Simulation Evidence and Branch Narrowing¶
This page summarizes the public role of CCT's simulation work without turning public documentation into bench recipes.
The purpose of the simulation layer is not to replace hardware. It is to make claims executable before hardware: define estimators, stress-test confounders, identify operating regions, reject weak branches, and write the question that a physical bench must decide.
What Simulations Have Already Done¶
| Lane | Public Result | What It Shows | Replication Surface |
|---|---|---|---|
| Measurement-regime / observer-mode sweep | Simulations and reduced models identify operating regions where record type and scaling become sensitive to observer mode under fixed-source controls. | The observer thesis can become an estimator test, not just ontology. | Publish the source/control assumptions, record-type metrics, bandwidth definition, nulls, and failure modes. |
| Field-control / structured field geometry | Simulation work narrowed the claim from broad controller superiority to a structured-geometry question under matched resources. | The simulation layer did real selection work by making the stronger claim smaller. | Publish the geometry-stability criterion, matched-resource comparison, drift/noise nulls, and narrowing rule. |
| Material-control / structured-vs-thermal benchmark | Simulations identify a structured-favorable operating wedge under energy and thermal controls. | Structured driving is worth physical exposure when the ledger and thermal nulls are declared. | Publish the task definition, thermal baseline, energy ledger, reversibility checks, and pass/fail thresholds. |
| Timing / metrology lane | Long-horizon timing and effective-metric work is held behind null-guarded metrology gates. | Speculative claims are converted into later-stage measurement questions rather than promoted directly. | Publish the null structure, held-out controls, and promotion gates. |
What This Means¶
The current simulation record supports a staged interpretation:
- Model results define what can be claimed inside bounded assumptions.
- Simulations turn selected claims into estimators, operating regions, and branch decisions.
- Bench protocols expose the engineering claims to instruments, materials, ledgers, drift, noise, and replication.
The most important public result is branch narrowing. Some lanes became stronger. Some became smaller. Some moved from anomaly-facing language into metrology and null-control language. Some became gated follow-ons. That is evidence of an evaluation program doing selection work.
What Is Publicly Reported¶
Public reporting should expose:
- the lane being tested;
- the claim status;
- the estimator or gauge family;
- the baseline and null controls;
- the operating-region or narrowing outcome;
- the hardware question that remains.
Public reporting does not need to expose bench recipes: specific materials, optical layouts, drive knobs, calibration recipes, or implementation notebooks can remain in lab protocol records.
Initial Public Replication Package¶
An initial public-safe replication package now exists at cct-public-replication/ in the project repository. It exposes the parts of the simulation/evaluation stack that can be checked without turning public documentation into a bench recipe:
- copied baby-theorem verifier code;
- copied toy-world simulation code with public smoke and reference configs;
- generic measurement-scaling estimators;
- generic steering-per-joule ledgers;
- uncertainty/null helpers;
- synthetic CSV examples for first reruns.
This package is evidence of rerunnable analysis machinery. It is not a hardware data release, and its examples are not bench data.
Artifact and Replication Status¶
This section answers a narrower review question: which simulation claims are currently reviewable as public artifacts, which are project-reported milestones, and what remains pending before stronger evidence labels are warranted.
| Result / lane | Claim level | Protocol status | Data status | Code status | Independent replication | Current public claim |
|---|---|---|---|---|---|---|
| E2-Prime horizon-style simulation | Pre-registered simulation milestone | Public summary; full protocol package pending publication decision. | Project-reported run result; public release package pending. | Project code path exists internally; public release package pending. | Not yet independently replicated. | Supports prediction workflow, operating-region selection, and simulation-to-protocol discipline. It does not confirm hardware behavior or universality. |
| Cross-domain measurement-scaling fits | Estimator portability / regime mapping | Method described publicly at the concept level. | Mixed source status: some fits use public external datasets, others remain project-reported summaries. | Initial public estimator package exists; source-data package remains lane-dependent. | Not yet independent as a CCT replication. | Supports gauge portability and regime-mapping hygiene, not a universal law claim. |
| Measurement-regime / observer-mode lane | Branch narrowing / bench aid | Public lane summary; bench-specific preregistration exists in lab records. | Public synthetic example exists; bench data package pending. | Initial public estimator/template package exists. | Hardware replication not yet applicable. | Defines a fixed-source observer-mode bench question and the nulls it must survive. |
| Field-control / structured field-geometry lane | Branch narrowing / bench aid | Public lane summary; bench-specific preregistration exists in lab records. | Public synthetic example exists; bench data package pending. | Initial public template package exists. | Hardware replication not yet applicable. | Narrows the claim to structured field geometry under matched resources before broader controller claims. |
| Material-control / structured-vs-thermal lane | Branch narrowing / bench aid | Public lane summary; bench-specific preregistration exists in lab records. | Public synthetic example exists; bench data package pending. | Initial public ledger/null package exists. | Hardware replication not yet applicable. | Narrows the hardware question to structured-vs-thermal task control under full ledger controls. |
| Timing / metrology lane | Later-gated metrology branch | Public lane summary; promotion gates remain later-stage. | Public exemplar package pending. | Generic null/uncertainty helper exists; lane-specific template pending. | Not yet applicable. | Converts long-horizon claims into null-guarded metrology questions rather than current evidence claims. |
Status terms:
- Public summary: enough information to understand the claim, lane, nulls, and current status.
- Initial public package: rerunnable code/templates or synthetic examples are available, while protected bench data and implementation details remain withheld.
- Public release package pending: the project has not yet exposed enough protocol/data/code for an outside reviewer to rerun the result from public artifacts.
- Independent replication: an outside group reruns the relevant analysis or hardware test without relying on project-only artifacts.
This ledger deliberately separates simulation milestone, public rerun package, hardware exposure, and independent replication. A result can be useful before all four exist, but its evidence label must stay tied to the highest stage it has actually reached.
How To Evaluate The Simulation Layer¶
A useful review should ask:
- Did the simulation make a claim more executable?
- Did it declare the estimator and baseline?
- Did it identify confounders and failure modes?
- Did it narrow or gate a branch rather than only add positive language?
- Did it produce a concrete hardware-facing question?
That is the standard. Simulations are not hardware confirmation, but they are not an afterthought either. In CCT, they are the translation layer between ontology, formal claims, and physical exposure.