The metric is in the eye of the beholder. ;)

I'd store all OK stringifications in a hash and fail if it's not included.

As an idea:

If you have different levels of nice-to-have probably try to SKIP / TODO those depending on a global threshold $NICE_LEVEL . Changing it should give you an indication of "more OK". (Disclaimer: never tried this)

If you log the level for each block, you can try to calculate a rational number metric afterwards.

Please clarify if I misunderstood your question.

PS: I'm so proud of me to have avoided any pun referencing "50 shades of ..." Oops! ;)