Frozen result
What was computed
Nonlinear-adjacent linear CGYRO test of whether the negative-triangularity quiet-basin stability advantage survives a collisionality scan (ν* ×0.25–4). It does: NT growth-rate suppression vs PT holds 56–66% across the ladder and never degrades — the NT benefit is not a collisionality artifact.
Linear γ across the ν* ladder
| ν* (×nom) | NT γ | PT γ | NT suppression vs PT |
|---|---|---|---|
| ×0.25 | 0.0528 | 0.1204 | 56.1% |
| ×0.5 | 0.0494 | 0.1180 | 58.1% |
| ×1 | 0.0439 | 0.1099 | 60.1% |
| ×2 | 0.0329 | 0.0972 | 66.2% |
| ×4 | 0.0294 | 0.0809 | 63.7% |
Both branches are linearly unstable at kᵧ=0.30 (γ>0, electron-diamagnetic direction) and stabilize monotonically as ν* rises. NT γ sits far below PT γ at every ν*, and the advantage does not degrade — it strengthens toward higher collisionality. Honest caveats: this is a single-kᵧ linear growth-rate proxy (not a full spectrum or nonlinear flux); NT modes are near-marginal and carry a few-% convergence band (NT cases re-run to MAX_TIME=300), but the NT-below-PT ordering is unambiguous.
Reproduce this
# Reproduce BR-L2-A13 against the frozen record pip install kronos-toolkit # github.com/KronosFE/kronos-toolkit python -c "from kronos_toolkit.verify import run_regression; run_regression()" # BR-L2-A13 is anchored in the 2026-08-27 freeze spine (verdict: reproduced).