56 checks for reviewing the appraisal behind a residential transition loan. Every one published in full, with the provenance of its threshold — which is more than can be said for any number currently in use.
The standard states what each check tests and when it fires. Where a check is original to Sphinx, the method used to compute it is part of the implementation rather than the specification — the same separation any standards body maintains between a spec and a reference implementation. Industry-baseline thresholds are shown in full; they exist in every review tool on the market. Sphinx-original firing conditions are stated qualitatively; the exact parameters are implementation detail.
Every threshold in mortgage lending is inherited convention — a comp beyond a mile, gross adjustments over 25%, a desk review that busts at 10%. Ask where any of them came from and nobody can tell you. Each check here carries its provenance: whether the number is a stated industry baseline, or fitted to realised sale prices. When a threshold moves, the change publishes with it. Calibration is fitted to each contributor's archive at onboarding and pooled as loans exit. The status page is the canonical record of where that stands; the governance page records how a threshold moves, and the validation protocol is the pre-registered test it has to pass.
How to cite: consistent with RTL-STD-001 v1.0, RTL-V-01 — document ID, version, check ID. Findings record the threshold values used, by value. The methodology note is written for a rating-agency or diligence reader; the governance page records how a version changes.
Whether the sales chosen actually support the conclusion. The recurring failure here isn't a bad comp, it's a thin set treated with the same confidence as a deep one.
Whether the grid's arithmetic is market-supported or reached for after the number was already chosen. The published research on appraisal overvaluation locates most of the distortion here.
No conventional review standard contains an equivalent to these, because a conventional appraisal values a house that already exists. Every one is computable from the appraisal and the budget alone.
The routine work that takes an underwriter twenty minutes and catches something roughly once a month. Low yield individually, which is exactly why it should be automated rather than skipped.
Licence and discipline are table stakes and already public. Whether this appraiser actually works this county is the one that predicts error, and the one nobody looks at. Outputs here are consumed as an internal adjustment — never as a ranking, never sent to the appraiser, and never a basis for requesting a revision.
Found something wrong? That's the point of publishing it. Tell us which check and why — standard@sphinx.credit.
RTL-STD-001 v1.0 · published · every threshold labelled with its provenance. Industry Baseline — initial threshold drawn from GSE guidelines where documented, common practice otherwise. Subject to outcome calibration. Sphinx-Original — check or threshold that exists in no other review standard. Calibrated — threshold fitted to realised outcomes, with its sample size.
The Sphinx standard describes analytical review of an appraisal report. It is not an appraisal, not a valuation, and not a review appraisal under USPAP Standard 3. Outputs are internal analytical inputs for credit decisioning. They are not communicated to appraisers and are not a basis for requesting revision to any report.