Scoring
When scoring the exam, the score for each of the questions is given only for fully correct answers. This means that the question will be deemed as correctly answered only when the correct answer or all of the correct answers are selected. A single incorrect answer in a multiple-response question invalidates the entire question.
The exam pass/fail grade is independent of the number of correct answers, and is calculated based on the score that's achieved out of 1,000. Each question carries a different number of points since each question is considered to be of a different level of difficulty. This means that even if you were to answer 70% of the questions correctly, you can get scored at 60% or lower if all the questions that were answered correctly carried a lower score. I encourage you to answer all the questions during the exam, as even a few unanswered questions can affect the scoring dramatically.
AWS does not disclose the passing score publicly, but you can discern the passing score by examining the pass/fail scores of the community of test takers that have posted their pass/fail scores on various forums. The official stance of AWS is that the certification passing score is heavily dependent on the statistical analysis of multiple metrics that AWS receives from several sources.
Historically, the community has determined that the passing scores of different AWS exams has been set to between 650 and 750 points out of a possible 1,000. The benefit of this variable passing score is that the more difficult the exam, the lower the passing score will usually be, although depending on being able to pass the exam due to a low threshold is not recommended. Once the exam is released, the metrics that are received from the exam takers are taken into account and the passing score of a particular AWS exam is adjusted periodically. This implies that the actual passing score at the time when you attempt the exam might not reflect the definition set out in this chapter. Because of this, I recommend that each exam taker has the confidence that they can answer at least 80% or more mock questions correctly before attempting the real thing.