Scores are calculated using the following metric explained:

We have 5 binary (yes or no) questions on content. We also have three Likert scale ratings (1 - 7) on Tempo, Contextuality, and Redundancy. We want to rate content more important, but we still want to give credit for the other factors.

7 * 3 = 21 for best possible scores for Tempo, Contextuality, and Redundancy (Redundancy is best at 1 so this would be flipped). 100 - 21 = 79 available for content. 79 / 5 content questions  = 15.8 possible (rounded) per question.

This gives a perfect summary 100 points. A summary with no relevant content but all perfect scores for the other factors would get 21 points. Overall this would give summaries a maximum score of 100 down to a minimum score of 3.