Pattern Fit Scoring Rubric¶
Scoring rubric — instantiates Archetype Overmatching Guardrail
Scores a candidate match on weighted fit criteria to produce a graded degree-of-fit, turning "it kind of resembles X" into a defensible number with its reasoning attached.
Pattern Fit Scoring Rubric grades how well a case fits a proposed archetype rather than gating on whether it fits at all. It fixes a set of weighted criteria, each scored on an anchored scale, and each score carries a one-line justification — so the output is a graded degree-of-fit plus an auditable, per-criterion breakdown. Where the Pattern Fit Checklist asks "are the necessary features present?" as a yes/no gate, the rubric asks "to what degree, and where is the fit strong versus thin?" Its distinctive value is measurement with the reasoning stapled on: a total nobody can inspect is worthless, so every point is defended in writing.
Example¶
An investment committee is drawn to a pitch: "this is a marketplace play — network effects will take over, like the great two-sided marketplaces." Instead of voting on the resemblance, they run the marketplace-fit rubric. Criteria — single-geography liquidity, cross-side network-effect strength, take-rate defensibility, fragmentation on both sides, winner-take-most dynamics — are weighted by how load-bearing each is to the thesis, and each is scored against anchored level descriptions with a short justification. The startup scores well on fragmentation but lands near the bottom (≈1 of 3) on cross-side network effects, because supply is commoditized and multi-homes. The weighted total settles mid-band: "partial marketplace fit, weak on the load-bearing criterion." The rubric doesn't cast the vote — it hands the committee a graded, itemized score to argue with, and it makes glaring that the one criterion carrying the whole thesis is the weakest one.
How it works¶
Criteria are weighted by importance and scored on an anchored scale — level descriptions, not a bare 0–5 — so scorers drift less. The weighted aggregate expresses degree of fit, and a sensitivity read shows which criterion is driving the total. Crucially, each criterion score is recorded with its justification, so the rubric doubles as the written argument for the match rather than an opaque number.
Tuning parameters¶
- Weights — how much each criterion counts. Concentrating weight on the discriminating criteria, versus spreading it evenly, decides whether a fatal miss can be outvoted by peripheral matches.
- Scale granularity and anchoring — 0–1 versus 0–5, and how explicitly each level is described; richer anchors cut scorer drift.
- Aggregation rule — a weighted sum versus a gate (a minimum floor on the key criterion). A pure sum is compensatory — many weak matches can mask one disqualifying gap.
- Scorer calibration — one scorer or several reconciled; multiple independent scorers reduce idiosyncrasy.
- Pass band — where on the total the "fit" threshold sits.
When it helps, and when it misleads¶
It suits graded, multi-factor patterns and makes matches comparable and auditable across cases. Its failure modes are false precision (a crisp number laid over squishy judgements), weight-fiddling to reach a desired total (the classic run-backwards abuse), and compensatory aggregation letting surface matches mask a structural miss — the representativeness trap, where resemblance to a prototype inflates the felt fit while base rates are ignored.[n1] The disciplines that keep it honest: set weights and anchors before scoring the case, keep a hard floor on the load-bearing criterion, and carry the per-criterion breakdown forward, not just the headline total.
How it implements the components¶
structural_fit_check— it grades each required structural feature on an anchored scale rather than passing or failing it, producing a degree-of-fit per feature.case_evidence_profile— each criterion is scored from the case's actual evidence, assembling a scored profile of what the case genuinely shows.fit_rationale_record— every score carries a written justification, so the completed rubric is the documented reasoning behind the match.
It does NOT set the binary necessary-feature gate — that's Pattern Fit Checklist; it does not compare the case against neighbour archetypes — that's Case Comparison Matrix and Differential Pattern Review; and it does not convert its score into the decision's confidence tag — that's Decision Confidence Label.
Related¶
- Instantiates: Archetype Overmatching Guardrail — the rubric supplies the graded, defended measure of fit the guardrail weighs.
- Sibling mechanisms: Pattern Fit Checklist · Decision Confidence Label · Case Comparison Matrix · Differential Pattern Review · Anti-Pattern Review · Counterexample Search Session · Precedent Distinction Memo · Red-Team Pattern Match Review · Review Queue
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: The rubric evaluates a candidate match against anchored weighted criteria and returns a justified degree-of-fit finding with sensitivity evidence.
Nearest alternative: Analysis, Modeling & Optimization — Weighted calculation supports the result, but the defining product is an assessment disposition on an existing candidate match.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Education & Pedagogy
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Weighted scoring rubrics are a canonical educational assessment form for making criteria explicit.
Related originating lineages:
- Psychology — Pattern Fit Scoring Rubric is rooted in psychology and behavioral science: Behavioral research on representativeness motivates replacing resemblance judgments with explicit weighted criteria.
- Statistics & Experimental Design — Experimental design and statistics materially shaped Pattern Fit Scoring Rubric through randomization, inference, sensitivity analysis, and validation.
- Systems Thinking & Cybernetics — System-archetype diagnosis supplies the structural features being scored.
Review resolution: Light authoritative-source research resolves the primary-origin disagreement in favor of education and learning science. University of Michigan: Rubrics, Scoring and Grading directly documents the defining practice or theory described in the selected origin rationale. Other listed domains are retained only where the blind reviews identify material co-development or translation; broader adoption remains separate as domain_reach=multi_domain.
Attribution caveat: The particular archetype-fit rubric is synthesized from assessment and systems practice.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
Notes¶
A rubric scores fit to one archetype in isolation; a high score does not rule out that a neighbouring pattern would score higher still. Pairing it with a comparison across candidates (Case Comparison Matrix or Differential Pattern Review) is what turns a strong single-pattern score into a defensible best-fit claim.
[n1] The representativeness heuristic (Tversky & Kahneman) — judging how well a case belongs to a category by how much it resembles a prototype, while neglecting base rates — is the core bias this whole archetype guards against, and the specific one a compensatory rubric can smuggle back in. ↩