That sounds like it's a bad benchmark, then? I mean, it's still interesting, but a synthetic benchmark that fails to map to real use is... I guess actually negative value since it's misleading.