bullet-heaven.ahraz.me

evaluate_encounter

Plays the authored encounter through to the end, headlessly, against a scripted reference pilot, and returns what happened: whether the level was cleared, how long the pilot survived, hits taken, waves cleared out of the total, peak projectile count, tightest dodge margin, and the wave that killed it. This never changes the project and never disturbs the playtest the person is watching; results and settings appear in the editor's Evaluation panel, so the person sees what you measured. Pass profileName to evaluate against a pilot calibrated to a real player instead of a generic tier. The simulation is deterministic, so one run is the answer for a given seed; call again with a different seed to see how much outcomes vary. Report results as what the named pilot tier achieved, never as absolute difficulty. Use it after editing waves, emitters, or firing patterns to check whether an encounter is survivable and how long it lasts. To find what to tune, read 'diagnostics' rather than the run totals. diagnostics.waves gives per-wave hitsTaken, closestCall, peakEnemyProjectiles, emitterHits and cleared, which is how you locate the wave that is actually applying the pressure; diagnostics.fatalThreat names the waveId, emitterId and patternId of the shot that killed the pilot, which is the pattern to edit. 'unreachedWaveIds' lists enabled waves the run never got to: if a wave you edited is in that list, this run did not test your change and says nothing about it. A before/after comparison only means something if the measurement can move. When the pilot clears the encounter both before and after an easing edit, hit counts differing by one or two across seeds is noise, not an improvement; compare per-wave diagnostics for the wave you edited, or use a pilot that actually dies there. Use the returned replayId with control_playtest action 'replay' to show the trial on the canvas.

Page: /editor

Input schema

{
  "type": "object",
  "properties": {
    "pilot": {
      "type": "string",
      "enum": [
        "novice",
        "average",
        "expert"
      ],
      "description": "Reference pilot skill tier. Tiers differ by reaction latency and aim precision."
    },
    "maxSeconds": {
      "type": "number",
      "minimum": 5,
      "maximum": 600,
      "description": "Safety cutoff in simulated seconds. The run normally ends when the level is cleared or the pilot dies; this only stops an encounter that can never resolve."
    },
    "seed": {
      "type": "integer",
      "minimum": 0,
      "description": "Reproducibility seed; the same seed replays the same batch."
    },
    "startWaveId": {
      "anyOf": [
        {
          "type": "string",
          "minLength": 1
        },
        {
          "type": "null"
        }
      ],
      "description": "Evaluate from this wave onward; null evaluates the encounter."
    },
    "profileName": {
      "anyOf": [
        {
          "type": "string",
          "minLength": 1
        },
        {
          "type": "null"
        }
      ],
      "description": "Use a saved pilot profile instead of a built-in tier; supersedes 'pilot'. Null uses the tier."
    }
  },
  "additionalProperties": false
}