p
The bootstrap p-value, or null when there is no statistic or no null to read it against.
Computed as one plus the number of replicates at least as large as the observed, over one plus the number usable. The added one is not a rounding convenience. Without it a finite bootstrap can return zero, which asserts that the observed value is impossible under the null — a claim no finite sample of replicates can support. The floor this form imposes is the honest resolution limit of the number of replicates drawn.