When two manuscripts share a rare characteristic our first reaction is to assume a common origin. This does not always work, but it is a better starting point than some others.
For instance, take co-blogger Michael Holmes's text of 1 Clemens ("To the Corinthians") in his Apostolic Fathers. On page 28 of my 1999 edition the text has a misspelling in the Greek text μελαγοπρεπες which should be μεγαλοπρεπες (the -λ- and -γ- have swapped places). Lightfoots's printed edition does not have this error, and neither has the TLG online, based on Jaubert's text.
For my Greek teaching I use the opening paragraphs of 1 Clemens. And at the time when I prepared my handout I was still working from the now defunct CD-ROM version of the TLG. When one of those clever students (yes, the same one as here) pointed out that μελαγοπρεπες in my handout looked like a misprint, I checked Holmes, found the same spelling, and initially thought that Holmes and me could not both be wrong. Till I realised that I had simply cut and pasted from the CD-ROM text, which might have been the same thing Michael had been doing. And rather than asking him (we are scholars after all, we don't talk), I checked the preface and found Holmes's acknowledgement of the technical assistance provided by the TLG, signed off in 1992, way before the online edition.
Agreement in error can point to common origin.
[This is the relevant word in Codex Alexandrinus, difficult to read, but in the correct spelling]