Environments & rehearsal · Replay
What actually happened
A run is a record, not a log line. It carries what it ran, so two of them can be compared.
What a run carries
| Scenario | What it ran, by name |
| Site | Where |
| Seed | The number that makes it repeatable |
| Environment | Live, Dome Lite or AirSim |
| Window | When it started and when it ended |
| Outcome | What was stopped, what got through, what it cost |
The environment is on that list for a specific reason: it stops you comparing a Dome Lite run against an AirSim one and drawing a conclusion. Two runs are comparable when they ran the same scenario, at the same site, from the same seed, in the same environment. Anything else is a different question.
Every event carries its run
Each event produced during a run is stamped with that run's id. Nothing has to be reconstructed from timestamps afterwards, and two overlapping exercises cannot contaminate each other's record.
That stamp is what replay reads.
Replay
Open a run and you get its window back: the fused picture as it was, the alerts as they opened, the plans as they were proposed and what happened to each.
Replay reads the recorded events, not a re-run of the simulation. That distinction is the whole reliability of it. Re-running the wire would give you something close to the original and not identical, because batch arrival order depends on how busy the host was. Reading the record gives you exactly what the system saw and exactly what it decided.
Comparing two runs
The workflow rehearsal exists for:
- Run a scenario. Note the outcome.
- Change one thing. A posture, a rule, a sensor position, a planning profile.
- Run the same seed again.
- Compare.
Because the seed fixes the raid, the difference in outcome came from your change rather than from a different fight. That is what turns a demo into evidence.
Statuses
| Status | Means |
|---|---|
| Running | Events are still being appended |
| Complete | Stopped cleanly |
| Aborted | Superseded by another run before it was stopped |
An aborted run keeps everything it recorded up to that point. It is a partial record rather than a discarded one.
Ask the Copilot to narrate one
The Copilot can read a finished run and describe it: what arrived, when the picture firmed up, which alerts opened, what was approved. Useful for writing up an exercise without scrubbing through it.
See The C2 Copilot.
Where to go next
- Configuring a simulation to change one thing and run it again.
- Solver configuration if the thing you want to change is how plans are scored.