Whenever a user answers to a question during a test quiz or an exam, the system estimates his level of knowledge on the basis of the question's difficulty and on the correctness of the user's answer.
The difficulty of questions is adaptively selected by the system to incrementally challenge the user among different tests.
Each question tests some specific skills. The system also estimates the user's knowledge on correlated skills.
For instance, if the user is tested on Amazon S3, the system infers his knowledge also on storage. Being an estimate, when the user will be tested on other services in the storage area - e.g., Amazon Glacier - the system will tune the estimated score on storage over time.
At the end of quiz or exam you will see your score result update for each category and the updated leaderboard.
Consequently, the user skill score might decrease or increase over time in accordance to the correctness of user's answers.