Gordon,
Water pilot, life, aviation, are probabilities.
MCAS failures must not occur in future operations - where ‘must not’ is the level of probability required for certification. This level is the internationally accepted value for equivalent safety - as low as reasonably achievable, but like all risk some small probability remains.
There is greater safety benefit in considering the wider system; thus tail trim can move to unacceptably high values - trim runaway, structural failure, … Then the important question is if pilot intervention will be able to reduce the risk; can the situation be recognised (any alerting), can the procedure be followed, is action physically possible, time available, level of knowledge / training required, … all balanced against risk - cost effectiveness.
A point can be reached where the situation is unacceptable, then the design must be changed; thence consideration of a ‘fully’ protected system, e.g. triplex FBW, but which also has a demonstrated manual backup.
It is this argument which rejects the need for specific MCAS training; although the trim runaway case probably still does. However, because the latter safety case depends pilot intervention, and with evidence from these accidents, there is the suggestion that pilot intervention is ineffective - it cannot be guaranteed.
You and others cite training - low professional standards. Alternatively, consider that the circumstances of the situation could be beyond pilot comprehension, an ambiguous emergency procedure, physically impossible to execute, … all beyond reasonable human ability. Also, if ‘reasonable’ is viewed as a probability, then this too cannot be established due ‘being human’, which even with training, cannot be sufficiently guaranteed to use in a system safety case.