Documentation

Everything the benchmarks do, written down. Every protocol, every formula, every coverage figure, every version pin — public, and versioned by the evaluation run that used it.

That is not a nicety. A benchmark asks vendors to be measured in public, and asks clubs to act on the result. Neither is reasonable unless both can read exactly how a number was produced.

What is here

Start here explains what the Arena is and how to take part. Getting started for the shape of it and how the live benchmarks differ; Quickstart to register, enter a benchmark and make a first submission with curl.

Protocol covers the rules that apply across every benchmark — how a participant enters, when an answer is fixed, what declining and arriving late cost, what is published — in the evaluation protocol, and how to read a leaderboard, which is the page to send anyone about to act on a number.

Tasks has one section per benchmark. Two are live:

• Transfers — its overview, the submission contract, scoring, data and coverage, and the version archive and changelog behind every published score.
• Game Value, team board — its overview, onboarding, the delivery contract, scoring, and data and coverage.

Every live section has the same core — what is asked, the contract, how it is scored, and the data it rests on — so integrating a second benchmark means reading a familiar layout rather than learning a new one. Each then adds what its task needs: Transfers its version archive; Game Value an onboarding path, because taking part in it is a weekly integration rather than one submission.

Sections for benchmarks still in design — Player Valuation, and Game Value's player and element boards — publish their protocol before any score exists.

Reference is the shared vocabulary. The glossary defines every term a board uses — the shared terms first, then each benchmark's own — so a term cannot mean two things in two places. The role taxonomy and metric vocabulary are the player vocabulary Transfers is written in.

Two rules that govern these pages

A version page is written once and never edited. Corrections are published as a new version with a changelog entry naming what they replaced. That is what lets a score from November still be readable against the rules that produced it. Transfers publishes a page for every version its scores name. Game Value names its contract version on every snapshot, and does not yet publish a page per version.

What the API publishes is not retyped here. Wherever a page shows something a participant also fetches from the API — Transfers' metric masks, band grids and sample payloads — it is rendered from that same specification, so the page you are reading and the endpoint a vendor calls cannot say different things.

Status

Two benchmarks are live: Transfers, and Game Value's team board. Pages appear as their content is settled rather than as placeholders, so an empty section means the decision is still open, not that the page is missing.