Yes, I am also confused. The OP "may" be asking how much information is needed to confirm the identification of a specific stamp variety. I believe their example is if they find a hair line or scratch, can they then say it is stamp A.
Some catalogs show known varieties. Those catalogs, by word or detailed photo, will show you which specific feature makes it that known variety. Therefore, for example, if the description is "scratch from upper left to lower right", you would probably assume you have the same, as long as you have a scratch from upper left to lower right of the exact same issue, color, denomination, watermark, etc. Better yet, if they show a picture and your scratch EXACTLY matches the picture, then you are probably good. (But read the wording as they may also say that there is another feature somewhere else, and you need both of these to be present.)
If the Ink from a postmark obscures part of the variety, then it matters how much. Every case would be different. However, we used to have a poster here that thought they had super software that could look through postmarks at the features underneath. I don't think they got much agreement from this group.
I have no idea what you are asking in point number 3. Perhaps you are asking how to certify such a variety? Send it in. Don't just state on
ebay that this is a newly discovered variety and give it some catchy name.
NOW, another thought. That is that the OP is asking if a special feature points to a new, unknown variety. I believe the consensus would be that you need to have at least one other confirming copy with the EXACT same feature to have a chance of proving a consistent, new variety.
At least this is how I am reading the OP's questions.