Skip to main content
The beta leaderboard is a presentation layer over evaluation responses.
Use evaluation.data.score as the beta row order.

​
Beta rule

Sort by:
Use evaluation.data.score as the beta leaderboard row order.
The router does not currently expose a server-owned cross-submission leaderboard endpoint. The UI or thin client assembles rows locally from simulation and evaluation calls.

​
What every row should show

A useful row should show more than one number:
FieldWhy it matters
scorecurrent public beta sort key
scenarioResult.annualizedMeanReturnreturn contribution
scenarioResult.maxDrawdownpeak-to-trough risk
scenarioResult.cvartail-loss diagnostic
scenarioResult.turnoverchurn and friction proxy
scenarioResult.executionCostmodeled cost drag
benchmarkUtilitycomparison denominator
eligibility.eligiblegate result
capitalScore.capitalScorestricter robustness diagnostic

​
Evidence artifacts

Evidence should be reproducible. Artifact mode should preserve:

Inputs

Input hashes, config hashes, dataset or snapshot hashes, and strategy registry snapshot hash when the menu affects outcomes.

Outputs

Output hashes, report hashes, and dependency/package version information.
If a report cannot be replayed from committed inputs and referenced artifacts, it is not strong evidence.

​
Interpreting eligibility

Eligibility means the strategy passed the configured gates for a review or next step. It does not mean the strategy receives capital. Use this line consistently:
Evidence creates eligibility. It does not guarantee capital.

Scoring and benchmarks

Understand score, benchmark utility, CapitalScore, and ContestScore.

Guarded vaults

See how future vault review should consume evidence.

Public settlement

Understand what can become public at execution time.