I suspect you might be right but I don't really know. Wouldn't these proposed regressions be trivial to confirm with benchmarks?