You need to run a quantitative test. Not "it felt different". Those are good things to note, but that isn't a quantitative result. How many rotations did it make, how many times did you test it, what kind of output is it creating. You have to clearly define a test and then live by the results. If it doesn't work by all means go back to the math and figure out where you went wrong, but don't go on a hunch of if "I just change..." or "If I could..." prove the concept with math. If you can't do that then you are taking the internet at face value... the world might as well be flat, the moon is made of cheese, and 9/11 was an inside job.