The score considers two main factors: the severity of the issues found (high, medium, or low) and your org size (the amount of metadata). High-severity issues undergo a logarithmic transformation that heavily penalizes their presence, while medium and low-severity issues are treated differently, with varying degrees of impact on the score. It also accounts for the ratio of number of issues which exist to total number of metadata, which help to make the score relevant to different org sizes.
