Bundle · Build & Ship
You said the value setswere closed.Go and look.
Every model assessment is a self-assessment. You mark your own homework, and the marks come back better than the data would. These three close that loop: one survey asks what you have declared, two sweeps go and measure the two claims that are cheapest to check, and the rollup reports where they disagree.
Deterministic and offline. Nothing connects, queries, or changes anything.
Grades models, fields and record pairs — never people.
Instant download. Three engines, three workbooks, two playbooks, one rollup.
01
One claim, two measurements
The survey is what you believe. The sweeps are what is there.
01
The claim
What you say you have
Six dimensions marked 0, 1 or 2 per domain — including whether your value sets are closed and your identity rules are stated. Self-assessed, on purpose.
MODELED · PARTIAL · UNMODELED · $89
02
The evidence, on values
What the columns contain
Every value a field actually holds, against the set you declared — and the variant carrying 6,488 rows that is a second status rather than a typo.
SETS HOLD · DRIFT FOUND · SETS BROKEN · $49
03
The evidence, on identity
Whether the key holds
Candidate record pairs your declared key cannot separate — and the ones your own data says can never be merged.
DISTINCT · SAME ENTITY · IDENTITY CONFLICT · $69
CONFIRMED, KNOWN GAP, OVERSTATED — in that order
A KNOWN GAP is work you already planned: you said the sets were partial, the sweep agrees, and nothing about your picture of the world was wrong. OVERSTATED is worse, because you said the sets were closed, the data says otherwise, and everything built on top of that answer was built on a belief. The shipped example carries one of each, from real runs of all three engines.
02
What the rollup prints
Run all three, feed their JSON to one script, and it tells you which of your claims your own data does not support.
MODEL & IDENTITY TRIO - CLAIM AGAINST EVIDENCE
==========================================================================
The survey covered 4 domain(s) and read NO MODEL.
Closed value sets OVERSTATED
claim weakest mark 2 of 2 - every branch-driving field has a closed, reviewed set
evidence SETS BROKEN (Enum & Status-Value Drift Sweep)
you claimed it was fully declared in every domain, and the columns contain values the set does not allow
that sweep says start at: order.status -- In Progress
Identity rules KNOWN GAP
claim weakest mark 1 of 2 - a stated functional key per type, and the systems use it
evidence IDENTITY CONFLICTED (Entity Identity & Functional-Key Sweep)
you did not claim this was fully declared (weakest mark 1), and the sweep agrees there is work to do
that sweep says start at: P-04
--------------------------------------------------------------------------
TRIO VERDICT: MODEL OVERSTATED
0 confirmed / 1 known gap / 1 overstated (of 2 pairs)
start at: value_sets (order.status -- In Progress)
A known gap is work you planned. An overstated claim is a belief
you have been acting on. That is why they rank in that order.
This rollup computes no score. It reports where a claim and a measurement
of the same thing disagree. Confirmed means these two checks agree - not
that the model is right.This is a real run against the three Systems’ own shipped samples, which is why the model reads overstated. Two things are worth noticing. It takes the weakest mark across your domains rather than the average, because a claim that holds in three domains out of four is not one you can act on everywhere. And it ranks OVERSTATED below KNOWN GAP on purpose — a known gap is work you already planned, while an overstated claim is a belief every branch and report downstream was quietly built on.
03
What you get
Three complete Systems, each usable on its own, plus the piece that only makes sense once you have more than one.
Three engines
Zero-dependency Python. Each runs offline, reads only the files you point it at, and prints a verdict you can paste into a ticket.
Three workbooks
Every engine reproduced as live formulas — 99, 362 and 66 of them — with no macros. Excel, Google Sheets, or Numbers.
Two playbooks and a runbook
How to run the survey with the people who know the domain, and what to do when a claim comes back overstated.
Three worked samples
Six domains, five fields across 220,936 rows, eight candidate record pairs — each built around a pair that isolates exactly what changes a verdict.
The claim-vs-evidence rollup
One script, three JSON files, one answer: which of your declarations your own data does not support. No score.
Editable rule packs
Dimensions, thresholds, signal weights and normalisation rules all live in JSON files you own.
04
What it will not do
Closing the gap between claim and evidence is not the same as being right.
Built for
- Teams who have written down a domain model and never checked it against production.
- Anyone whose status fields and customer records are about to be handed to an agent.
- Architects who need to show a stakeholder the difference between what was declared and what is there.
It will not
- Query your database. You run the distinct-value query and bring the candidate pairs; they grade what you bring.
- Check that a declared set contains the right values. A closed set with the wrong values in it reads as closed to all three.
- Merge, migrate or fix anything. They name the field, the pair and the claim; the work is yours.
- Turn agreement into correctness. MODEL HOLDS UP means these checks agree with what you claimed.
These Systems grade a model you describe and data you paste in. They are working aids, not legal or compliance advice, and not audits or certifications. None of them scores, ranks, or assesses people. MODEL HOLDS UP proves these checks agree with what you claimed — not that your model is correct.
05
Questions people actually ask
Short answers first, with the caveat after it rather than in front of it.
Three complete Systems, each sold separately, plus a claim-against-evidence rollup. Domain Ontology Readiness Survey ($89) asks what you have declared. Enum & Status-Value Drift Sweep ($49) goes and looks at what your columns actually contain. Entity Identity & Functional-Key Sweep ($69) goes and looks at whether your key can actually tell two records apart. $207 separately, $159 together.
One is a claim and two are measurements of that same claim. The Survey is self-assessed — you mark whether your value sets are closed and your identity rules stated. The two Sweeps read real data. The rollup pairs them up and reports where a claim and a measurement of the same thing point in different directions. It computes no score; the verifier asserts the engine contains no blending arithmetic at all.
Because a known gap is work you have already planned. You said the value sets were partial, the sweep agrees, and nothing about your picture of the world was wrong. OVERSTATED means you said the sets were closed, the data says otherwise, and everything built on top of that answer has been built on a belief. The shipped example carries one of each so the difference is visible.
Because the rollup takes the weakest mark across your domains rather than the average. A claim that only holds in three domains out of four is not a claim you can act on everywhere, and averaging would round it back up to looking fine. The verifier asserts this directly: two domains at full and one at partial gives partial, where an average would have given full.
Yes — all three are sold individually at $89, $49 and $69, and each is complete on its own with its own engine, workbook, worked sample and playbooks. The bundle is those three at $159 plus the rollup, which only becomes useful once you have run more than one of them.
That these two specific checks agree with what you claimed. Not that your model is correct — a closed value set with the wrong values in it will read as closed by both the Survey and the Sweep. It finds the gap between what you believe and what your data shows, which is a different and narrower thing than being right. All three grade models, fields and record pairs — never people.
Mark your own homework, then check it.
Three complete Systems and the rollup that finds the claim your data does not support. $159 together, $207 apart.
These Systems grade a model you describe and data you paste in. Working aids, not legal or compliance advice, and they never score or rank people.
Sold by RedHub AI LLC · Secured by Stripe · redhub.ai