Questions about survival prediction

Hi, thanks for your great work. I wonder for these metrics in EvalSurv, how to use them evalaute model prediction performance directly.

I think only these three metrics: Brier Score, Negative Binomial Log-Likelihood, and Administrative versions of Brier Score and NBLL, these three metrics' higher values mean worse model.

Is my understanding correct? Thanks.

Start with EvalSurv and the definitions of Brier Score, NBLL, and their administrative variants mentioned in the issue. Clarify how each metric should be interpreted for survival-model predictions, and document whether higher values indicate worse performance.

Written by the indexing model from the issue text.