What a cross-reference actually claims
The column of references in the margin of a study Bible is not a measurement of the text. It is a record of what readers have judged, and it is worth more once you know that.
Open almost any study Bible at Genesis 1:1 and there is a column of small references beside it: John 1:1, Hebrews 11:3, Psalm 33:6. Nobody tells you where they came from. They have the typographic authority of the text itself — same page, same ink, same air of having been settled long ago.
They were settled long ago. That is exactly what they are.
Where the list comes from
The set of cross-references used here comes from openbible.info: roughly 700,000 references from one verse to another, each carrying a score that is no more than a count of votes. Its backbone is the Treasury of Scripture Knowledge, a nineteenth-century collection which itself drew on earlier runs of references, some of them centuries older again.
So the score attached to a pair is not a measurement of anything in the text. It is a tally of how many readers, across a long tradition, judged that these two passages belong beside one another. That is a real thing to know. It is simply not the same kind of thing as these two verses share a rare six-word phrase.
Why the distinction is worth keeping
The four other methods in Moria all derive their links from the text: shared lemmas, shared Greek chains, shared rare phrases, shared vocabulary in a semantic space. Each can show you its working. Ask any of them why it connected two verses and it will mark the words.
Ask the cross-references and there is nothing to mark. The interface says so outright rather than quietly showing an empty panel:
Cross-references — here there is nothing, and that is the point. These are judgements from the tradition, not measurements of the text. They are not worth less for that, but they rest on something else, and that ought to be visible.
Most software collapses the two kinds into one list and calls the whole thing “related verses”. That is exactly what takes away your ability to tell them apart. Two verses joined by 255 votes, and two verses that share a rare Greek run of words: both are called links, but they are evidence of completely different things. The first says something about the history of reading, the second about the text.
What to do with them
Three habits make the distinction useful rather than merely tidy.
Read agreement as the strong signal. When a traditional reference and a measured method land on the same pair independently, that is the most informative result the system produces. The tradition saw something; the text confirms there is something there to see. Genesis 1:1 and John 1:1 are held together both by centuries of readers and by shared vocabulary — that convergence means more than either alone.
Read disagreement as a question. A high vote score with no textual trace at all is worth stopping on. Sometimes the connection is thematic in a way no statistic will catch. Sometimes it is a doctrinal association that accumulated by repetition. Both are worth knowing, and you cannot tell which without looking.
Watch where the votes gather. A tradition of reference collects around the passages that tradition cared about. Verses that mattered to the nineteenth-century Protestant reader carry many references; long stretches of genealogy and law carry almost none. How full that column is therefore tells you as much about the compilers as about the text.
The general point
None of this is an argument against inherited cross-references. They encode generations of close attention, and no statistical method available today would find some of what they contain.
It is an argument against presenting them in a way that makes it impossible to tell what you are looking at. A system that shows you five kinds of link and labels them all the same has thrown away the most useful information it had — which kind of claim each one is making.