Layer II · the companion volume

Named and Counted,
Not Blamed nor Accounted (v7)

A metrological instrument for measuring directional dysfunction: naming what had no name

Briar Greenway
ThinkWell Labs Metrology · 2026
Companion to the live instrument at toroid.fyi/sabbot.

Some things stay invisible until they are named.

Four objects, one discipline: this manuscript (the argument); the instrument (the apparatus); the rubric (the replication manual); and the evidence anchors (the audit trail). The live version at toroid.fyi/sabbot is the working tool; it will keep moving after the page has set. Read the book for the reasoning and the instrument for the readings—and where they disagree, trust the instrument. It is newer.

Contents

1. The Case Against the Book. Begin where a hostile reader would.
2. Preface. What this is; what it is not.
3. The Doctrine of Restraint. Measure, do not pronounce; intent in two registers.
4. Terminology and the Ladder of Claims. Directional asymmetry, obfuscratic, Sabbotarchy.
5. Part I. The five signatures.
6. Part II. The composite, the bands, and the uncertainty I will not fake.
7. Part III. Evidence discipline: occurred, projected, contested, correlation.
8. Part IV. Temporal dynamics: velocity, breadth, and the synchrony test.
9. Part V. Resource operations: the adjunct layer, held outside the number.
10. Part VI. Validation: the controls, and the reform that scores low.
11. Part VII. Application: the seven architectures, dated and frozen.
12. Part VIII. What would falsify this; and the two tests it has not yet faced.
13. Appendix. The two-instrument stack, distribution, citation.

The readings

What follows is the instrument’s output, not the argument. Every number, table and definition on this page is quoted from Part II, Part III and Part VII of the book. The three figures are drawn from those same numbers; they are renderings of the tables, not additional findings. Where a figure and the book disagree, the book governs — and where the book and the live instrument disagree, the instrument is newer.

The composite and its bands

The five signatures fold into one number per architecture. Each is scored zero to one hundred; the composite is the weighted average.

Index = 0.25·S1 + 0.20·S2 + 0.15·S3 + 0.10·S4 + 0.30·S5

The index sorts into four bands.

The thresholds are not pulled from air. Reaching 70 takes the high range on at least four of the five dimensions; reaching 85 takes it on all five at once. They will move as more architectures are scored, and any move gets logged, not slipped in.

On the precision I will not fake. These scores are not measurements drawn from a sampling distribution. They are one analyst’s judgments, and a confidence interval computed over a sample of one is fabricated precision. So no score in this book carries a ±. Read every number here as a claim about which band, never as a precise quantity.

The seven architectures, and the controls

Set against the reform lineage, the separation is total. Grace Commission, 14. Reinventing Government, 19, on a larger workforce cut than DOGE. President’s Management Agenda, 20. Three prior reforms in the low genuine band; seven current architectures from the high band into the extreme. The gap between the two clusters is the finding, and it is a gap the instrument was demonstrably able not to produce, since the lineage proves it can score a large hard reform at 19.

The current dimension scores and composites

Two features of this table outweigh any single row. The spread is the first: the composites run from the mid-seventies to about ninety, which means the instrument is grading rather than slamming every case to the ceiling. The VA architecture scores meaningfully below the IRS one, and an instrument that could not tell those two apart would be worthless. This one tells them apart. The floor is the second. Two architectures sit in the high band rather than the extreme, and they sit there because their readings really are lower.

How fast they arrived

The OBBBA row is the one to sit with, because it is the cleanest evidence in the book that the instrument is not simply scoring my politics. At the first checkpoint, before enactment, OBBBA scored 16: squarely in the genuine band, below the reform lineage’s Grace Commission. The instrument looked at the thing in its latent state and declined to call it a signature. A measure that wanted to convict would have convicted at 16. This one waited for the evidence and moved when the evidence moved.

The evidence tags

• Occurred. A documented event, verifiable against the contemporaneous record: a statute, a court filing, an agency notice, dated reporting. Eligible to anchor a claim about a realized outcome.
• Projected. A modeled or forecast outcome, a CBO estimate, a mortality model. Legitimate evidence, but it may anchor only a projected claim, and it must say so in the row it sits in.
• Contested. Magnitude, attribution, or characterization is disputed. The tag does not, by itself, lower confidence. Whether it does depends on the quality of the contestation.
• Correlation. A relationship has been measured between variables, with no claim about causal direction.

A note on the numbers

The scores, the correlation, and the projected ranges in this part are frozen as of June 2026 for print. The live instrument at toroid.fyi/sabbot keeps updating against fresh evidence, so its current figures may differ from these. Where this part and the instrument disagree, the instrument is newer. This part is the dated record of where the readings stood when the book went to press.

The two tests it has not yet faced

Part VIII states both in advance, with their predicted outcomes on the record, before any scoring. Convict a friend: point the five signatures, S5 above all, at an architecture I am politically sympathetic to. Acquit an enemy: find one action of the administration I oppose that, measured honestly, is ordinary symmetric incompetence, and score it Genuine or Mixed. Neither has been run.

Until the book contradicts its author somewhere you can see it, you are right to suspect it has only ever agreed with them, and a measurement that only ever agrees with its maker has measured, among other things, its maker.