---
title: "Blind rank test — is the order real? · Undominated.ai"
canonical: https://undominated.ai/blind-rank/
description: "Order two anonymised models, then see their published confidence intervals. On this board 94% of adjacent steps are inside the measurement error, so the integer rank is often invented."
---

# Blind rank test — is the order real? · Undominated.ai

> Order two anonymised models, then see their published confidence intervals. On this board 94% of adjacent steps are inside the measurement error, so the integer rank is often invented.

# Blind rank test

Two models from the board with their names and scores hidden. You get the one thing a buyer sees first: the price. Pick the one you think scores higher, then see whether the measurement can tell them apart at all.

**94%** of the 144 adjacent steps on this board are inside the published error band. Computed from the catalogue on this page load, not written down.

Pair 1 of 8

Which of these two scores higher?

 A $1.20 /M out name and score hidden vs B $10.00 /M out name and score hidden

Scores are LMArena Elo, used under CC BY 4.0 from the official dataset. Nothing you click is recorded.

## Continue your investigation

 - [Read uncertainty first](/significance/)
- [Test a shortlist](/compare/)
- [Inspect the experiment](/methodology/)
- [Compare intervals](/ranges/)