How FP1 turns evidence into an auditable decision record.
FP1 publishes the rules used to build, revise, grade, and retire each instrument. The evidence, thresholds, judgments, disagreements, and falsifiers remain visible so another reader can challenge the result.
Eight stages with accountable judgment.
Eight stages sit between an observation and a decision brief. People execute every stage today, with software assisting retrieval and formatting. The labels show where software assists, where published rules apply, and where a named person remains accountable. No stage runs unattended.
Software can organize evidence and apply explicit rules. It cannot own the choice of the pivotal uncertainty or the argument about how that uncertainty changes a decision. A named analyst remains responsible for both.
Today, people execute all eight stages under the rules in NCB-004. Software assists retrieval and formatting. “Rule-guided” means a person applies published grading rules consistently; it does not mean a program makes the judgment. A computational engine intended to assist stages 1 through 4 is in development. It is not producing the material currently published.
Six rules that make an FP1 conclusion auditable.
Pre-registration
Calls, tiers, thresholds and falsifiers are published before the tracking window opens, with the publication date on the record. A threshold re-set inside its own window voids the pre-registration claim for that signal: the signal is retired in public and the change is logged, rather than quietly improved.
Source and independence grading
Every claim carries two separate grades. Quality asks how good the source is. Independence asks whether the citations behind a claim are actually different sources, or one source restated. Five outlets carrying one wire story is a single source wearing five hats, and the grading says so.
Relationship mapping
Signals are not read one at a time. Each relationship asserted between signals carries its own confidence grade, separate from the confidence in the underlying observations, because a well-measured fact can support a weakly argued relationship.
Disagreement preserved
Analytical lenses remain separated by method. Where they disagree, the brief records convergence, documented dissent, or insufficient evidence. Editorial synthesis must not erase a documented conflict.
Grading discipline
Grading happens against dates fixed in advance, not when the news is convenient, and the grade date is separate from the publication date so that lateness is visible rather than absorbed. Verdicts come from a closed vocabulary: held, held and strengthened, held and watched, tripped, retired. A closed set prevents a grading from softening its own failures.
Self-falsification
Individual calls fail by their falsifiers. The method fails by four stated triggers: threshold churn, a falsifier tripping with no prior drift toward the line, a reference class that cannot represent what actually moved, and unresolvable grading drift between reviewers. Each forces a public restructuring rather than a quiet patch.
Limits that remain visible.
Reference-class selection is judgment
Two honest analysts can assign the same dimension to different historical populations. The method makes that judgment inspectable, not infallible. The class is named so it can be disputed.
Some histories are thin
Where a threshold is anchored to a benchmark rather than to a base rate, it is labelled as such and carries at most medium conviction.
Position is ordinal, not calibrated
A chart shows ranking and direction, not probability. A marker twice as far from a line is not twice as safe, and nothing here claims otherwise.
The grader grades itself
FP1 both places the thresholds and scores them. The mitigations are structural: separation of analytical roles, a closed verdict vocabulary, and public receipts. They are not a substitute for external audit.
Method documentation and implementation notes.
The method summarised above is documented in full in the methodology papers. These are the reference texts, not the entrance.
Signal Scoring and Historical Thresholds
The anatomy of a call, the derivation of a historical threshold from a named reference class, the scoring model, the two board geometries and the prohibition on sharing one graphic, the grading discipline across sweeps, the evidence-to-decision pipeline, and the conditions under which the method is revised or retired.
The Novacene Composite (FNC-1)
The running v0.1 measurement instrument used a seven-stock public-equity proxy across four substrates, paired with a four-market belief overlay. The paper documents constituent selection, weighting, normalization, limitations and falsification criteria. A broader Tier-2 constituent set was drafted but was not the production instrument for the 2026 Register window. See the implementation note.
The Register
What FP1 tracks, the threshold that counts as movement, the direction each dimension can be proven in, and the date it is graded. Also what FP1 declined to register, and why.
Analytical lenses
Three optional analytical lenses, each with a question, a method, an output, and a failure condition. Verification, Context, Integrity, and the Rāwī translation pass are documented separately.