Strategy Lifecycle

The Strategy Arena lifecycle is intentionally staged. Each stage has a different owner and a different evidence requirement.

​

Lifecycle

​

1. Submit

The builder creates a candidate strategy for a supported market family. In public beta, the candidate is sent through the Strategy Arena API. In the encrypted intent path, protected inputs are submitted through an on-chain intent record and processed by the private operator runtime.

​

2. Simulate

simulation-service resolves the official market dataset and execution profile, runs strategy components through spot-strategies, applies execution adjustment through amm-math, validates invariants, and emits reproducibility metadata. The public competition profile uses server-controlled datasets, costs, and scoring windows so candidates are comparable.

​

3. Evaluate

ranking-evaluation computes scenario utility, benchmark-adjusted score, CapitalScore diagnostics, and eligibility output. Evaluation failures are typed and explicit.

​

4. Score

The beta score is the public display score returned by the evaluation endpoint: $$\mathrm{score} = 100 \times \frac{\mathrm{strategyUtility}}{\max(\mathrm{benchmarkUtility}, 0.0001)}$$ CapitalScore is a stricter diagnostic for robustness across scenarios. It is not the same as contest scoring.

​

5. Evidence

Evidence is the set of replayable outputs that let another reader understand why a candidate ranked where it did:

• simulation output hash,
• benchmark utility,
• utility breakdown,
• eligibility reason codes,
• report hash,
• dataset and registry metadata.

​

6. Eligibility

Eligibility means a strategy passed configured gates. It is not a capital commitment.

​

7. Guarded path

Future guarded deployment should follow vault constraints for assets, size, execution permissions, and operational risk. The arena evidence is the input to that decision, not a bypass around it.

​

Related pages