Is Hi-Opt II way better than halves

I have been using the halves counting system for years. I was just messing around running simulations with CVData and consistently got a 30% or more better return using the Hi-Opt II system instead of halves. I was really surprised! I had thought that halves was one of the most effective systems. Is this news to anyone else or is there something wrong with my simulation. I will have to learn this system if the difference is that great