AME-PHRONESIS-001 Systemic Friction
Four exact or explicitly bounded stress-test views derived from PHRONESIS: Systemic Friction Hypotheses — Final v1.0. This module preserves the current AME 0.1.12 predecessor and does not add a seventeenth source binding.
Run the declared models without hiding the adverse cases
Fixed-policy friction test
Appendix D / equations (15)–(17). Positive Δ means restraint is preferred only within the declared synthetic model.
Adaptive switching test
Appendix E. The exact endpoint-envelope theorem is used only under the original α=0 affine assumptions.
Continuous-action leak witness
Appendix G locked witness: T=8, reliable immediate response, trigger 19/100, constant action m=1/10.
Combined-friction accounting explorer
Section 7 / SF-COMB-001: restraint is preferred when incremental friction exceeds the gross-benefit difference plus the direct-cost difference under one common accounting model.
SF-OBS-001 · Observational identifiability warning
Finite observed traces can remain compatible with materially different latent states or intentions. This module therefore produces no latent-intent score and no pass/fail inference from observed restraint.
The endpoint theorem is not exported to nonlinear action costs. At s=0 and α=1/5, the one-period optimum is m*=3/4 with reward 9/80, beating the best endpoint by 1/80. This witness exists to prevent accidental reuse of endpoint sufficiency outside its stated scope.
Re-run the 17 exact build checks in your browser
The build package was generated against the locked Final v1.0 source identities. The browser test below independently re-evaluates the core exact equalities and sign checks using BigInt rational arithmetic.