Just my two cents, but you might want to update your naming conventions for the types of missingness.
You use completely random, at random, and not at random. The first is truly random, the second is non-random and associated with known variables, and the third is non-random and associated with unknown, or latent variables. It might clearer if you call these:
missing at random
missing not at random and observed (or explainable)
missing not at random and unobserved (or not explainable or latent)