Thanks! It's still a little vague (they only state "Each item, task, or simulation is worth one point, regardless of what you have to do to answer the question."), but since they may not use all 50 questions in the scoring (some may be test
questions), that could account for the odd scores (49 equally valued questions come close to matching the two scores). I guess just do better is the short answer.