Arena / planned experiment

Two outputs. One brief. Brands hidden.

The Arena will test which result better supports a defined research decision—without revealing the vendor, model, or visual polish until after evaluation.

OUTPUT / Aidentity withheld

Three themes, high agreement

Reports clear price sensitivity and feature comprehension. Evidence coverage is incomplete for one subgroup.

vs
OUTPUT / Bidentity withheld

Four themes, wider variance

Reports mixed price response and a novel trust barrier. One recommendation is unsupported by the source record.

Arena rules

01

Task before tool

Every match begins with a preregistered question and decision rule.

02

Blind before vote

Names, styling, and vendor language stay hidden during assessment.

03

Experts before crowds

Method-qualified reviewers evaluate evidence, not entertainment value.

04

Reveal every limitation

The result includes disagreement, uncertainty, and evaluation provenance.

Founding cohort

Volunteer a system for the first blinded protocol.

Participation requires configuration disclosure, reproducible outputs, and acceptance that unfavorable results remain publishable.

Enter the review queue