Since we’re dealing with BLAST searches here, I’m assuming that the search space is all possible sequences in a genome of the same size. (I’ve been assuming we’re using nucleotide sequence, but the argument is the same if instead we use amino acid sequence for proteins.)