# E2E W16.5 Closure Confirmation

**Date**: 2026-05-29  
**Scope**: Anderson BOUNDARY physical perturbation landing loop  
**Status**: CONSTRAINT / NO PROMOTION

## Question

Does the intermediate Anderson window `W=16.5` close as a real landing
coordinate, or only as a transient artifact of the first perturbation loop?

## Runs

### R1 - narrow isotropic replay

Artifact:
`tools/data/landing/boundary_anderson_w165_closure_confirm_20260529_r1.json`

- variants: 15
- sizes: 5,6,7
- W16.5 closures: 2
- closure coordinates: `onsite_tilt=-0.065`, `onsite_tilt=-0.060`
- W20 preservation: 1.0

Reading: the first closure map did not collapse. The closure is concentrated
near a negative onsite tilt band.

### R2 - stricter isotropic confirmation

Artifact:
`tools/data/landing/boundary_anderson_w165_closure_confirm_20260529_r2.json`

- variants: 15
- sizes: 5,6,7,8
- reps: 10
- scramble trials: 160
- W16.5 closures: 1
- closure coordinate: `onsite_tilt=-0.0625`
- W20 preservation: 1.0

Reading: adding size 8 and stronger null pressure does not remove the closure,
but narrows it to a precise isotropic coordinate.

### R3 - independent seeds plus mild anisotropy

Artifact:
`tools/data/landing/boundary_anderson_w165_closure_confirm_20260529_r3.json`

- variants: 11
- sizes: 5,6,7,8
- base seeds: `2026052911,2026052912,2026052913`
- anisotropic hopping: `1.04,0.98,0.98`
- W16.5 closures: 0
- W16.5 global-only: 1 at `onsite_tilt=-0.066`, anisotropic family
- W20 preservation: 0.909090909

Reading: the full local/global closure does not survive this seed/family
change. A global-only residue remains in the same negative tilt neighborhood,
and W20 loses one case under anisotropic stress.

## Result

The useful result is not a promoted physics claim.

The Lab has isolated a narrow intermediate landing coordinate:

```text
Anderson W16.5 closes under isotropic negative onsite tilt near -0.0625,
while W20 remains the robust local_global reference.
```

This coordinate is real enough to survive the stricter isotropic replay with
size 8, but not robust enough to survive independent seeds plus mild
anisotropy as a full family.

## Interpretation

- `W20` remains the reference pole.
- `W16.5` is a sensitive intermediate channel.
- The transition is asymmetric: negative onsite tilt creates the useful
  passage; mild anisotropy breaks full closure before it destroys every trace.
- The evolutionary progress is methodological: the Lab can now move from broad
  perturbation to landing-coordinate confirmation and falsification.

## Boundary

No law promotion. No public discovery claim.

This is a verifier result:

- candidate: narrow isotropic landing coordinate;
- evidence: repeated local/global closure under stricter isotropic replay;
- falsifier: independent seed/family stress prevents promotion;
- next useful test: map the basin around `-0.0625` with controlled seed families
  and a separate anisotropy axis, then decide whether the coordinate is a
  stable basin or only a gate crossing.

## Verification

`python3 tools/lab_surface_invariant_check.py`

Result:

```text
lab_surface_invariant_check: PASS
failures: []
warnings: []
```
