# [2019.10.12] Probabilistic Associator Didn't Work

Unfortunately, the probabilistic associator, that I tried to use,
didn't work for the case. It is zero for a uniformly random
distribution of items in a Cayley table, which means that such a
table is associative in a sense. It probably is, but now the training
process tends to converge somewhere near this distribution, which is
uninformative. I should try something else.