Most mutations are neutral. Nachman and Crowell estimate around 3 deleterious mutations out of 175 per generation in humans (2000). Of those that have significant effect, most are harmful, but the fraction which are beneficial is higher than usually though. An experiment with E. coli found that about 1 in 150 newly arising mutations and 1 in 10 functional mutations are beneficial (Perfeito et al. 2007).
You missed the part about cause and effect directing the mutations coming into being. This means that they don't really even fit the definition of mutation, even though they are something not normally seen.
When you prove that everything has a cause, which you haven't yet I will agree with you.