Adaptive Experimentation Agent

Dell · Analyst Workbench · v0.1

BUILD: wb-2026-06-02-v11
API: OFF
BENCH: PEND

Experiment Workbench

1) Console specs & variants → 2) Experiment & telemetry → 3) Validation & approval after you run.

DELL · ANALYST CONSOLE

API is offline — use Refresh in the top bar, then run analysis.

Step 1

Console specs & variants

Each card is a different test variant (arm) in the benchmark environment — control plus three product profiles. Pick the profile that frames how we interpret difficulty, rewards, and onboarding in step 2.

Variants in the benchmark

A full multi-arm run (e.g. exp_sanity_001_calibrated) compares all four arms side by side. The cards below are the same variant types you assign when building a custom experiment.

  • Baseline (control)(control)Standard difficulty, light onboarding, no aggressive promos.
  • Fast progression(fast_progression)Higher rewards and faster progression for engagement tests.
  • High challenge(high_challenge)Harder curve for skilled players; minimal hand-holding.
  • Guided onboarding(guided_onboarding)Lower friction and guided flows for new-user retention.

Select your framing profile for step 2

Selected profile

Baseline (control)

difficulty
0.0
reward rate
1.0×
matchmaking
12s
onboarding
light
promotions
none

Experiment source

Step 2

Experiment outline

Select which test to analyze and run telemetry. Console profile from step 1: Baseline (control). Preview below lists the variants in that experiment.

Loading experiment details…

After you run analysis, validation and recommendations appear here (step 3).

New here? Read how the system works