Falsifiability

The test
register.

A scientific test requires a specified model and a result that could count against it. PDT’s test programme therefore identifies the assumptions, calculations, observations and rejection criteria for each proposed test.

Kinds of test

Mathematical verification: Checks that a construction or implementation satisfies its stated mathematical properties.
Physical model test: Tests whether an apparatus or dataset agrees with a specified instrument, dynamics or noise model.
Prospective prediction test: Evaluates a fixed forecasting rule using information available before the outcome.
Engineering comparison: Tests net performance against a strong matched controller.
Distinctive physical prediction: Requires a calculated difference from an appropriate competing physical model.

Rejection criteria apply to the specified model and assumptions being tested. A null result against one historical proposal does not by itself refute the framework as a whole.

Current register

Proposed and completed tests

Mathematical verification

Mathematical exampleDossier

The exclusion operators form a complete instrument (Theorem R05.1).

Inputs and assumptions: Orthogonal projectors summing to the identity; an active set of m ≥ 2 labels.
Observable or quantity: Sum of the operators K†K and the survivor probabilities.
Predicted result: The sum equals the projector P_S; with populations 1/2, 1/3, 1/6 the six histories sum to 1 and survivor totals equal the populations.
Comparator: Exact arithmetic.
Uncertainty and calibration: None; exact fractions.
Rejection criterion: Any history set whose probabilities do not sum to 1, or survivor totals that differ from the populations.
Source: R05

Physical model test

Proposed protocolDossier

A specified apparatus implements the exclusion instrument.

Inputs and assumptions: Preparation, apparatus interaction, noise and calibration model, record rule.
Observable or quantity: Measured outcome frequencies and post-measurement states.
Predicted result: Not yet calculated; requires the apparatus model.
Comparator: An appropriate conventional quantum model of the same apparatus.
Uncertainty and calibration: Calibration errors, full-rank noise (Theorem R05.4), finite sample size.
Rejection criterion: Frequencies outside pre-registered tolerance of the specified instrument.
Source: R04 R05

Prospective prediction test

Model specified, evaluation not yet runDossier

A fixed tolerance rule bounds disagreement with a later readout.

Inputs and assumptions: Information h before the forecast; calibration set Θ; tolerance δ.
Observable or quantity: Prediction rate, error rate among issued predictions, no-prediction cases.
Predicted result: Under the stated model with the true parameter in Θ, disagreement is bounded by δ.
Comparator: The stated bound and simpler forecasting baselines.
Uncertainty and calibration: Calibration-set coverage, dataset size, interval estimates.
Rejection criterion: Error rate among issued predictions significantly above δ on held-out data under the stated coverage assumptions.
Source: R04 R06

Engineering comparison

Proposed protocolDossier

PERSIST-Q delivers a net control improvement.

Inputs and assumptions: Control problem, estimator with source version, matched resources and latency.
Observable or quantity: Net task performance, end-to-end latency, deadline misses.
Predicted result: Not yet calculated.
Comparator: Strongest implementable conventional controller and simpler baselines.
Uncertainty and calibration: Held-out evaluation, device or backend transfer, paired or randomised design.
Rejection criterion: No significant net improvement over the strongest matched controller.
Source: R04 R06

Distinctive physical prediction

A distinctive physical test requires a quantitative difference from the relevant comparator. The current entries below identify the models and calculations needed to establish such a test.

Historical proposals

The earlier experimental trio

Earlier PDT papers proposed a coherence floor, a Yukawa-type force signature and photon parity or helicity effects. These proposals are retained in the historical research record, but are not current predictions of the foundational suite. A testable prediction requires a specified physical model, a quantitative calculation and an experimental protocol.

Coherence floor in matter-wave interferometry
Historical proposal, withdrawn
Residual visibility loss is not the same observable as a surviving coherence floor.
Yukawa-type short-range force
Historical proposal, withdrawn
No supporting derivation and protocol in the current foundations.
Photon parity or helicity effects
Historical proposal, withdrawn
No supporting derivation and protocol in the current foundations.

QM V (original reference)QM III (original reference)

The testregister.