~TRAINING LAB

Pick a build, choose a scenario, evolve. The best neural weights survive.

TRAINING LAB SETUP
configure scenario, build pairing, and runtime controls before launch
1

TRAINING MODE

LOCOMOTION: Start here! Teaches route stability and anti-stall motion. · Start here if your agent wall-rides.

Auto-fit setup for LOCOMOTION: mode + objective + environment + runtime.
2

SELECT AGENT

Pick a build

BASE BUILD (PARENT)

LOADING...
3

CONFIGURE & RUN

GUIDED GOALS
Reduce wall-riding and stall loops before advanced combat.
scenario: LOCOMOTIONmode: EVOLUTIONobjective: mobilityenv: cover_parkour
Critical spatial sensors must pass.
Keep stall ticks under 120.
Keep wall collisions under 140.
ONE-CLICK PIPELINES

Anti-meta guard is ON: save step re-validates top candidates across varied opponents/modifiers and keeps the most generalist performer.

12
4
0.30
TRAINING INTEGRITY CHECKS2 blocking · 1 advisories
Base build selectedFAIL
Pick the build you want to improve.
Opponent configuredPASS
Scenario uses curriculum opponent setup.
Runtime budget is safePASS
Complexity 240/30000 is within STANDARD budget.
Critical sensor groups presentFAIL
Missing: Body-space awareness.
Recommended sensor depthWARN
Weak coverage: Resource and timing awareness, Motion prediction awareness.
Mode aligns with scenarioPASS
EVOLUTION is suitable for LOCOMOTION.
Mode sample volumePASS
Sample volume is in a healthy range.
Training depth expectationPASS
Good depth for an early-stage build.
Mutator profile fitPASS
lab_standard matches this scenario's intent.
SENSOR COVERAGE
Core combat awareness0/3
Projectile awareness0/4
Pickup control awareness0/3
Hazard discipline awareness0/5
Body-space awareness0/7
CRITICAL
Resource and timing awareness0/3
RECOMMENDED
Ability timing awareness0/9
Motion prediction awareness0/3
RECOMMENDED
Multi-target contest awareness0/5
LAB COACHBLOCKED
Select the build you want to train.
Recommended goal stack is selected but not applied.
RUN MANIFEST: standard · seed 424242 · profile balanced · env lab_standard · 12 pop · 4 gens · 5 matches/eval · 600 tick cap · complexity 240/30000
~36s · 240 matches
Select a build
TRAINING RESULTS + TACTICAL ANALYSIS
live telemetry, replay comparison, and tactical before/after deltas
@ PROFILE