What proportion of vaccine trials are successful usually? Could the probability of efficacy be estimated and used to judge if the study goes ahead?

There needs to be a reasonable amount of risk taken in research, otherwise we will find it impossible to make progress.