SOLVED - apologies for posting.

The answer is simple in the end - I have one matching question component in the article and so the assessment component is only able to calculate the score based on one question - hence 0% or 100% as possible outcomes.

To ensure the result was as required I had to separate the matching component items into 3 separate matching components - which gives the correct feedback once the article is completed as the assessment is derived on a PER COMPONENT basis.

So a minor noob oversight - but we all learn by our mistakes right?