If you have the chromosome/segment details and you wish to check if the "Johns" are one and the same person, I'd have thought a 1to1 on Gedmatch/Genesis would have provided pretty conclusive evidence? Scales can be expanded to look even more closely at the graphical results and Genesis Tier1 Q match (experimental) might be worth a look. What about "X" match. Perhaps clutching at straws; how do they relate to other "tested" relatives... but I expect you've covered all bases.
I don't have DNA for the Johns.....All I know is that I have 3 people who share a number of cM on the same segment, and that I have a confirmed/known relationship between A & B. However, the predicated relationships between A & C, and B & C, for the likely hoods with the highest percentages don't align together based on that known relationship. What I was looking for was something that would therefore make it easier to work through all the possible combinations of the lower level relationships and say these ones are the best fits. Using what I know of my tree and the cousins tree, those relationships could then be tested to see if John is a likely candidate as the match so we know where to focus on with the paper trails or if it points in the direction that he cant be.
Richard