inalpha

E1 candidate evaluation references

• —run_cv_backtest
• —deflated_sharpe_ratio
• —probability_of_backtest_overfitting

Three candidate slots were requested and three distinct candidate sources were generated.

All three distinct candidates reached the real E1 candidate evaluation path and completed successfully.

The candidate evaluator exposed no parameter for evolutionary trial count, candidate count, search count, or equivalent multiple-testing count.

No evolutionary trial-count field appeared in any of the three candidate evaluation snapshots.

No CV, DSR, or PBO-related field appeared in any of the three candidate evaluation snapshots.

The E1 evolution generation and slot evaluation modules contained no direct references to run_cv_backtest, deflated_sharpe_ratio, probability_of_backtest_overfitting, or evolutionary trial-count parameters.

The paper backtest module contains separate CV and Deflated Sharpe machinery.

Therefore this run found no evidence that the E1 evolutionary candidate population is passed into the CV/DSR/PBO machinery at the candidate-evaluation boundary tested.

This does not establish that Inalpha has no multiple-testing correction elsewhere in its broader research system.

The deterministic mutation boundary did not invoke an LLM provider.

The dataset was synthetic, so the run provides execution and architectural evidence rather than strategy-quality or real-market evidence.