I posted a very simple overview of the @Havelock_AI methodology for deriving scores.