Level complet · 3 run(s) · 6000 ticks (30.0 s of emulated time) · generated 2026-11-14 14:02:35
GLOBAL VERDICT: ZERO VIOLATION — 3 runs, the handover bound holds everywhere (level complet).
Acceptance criterion (spec §13)
Measurement
Statut
Zero invariant violation tolerated
0 violation(s), 0 malformed line(s)
OK
Handover after a death: 0 ticks without a fresh order (certified second)
worst measured gap: 0 tick(s)
OK
Deviation from the twin without fault ≤ delta (0.02)
worst measured deviation: 0
OK
Campaign scale (spec 7.8): real cuts
deaths 1/2000, freezes 1/10000, cuts 0/5000
FAIL
Mutants M1–M8 flagged
no mutant bench result supplied (--mutants)
not measured
Per fault class
Fault class
Runs
Violations
Worst gap (ticks)
Bound
Status
death
1
0
0
0 (certified second)
OK
freeze
1
0
0
0 (certified second)
OK
nominal
1
0
0
0 (certified second)
OK
Worst cases
death death
Worst gap 0 tick(s) at tick 0; worst deviation from the twin 0; command served by: fallback=1, holder=1997, shadow=2. Handovers: tick 2 → A (mandate 1, 0 tick(s)), tick 502 → B (mandate 2, 1 tick(s)).
freeze freeze
Worst gap 0 tick(s) at tick 0; worst deviation from the twin 0; command served by: fallback=1, holder=1997, shadow=2. Handovers: tick 2 → A (mandate 1, 0 tick(s)), tick 702 → B (mandate 2, 1 tick(s)).
nominal nominal
Worst gap 0 tick(s) at tick 0; worst deviation from the twin 0; command served by: fallback=1, holder=1999. Handovers: tick 2 → A (mandate 1, 0 tick(s)).
Runs
Run
Class
Ticks
Violations
Max gap
Max deviation
Status
death
death
2000
0
0
0
OK
freeze
freeze
2000
0
0
0
OK
nominal
nominal
2000
0
0
0
OK
Emulated timings, measured worst cases — never proven worst cases. Single-fault hypothesis: one fault at a time until redundancy is back. The actuator is a declared single point of failure; its arbitration core is tiny and exhaustively explored. See docs/SAFETY.md §7 and docs/RESULTS.md.