My recollection is that 3x3 is a goal to be 3x faster than Ruby 2.0–presumably many of those gains have already been realized, so best not to depend on tripling _current_ performance.