Probability distributions of success and failure runs related to Bernoulli and Markov dependent trials