Comments on the sequential probability ratio testing methods

In this paper the classical sequential probability ratio testing method (SPRT) is reconsidered. Every individual boundary crossing event of the SPRT is regarded as a new piece of evidence about the problem under hypothesis testing. The Bayes method is applied for belief updating, i.e. integrating these individual decisions. The procedure is recommended to use when the user 1) would like to be informed about the tested hypothesis continuously and 2) would like to achieve his final conclusion with high confidence level.