Metryval

A 19-dimension personality measurement instrument, synthesizing 36 peer-reviewed frameworks.

Metryval is a measurement instrument, not a categorical typology. It synthesizes 36 peer-reviewed psychological frameworks into 17 dimensions across 5 domains (the PAREN architecture: Perception, Agency, Responsiveness, Engagement, Navigation), and cites 9 additional frameworks as methodology precedents. Per-framework reception is tracked publicly in the Framework Reception Ledger and framework-reception-ledger.json. When a source framework’s academic reception changes (Polyvagal Theory was withdrawn in May 2026 following a 39-expert consensus paper; MFT MFQ-30 was retired as an operationalized instrument in our v1.18++ cascade), the ledger updates and a corresponding entry lands in the change log.

What it is

  • 19 dimensionsacross 5 PAREN domains (asymmetric 5+3+4+4+3). Big Five’s OCEAN is the bedrock; PAREN extends it by integrating frameworks that OCEAN does not directly measure (moral foundations, moral reasoning, emancipative values, regulatory focus, autonomy, authenticity, interoception, temporal orientation, prosocial orientation).
  • ~370 items in production(post-Decision-F native N+E wiring; to be cut toward the ~200-item launch target after the pilot). Pilot battery ~430 items. ~110–128 min full instrument; ~12 items / ~3 min Lite snapshot above the funnel.
  • Deterministic scoring engine.No LLM in the scoring path. Reference-centroid distance scoring in 19-dimensional space (Euclidean at launch, Mahalanobis once a norming sample of N≥500 establishes empirical Σ).
  • Audit-tracked. The Framework Reception Ledger and the Policy Decisions Ledger are committed to the repo, versioned, parser-checked against the markdown source on every build, and published as fetchable JSON.

What it is not

  • Not a categorical typology.Metryval outputs reference patterns, not types. The framing throughout the product is “your pattern resonates with X” rather than “you are X.” This is structural: the empirical natural cardinality of personality clustering is small (k = 3–4 per Asendorpf 1999 and Gerlach 2018), and Metryval avoids the MBTI categorical-type failure mode by delivery-time assembly against framework-validated reference centroids.
  • Not a generative-AI product. The scoring engine is deterministic. LLMs are used in the Blueprint long-form rendering layer (clearly disclosed at /methodology/transparency), never in scoring.
  • Not a hiring or selection instrument. The architecture is consumer self-understanding. B2B employment, selection, and personnel use cases are structurally excluded across all surfaces. The lock is Policy Decisions Ledger entry P-01, and it is binding regardless of pricing tier or contract type.
  • Not yet empirically validated at scale. Reference centroids are theoretical placements in a framework-validated 19-dimensional space. The pilot validation roadmap (Wave 1 + Wave 2 Prolific; pooled N≥1,500 powers the four-gate retirement workflow) is documented in the methodology pages. The Y3 methodology paper publishes empirical convergent, discriminant, and incremental-validity results including failures.

The open data

Two canonical JSON artifacts back every claim on this page:

Both ledgers are parser-checked against their markdown sources on every build, so the JSON cannot silently drift from the canonical state. A standalone open-source repository carrying the deterministic scoring engine + reproducible sample tests is the next Track 4 deliverable.

Read the methodology

Five structural defense surfaces document the architecture for methodology-curious readers:

  • /methodology/anti-barnum — the “specific enough that only ~1 of 53 reference patterns recognizes themselves” threshold, operationalized as ≥1.5 SD deviation on ≥2 dimensions
  • /methodology/frameworks — the framework canon and per-entry reception state
  • /methodology/language — the public banned-vague-qualifier list and the somatic-specificity requirement on narrative outputs
  • /methodology/transparency — California AB 2013 disclosure, LLM-usage scope (no LLM in scoring), data handling
  • /methodology/policy — Policy Decisions Ledger; P-01 hard exclusion of B2B selection / hiring; EU AI Act Annex III item 4 non-applicability

Take the Lite snapshot

The free 12-item snapshot returns your nearest reference pattern in approximately 3 minutes. It is the only consumer-facing path that does not require an email. The full ~370-item instrument is the paid path; Blueprints unlock at $19. Pricing canon at /pricing.

Take the Lite snapshot →

Who runs it

Metryval is operated by Metryval Pte Ltd, a Singapore private limited company. The public byline on Substack, X (@metryval), and editor correspondence is the pseudonym Tomas Cantor. The pseudonym is locked and disclosed openly here; legal accountability runs through Metryval Pte Ltd directly, registered with Singapore’s ACRA and publicly searchable.

Contact: tomascantor@metryval.com (operator) · hello@metryval.com (general).