One challenge of XP agile method in software development is this method's underestimation over software quality attribution; these attributes are of main indices of software architecture. In this study, a method is represented for responding this challenge on the bases of probability theory. In this method, firstly, the rating matrix is structured on the bases of quality attributions an architecture solutions. Each element of this matrix shows a rating for every solution and the ratings will be initialized through analytic hierarchy process (AHP). Via quality weight implement on above matrix the rating vector is created whose each element represents the rate of each solution in reaching those quality attributes in rating matrix. Because rating vector follows normal probable distribution, its elements' rating probable density is mostly gathered around the mean.
In this study, the probable density of architecture will be defined, the appropriateness of a solution in comparison to other solutions will be evaluatedor and the respond to XP challenge will be done easily in order to reach architecture solutions in the frame of rating vector. A complex of data are gathered and the results are compared in an experimental method in order to investigate the represented method of this paper.