Experiments / Benchmarks / Falsifiability

Experiments: Executed Benchmarks vs. Proposed Tests

This page separates executed, reproducible residual-first benchmarks from proposed falsifiable tabletop tests. The executed benchmark records are methodological and make no claim of new physical effects. The proposed roadmap describes discriminants intended to test QDL-style closure and scaling against standard parameterizations.

Use this page after reading the Framework page. Framework defines the closure and admissibility rules; this page states how those ideas are being tested, benchmarked, or turned into future experimental targets.

Residual-first adequacy tests Executed benchmark records No new effects claimed from benchmarks Proposed tabletop discriminants Pre-stated failure conditions Torsion / NV / cavity / metamaterial platforms
Track A: Executed Benchmarks Track B: Proposed Tests How to Interpret Results Framework Research Hub Animation
Executed Benchmark records
These records demonstrate benchmark discipline: declared datasets, declared families, declared budgets, and residual-first diagnostics.
In Progress Scaffolding
  • Pre-registration templates for tabletop discriminant tests.
  • Replicability scaffolding for artifacts, scripts, plots, and expected outputs.
  • Partner outreach for platform-specific sensitivity studies.
“In progress” means support work for falsifiability and auditability, not experimental confirmation.
Proposed Falsifiable tests
  • Torsion-balance discriminants using geometry and mass-configuration sweeps.
  • NV-center discriminants using geometry and field sweeps.
  • Cavity discriminants using length, frequency, geometry, and material sweeps.
  • Metamaterial discriminants using dispersion-collapse signatures.
“Proposed” means specified as a test target. It is not a demonstrated empirical effect.
Jump to what you want Quick nav
Use Track A for executed records. Use Track B for proposed platform tests. Use Interpretation Logic to separate benchmark adequacy from empirical support.
Track A — BenchmarksExecuted → Track B — RoadmapProposed → Benchmarks and DOIsList → Interpretation logicRead → Torsion balancePlatform → NV centersPlatform → CavitiesPlatform → MetamaterialsPlatform →
Two-track layout: Track A is executed and methodological; Track B is proposed and falsifiable.
Track A: Benchmarks Track B: Roadmap Interpretation Collaboration

Program Overview

How QDL maps framework claims into auditable benchmarks and proposed discriminants.

What this page does and does not claim

QDL proposes a closure-based admissibility rule in a 3L + 2F basis and asks whether this rule can be tested through residual structure, scaling behavior, and discriminant signatures across multiple platforms.

  • Executed benchmarks are methodological and do not claim new physics.
  • Proposed experiments describe falsifiable discriminants and explicit failure conditions.
  • Empirical support requires pre-stated patterns, controlled sweeps, and comparison against standard model families.
  • Null results are meaningful if the failure condition was declared before the test.

For the framework definition, start at Framework. For the broader paper library, use Publications.

Key references

Framework-first references and experimental roadmap anchors.

Publication Record Resources

Track A — Executed Benchmarks and Null Tests

Residual-first, auditable comparisons using declared model families and public or reproducible data.

Executed · reproducible records Residual-first · structure matters Declared families · stated budgets No new effects claimed
What “benchmark” means here

Benchmarks are explicit comparisons between declared model families under a stated parameter budget, evaluated primarily by residual structure rather than aggregate fit magnitude. Each record is designed to be auditable: dataset provenance, model forms, fit controls, and residual plots are published alongside the analysis.

  • Decision rule: coherent residual structure indicates model inadequacy.
  • Declared parameter budget: model families are specified before fitting.
  • Reproducibility: scripts, CSV files, figures, and reports are versioned with stable DOIs.
  • Scope limit: benchmarks test method discipline, not final physical interpretation.

These benchmarks are evidence for audit discipline; they are not presented as standalone validation of QDL’s strongest cross-domain physical claims.

Quick links: executed records

Stable DOIs for immediate citation and replication.

Each record is intended to make provenance, preprocessing, model families, and residual structure inspectable.

Executed benchmark summaries

Scope-limited summaries for orientation.

1. Optical Cavity Resonance Benchmark
Domain: precision photonics / metrology Residual-first diagnostics distinguish stabilized traces from uncontrolled resonance scans. Coherent residual structure is treated as a sign of model inadequacy rather than automatically as new physics.
DOI: 10.5281/zenodo.18076864

2. NV-Center ODMR Benchmark
Domain: solid-state quantum sensing Residual-first benchmarking compares baseline spin-Hamiltonian fits with a strictly reduced comparison family using public ODMR data, emphasizing structural adequacy rather than fit magnitude alone.
DOI: 10.5281/zenodo.18069870

3. Benchmarks and Null Tests
Domain: method / null-test discipline Defines residual-first reporting conventions and provides executed null-test anchors where a reduced family is expected to pass if no additional structure is present.
DOI: 10.5281/zenodo.18057668

Publishing discipline

What stays fixed across benchmarks to preserve auditability.

  • Public provenance: datasets are linked with stable identifiers.
  • Declared preprocessing: no hidden filtering, windowing, or cherry-picked segments.
  • Declared model families: reduced and baseline comparisons are explicit.
  • Declared parameter budget: added flexibility must be disclosed.
  • Residual-first reporting: residual structure is treated as primary evidence.
  • Scope discipline: method results are not overstated as new physical effects.

This discipline is designed to avoid ambiguous “fit improvement” narratives and enable clear success or failure judgments that other groups can reproduce.

How to Interpret Results

A compact decision tree separating benchmark adequacy from empirical support for QDL-specific discriminants.

Decision logic

Benchmark passes: residuals are noise-like under the declared reduced family. This supports adequacy of that reduced family for the dataset under the stated pipeline choices.

Benchmark fails: coherent residual structure persists. This indicates model inadequacy or pipeline artifact; it does not by itself identify missing physics.

Discriminant test supports QDL: a pre-stated QDL-distinct scaling or pattern appears under controlled sweeps, while the standard comparison family fails to match it within uncertainty without erasing the constraint by adding flexibility.

The page is structured to reduce scope creep: benchmarks are method; discriminants are physics tests.

Transform-invariance discipline

QDL-adjacent analyses often involve re-expression through units, bases, or measurement-chain parameterizations. This page treats unit changes, basis reparameterizations, and explicit measurement-chain re-expression as audit transforms when they are declared in advance.

Transforms that change model class, introduce undisclosed calibration degrees of freedom, or alter data selection are interpretation changes and must be declared as such.

This rule is intended to prevent post-hoc reinterpretation under a different analysis choice.

Claim-status summary
Executed benchmark A reproducible methodological test of residual structure or audit logic. It does not claim a new physical effect.
Proposed discriminant A pre-stated test target that could support or weaken a QDL-specific scaling or closure claim if executed under controlled conditions.
Positive result Requires pre-stated pattern, controlled sweeps, residual-first reporting, and failure of standard comparison without extra flexibility.
Negative result Counts as meaningful if the failure condition was declared in advance and the sensitivity was adequate.
Ambiguous result Requires more controls, better provenance, tighter budgets, or stronger separation between constrained and unconstrained model families.

Track B — Proposed Falsifiable Tabletop Tests

Four complementary platforms. Each block states a proposed discriminant and an explicit failure condition.

Torsion balance
Scaling discriminants versus geometry and mass configuration.
Proposed geometry sweep
  • Declare constrained and standard model families.
  • Run a geometry or configuration sweep.
  • Evaluate residual structure with fixed parameter budgets.
Go to torsion →
NV centers
Frequency-shift discriminants under field and geometry sweeps.
Benchmark exists Proposed test
  • Use benchmark workflow as method precedent.
  • Pre-register a sweep and constrained discriminant.
  • Hold calibration degrees fixed and report residual-first.
Go to NV →
Cavities
Length-frequency sweeps as clean L-F probes.
Benchmark exists Proposed test
  • Use existing apparatus for geometry or material sweeps.
  • Test constrained scaling against standard families.
  • Failure means no discriminant or only extra-flexibility success.
Go to cavities →
Metamaterials
Dispersion-collapse candidates as discriminant signatures.
Proposed sweep / collapse
  • Find or construct a dataset with a clean sweep axis.
  • Declare a constrained collapse target.
  • Collapse must beat flexible refits under fixed budgets.
Go to metamaterials →
Roadmap anchor

QDL Experimental Validation Protocol: 10.5281/zenodo.17654442

The protocol describes candidate discriminants across platforms. This page presents them as proposed tests until independently executed by experimental groups.

Torsion-Balance Experiments

Scaling discriminants versus geometry and mass configuration.

Concept

Precision torsion balances are sensitive to small forces and allow controlled sweeps of geometry and configuration. QDL proposes that certain admissibility or closure constraints may translate into discriminant scaling patterns under these sweeps.

QDL-distinct discriminant: a pre-stated scaling relation or exponent constraint tied to ledger closure under geometry or configuration sweeps.
Standard comparison: conventional parameterization plus known systematics fits equally well without the QDL constraint.
Minimum viable test: a geometry or configuration sweep designed to separate the constrained QDL family from the unconstrained family.
Failure condition: the QDL-constrained scaling is not observed within uncertainty, or it cannot outperform the standard family without adding degrees of freedom.
Status and next steps

This is presented as proposed. The recommended pathway is sensitivity-first.

  • Identify an existing torsion dataset suitable for a geometry or configuration sweep.
  • Declare the standard and constrained model families.
  • Declare parameter budgets before fitting.
  • Run the sweep and report residual structure as the primary diagnostic.
  • Publish null results if the failure condition is met.

Roadmap reference: 10.5281/zenodo.17654442

NV-Center Frequency Shifts

Discriminants under controlled field and geometry sweeps.

Concept

NV centers provide a precision probe with controllable environments. QDL proposes that constrained combinations of fields, geometry, and constants may yield discriminant patterns under structured sweeps.

QDL-distinct discriminant: a pre-stated pattern in frequency shifts or residual structure across geometry and field sweeps consistent with a constrained ledger family.
Standard comparison: baseline spin-Hamiltonian family plus known corrections matches the sweep equally well without the constraint.
Minimum viable test: a sweep designed to isolate the constrained pattern while holding calibration degrees of freedom fixed.
Failure condition: no constrained pattern appears, or it appears only after adding ad hoc degrees of freedom that erase the discriminant.

Phenomenology anchor: 10.5281/zenodo.17803804

Status and next steps

This is presented as proposed physics testing. The executed benchmark below is methodological.

  • Executed benchmark: 10.5281/zenodo.18069870
  • Proposed discriminant: pre-register a sweep and constrained family.
  • Analysis discipline: keep benchmark and physics discriminant logically separate.
  • Reporting: publish residual-first results, including null results.

The existing benchmark shows workflow discipline. It does not claim a new NV-center physical effect.

Cavity Length-Frequency Scaling

Geometry and material sweeps as clean L-F probes.

Concept

Cavity resonators tightly couple length and frequency and are well-suited to controlled sweeps. QDL proposes discriminant scaling constraints in an L-F basis that can be tested against standard parameterizations.

QDL-distinct discriminant: a pre-stated scaling constraint across geometry or material sweeps that is not reducible to unit-balance alone.
Standard comparison: conventional fit family explains the sweep without the QDL constraint at equal or better residual adequacy.
Minimum viable test: a geometry sweep designed to stress the constraint while keeping calibration fixed.
Failure condition: no discriminant scaling is found, or the constrained family only succeeds after adding degrees of freedom that erase constraint meaning.

Teaching / lab anchor: 10.5281/zenodo.17663340

Status and next steps

This is presented as proposed physics testing. The executed benchmark is methodological.

  • Executed benchmark: 10.5281/zenodo.18076864
  • Proposed discriminant: pre-register sweep and constrained model family.
  • Minimum practical route: repurpose existing cavity apparatus into geometry or material sweeps.
  • Reporting discipline: publish residual plots, null tests, and parameter budgets.

The roadmap aims for low-barrier repurposing of existing cavity apparatus into discriminant sweeps.

Metamaterial Coherence and Dispersion

Dispersion-collapse candidates as discriminant signatures.

Concept

Metamaterials allow engineered dispersion and controlled effective parameters. QDL proposes that certain “locked” combinations may yield discriminant collapse behavior under designed sweeps.

QDL-distinct discriminant: a pre-stated dispersion-collapse or scaling-lock signature under a defined sweep.
Standard comparison: conventional effective-medium modeling reproduces the signature without the constraint.
Minimum viable test: a sweep designed to separate constrained collapse from flexible parameter fits.
Failure condition: no constrained collapse appears, or it is explainable only by unconstrained refits that destroy discriminant meaning.

Phenomenology anchor: 10.5281/zenodo.17803804

Status and next steps

This is presented as proposed. Suggested pathway:

  • Identify an existing metamaterial dataset with suitable sweep structure.
  • Declare constrained versus unconstrained model families.
  • Declare parameter budgets before fitting.
  • Report residual-first and treat collapse claims as falsifiable targets.
  • Publish null results if constrained collapse does not appear.

The goal is to avoid “it could be” narratives by making the discriminant and failure condition explicit.

FAQ: Scope and Falsifiability

Short answers to prevent over-interpretation.

Are any new physical effects claimed on this page?

No. The executed benchmarks are methodological and do not claim new effects. The tabletop discriminants are proposed tests with explicit failure conditions, presented as targets until independently executed.

What would count as a QDL-relevant hit?

A pre-stated discriminant scaling or pattern appears under controlled sweeps, and the standard comparison family fails to reproduce it within uncertainty without extra degrees of freedom that erase the discriminant.

What would count as a clear failure?

The QDL-constrained family fails across the sweep within uncertainty, or it only matches after adding ad hoc flexibility that nullifies the constraint’s meaning.

Are benchmarks evidence for QDL?

Benchmarks provide evidence of auditability and a residual-first adequacy workflow. They are not presented as standalone validation of the strongest cross-domain QDL physics claims.

Collaboration and Data

How experimental groups can engage, replicate, or refute.

For experimental groups

Collaboration is most valuable when it sharpens falsifiability and reduces interpretive wiggle room.

  • Replicate an executed benchmark using the same record and artifacts.
  • Propose a platform sweep with a pre-stated discriminant and failure condition.
  • Publish residual-first reporting, including null tests, regardless of outcome.
  • Declare standard and constrained model families before analysis.
  • Declare parameter budgets, calibration degrees of freedom, and sweep axes in advance.

The aim is to make success and failure equally publishable by design.

Contact and resources

Email: james.bourassa@qdlphysics.org

Additional materials:

For a proposed discriminant test, the most helpful first step is a short sensitivity note: what sweep is feasible, what noise floor dominates, and what failure condition is credible.