FIELD MANUAL · SYSTEM

Build Testing Method

DIRECT ANSWER

Build Testing Method: current answer

An interpretable build test records the patch, encounter, loadout and target conditions, then changes one variable at a time. This is an editorial testing method: it helps separate documented balance changes from visual changes and unmeasured impressions.

Quick facts

  • AG-14 changes Vortex Shell damage and Thunderbolt Arrow area, so those properties deserve separate comparisons.
  • AG-12 changes Guided Bolts projectile effects and Lethal Purge HUD effects; a visual change alone is not a measured damage improvement.
  • ST-24 changes when failed captures resolve. A quicker failed attempt is not evidence of a higher overall success rate.

Practical workflow — editorial guidance

State one question before the run, such as whether the same target takes more damage or whether a spaced group is affected more often.

Record weapon and Core names, other equipped effects, target type, difficulty and relevant positioning. Keep the baseline conditions as similar as possible.

Change one Core or setting, repeat the comparison, and write down both successful and unsuccessful outcomes rather than only the best example.

Keep sample size and uncertainty with the result. Use the planner as a notebook; it does not calculate unverified damage formulas or establish capture probabilities.

Evidence limits

This method has not itself been validated by live-game experiments here. Observations remain conditional, and no minimum sample count guarantees a statistically reliable result.

Patch history

Source ledger

Last checked 2026-09-26. Official-source verification is not independent live-game testing. Practical workflows are editorial guidance.