Effectiveness of explanation facilities for intelligent systems