|---|---|---|---|---|---|
| Biomarker quality | NSCLC · Phase II/III | Lined up with outcomes | Uses the published Schwaederle 2016 values (1.35× genomic, 0.85× protein-only) for oncology solid tumor at Phase II/III, after review. | +5.2pp held-out AUC over the structural baseline, stable across cohort sizes. The genomic_validated cohort odds ratio was 5.59 vs the Schwaederle (2016) literature anchor of 1.35 — we ship the lower anchor with the overshoot disclosed. | JSON ↗ | Backtest methodology (M2, retired)PoS reference data |
| Phase 1 objective response rate (ORR magnitude) | Oncology solid tumor · Phase II/III | Added little | Shown as a flag beside the valuation. Does not change PoS. | A joint biomarker x ORR-bucket model beat the biomarker-only baseline by only +0.5pp held-out AUC (paired DeLong p=0.48), and the comparison is statistically unpowerable — detecting a +3pp gain at this baseline needs ~830 drugs; the cohort is 85. This followed the Phase 1 finding that the two signals combined fell below baseline (-0.3pp at 43-drug coverage) — the double-counting that motivated the joint-table test. | JSON ↗ | Backtest methodology (M2, retired)PoS reference data |
02How the checks were run
Small, internal and dated.
Each check held back 30% of a set of 85 oncology programs with known outcomes and compared how well the model separated approvals from failures with and without the signal. A set this small can suggest a direction but cannot settle one, and the analysis was ours alone. The model does not use values fitted to these programs: it uses the published figures described in the methodology.