some time between 1.06 and now they have sorted out the non-random random - now you do get a different initial value starting at each trig, it used to commence from zero every time, previously the workaround was to reverse the LFO, that always started with a different value