Where the numbers come from, how each one was computed, and what they cannot tell you. Every cross-reference figure on this site is reproducible from a single public dataset; the “Jesus is” pages count words in Google’s data instead, and say so.
This site has a front door and a workshop behind it. Topics pages set out a reading of scripture and name the passages it rests on. Data pages count what a public cross-reference compilation records. They make different claims, and a reader moving between them should know the ground has changed.
Topics · authored
A reading, argued by a person. Groupings, contrasts and conclusions are editorial. The passages are named so you can weigh them for yourself, and nothing here is drawn from the dataset below.
Data · compiled
Counts of what nineteenth-century editors and a modern community recorded. No interpretation is added. The figures describe the compilation as much as they describe scripture.
The two meet at the verse. A topic leads you to a passage; the data pages show you everything else that passage is tied to. That handoff is the reason this page exists.
A single row in the source file: a starting verse, an ending verse, and a vote count recording how many people agreed the connection is real. Nothing more. There is no category, no explanation, and no statement of what the two passages have in common.
Every figure on this site is built by counting rows of that shape. 344,799 of them.
Isaiah 53:5 points to 1 Peter 2:24. The reverse row does not exist. That is not a claim that Isaiah depends on Peter — it records which page an editor happened to be annotating when they wrote the note.
Three quarters of all rows have no partner running the other way. Filtering to a single direction would therefore discard most of the evidence and measure editorial habit instead of scriptural connection.
Every page consolidates both directions, with one labeled exception: the traffic page, which shows recorded direction deliberately and says so.
Consolidating is also why the Chapter Atlas draws 301,539 arcs rather than 344,799. An arc from Isaiah to Peter is the same arc as one from Peter to Isaiah, so each pair of verses is drawn once. Of the 344,799 rows, 258,324 run one way only and become one arc each; 86,430 are recorded in both directions and pair up into 43,215 arcs; and 45 rows point a verse back to itself, which cannot be drawn as an arc and are set aside. 258,324 + 43,215 = 301,539.
A connection touches two books, and both are credited with it. That is why the link-ends total, 529,346, is larger than the row count — each row is counted once for each end it touches, minus the rows that stay inside a single book.
This is also why book totals cannot be added together to recover the number of references. Genesis and Psalms each count the same connection.
Psalms carries more cross-references than any other book and is also the longest book. Dividing by a book's own verse count asks a better question: how heavily is a single line tied to the rest of scripture, regardless of how many lines there are.
It is the same move as reporting income per person rather than national income. On that measure Psalms falls from first to forty-fourth, and 1 Peter — a hundred and five verses long — comes first.
Neither number is the right one. Totals show where the mass of connection sits; rates show where it is concentrated.
Of the Bible's 31,103 verses, 30,304 are named somewhere in the compilation. 799 appear nowhere at all — among them Genesis 7:3, Psalm 115:6 and John 11:12.
That is a fact about the editors rather than about scripture. They annotated devotional and messianic material far more closely than legal, genealogical or census text, and the gaps fall where you would expect.
An earlier version of this site claimed complete coverage. The verse counts had been derived from the dataset itself, which made the claim circular. It was wrong and has been corrected.
Reading a verse in context usually means reading what sits immediately around it — the sentences before and after, the argument they belong to, the situation the writer is addressing. That is the first and more important kind of context, and nothing on this site replaces it.
A cross-reference supplies a second kind. It points at a passage somewhere else in scripture that uses similar language, makes a similar claim, or takes up the same idea. It says: this has been said before, or will be said again, in these other places.
The two do different work, and neither substitutes for the other. A verse read only through its cross-references loses the flow of its own chapter, and can be made to mean almost anything by choosing which links to follow. A verse read only within its chapter loses how the rest of scripture handles the same theme.
So the counts on this site measure the second kind and say nothing about the first. A verse with two hundred cross-references is not more important than one with two — it is a verse whose language and subject recur widely, which is a different observation entirely.
Three levels, one inside the next. 66 books contain 1,189 chapters, which contain 31,103 verses. Every figure on this site counts at one of those three levels, and mixing them up is the easiest mistake to make — so each page names the denominator it is working against.
| Books | Chapters | Verses | |
|---|---|---|---|
| Old Testament | 39 | 929 | 23,145 |
| New Testament | 27 | 260 | 7,958 |
| Whole Bible | 66 | 1,189 | 31,103 |
The distribution is uneven at every level. Psalms alone holds 150 chapters, more than twice any other book; Isaiah has 66 and Jeremiah 52. Five books — Obadiah, Philemon, 2 John, 3 John and Jude — have a single chapter each.
Chapters average 26.2 verses, but that average hides as much as it shows. Psalm 119 runs to 176 verses and Psalm 117 to two.
This unevenness is why several pages offer both a raw count and a rate. A long book scores high on totals for reasons that have nothing to do with how densely it is connected.
| Figure | What it is |
|---|---|
| 344,799 | Rows in the source file. One starting verse, one ending verse, one vote count. |
| 529,346 | Link-ends. Each row credited to both books it touches, excluding rows that stay inside one book. |
| 31,103 | Verses, taken as the highest verse number recorded in each chapter. Matches King James versification, which counts 31,102; the difference is 3 John, whose closing verse some editions split. |
| 30,304 | Verses named at least once as the end of a reference. The remaining 799 appear nowhere. |
| 17.02 | Link-ends divided by verses. A mean across the whole Bible, not a typical verse. |
| 16 | Median number of distinct verses a verse connects to. The mean is 19.9 and the maximum is 212. |
| 73,313 | Links joining an Old Testament passage to a New Testament one, both directions counted. |
| 43,087 / 30,226 | The same cross-testament links split by recorded direction: from the Old side, and from the New. |
| 2,036 | Distinct book pairs sharing at least one reference, out of 2,145 possible. 109 pairs share none. |
| 1,189 | Chapters, derived the same way as verses. |
| 258,324 | Rows with no partner recorded in the opposite direction. 74.9% of the file. |
| 301,539 | Distinct verse-to-verse links, with both directions merged into one. The Chapter Atlas draws one arc for each: the 258,324 one-way rows, plus 43,215 pairs recorded in both directions (86,430 rows). The 45 rows that point a verse to itself are not drawn. |
All counts come from the cross-reference dataset published by OpenBible.info, licensed CC BY 4.0. Most of it derives from the Treasury of Scripture Knowledge, a reference work compiled in the 1830s and long in the public domain, supplemented and scored since by community voting.
The file is tab-separated with three columns and uses OSIS book codes. References on this site follow the versification the dataset uses, which addresses 31,103 verses.
The text you read is selectable, and every translation is named beside the passage it belongs to. The default is the King James Version — the 1769 Blayney edition, in the public domain. Words the translators supplied are shown in italics, as a printed King James shows them.
The Berean Standard Bible is the other choice: a modern-English translation placed in the public domain on 30 April 2023. It is offered with thanks to Bible Hub, Discovery Bible, OpenBible.com and the Berean Bible Translation Committee. Its verses are shown as published. Where it keeps a verse only in a footnote (sixteen verses, such as Acts 8:37), a note stands in its place with the King James words. It prints 3 John 1:14 as one verse and notes that some translators begin verse 15 after “face to face”; the dataset numbers it that way, so it is divided there, with that note. When it is selected, the stories’ headings, notes and highlighted words follow its wording; those are this site’s words drawn from it, not part of the translation.
Both are included with the site rather than fetched from elsewhere, so both work with no internet connection.
Earlier versions of this site also offered the World English Bible Updated (public domain, from eBible.org, edition of May 14, 2024, stored unaltered; “World English Bible” is a trademark of eBible.org, used here of a faithful copy) and the NET Bible (requested from Bible.org as you read). Links that name them still open them.
Two verses can share references without the compilation recording any reference between them. Where they do, this site calls it an inferred cross-reference: the overlap implies a relationship nobody wrote down. That is this site’s own arithmetic, not something the dataset states, and a suggestion is not a citation.
Sharing is common, so the count alone means little. 6,733,882 cross-book pairs share at least one reference. Four or more is where coincidence stops being a plausible explanation, and 118,882 pairs reach it, across 19,011 verses. Four is the threshold; five is the stronger claim, reached by 55,914 pairs across 14,247 verses, and the difference between them is roughly a factor of six in how far each stands above chance. Below four this site says the connection is inferred but declines to put a figure on its strength.
Copyright and use. The inferred cross-reference originated here. The term, the thresholds, the degree-preserving method behind them, the account of it written on this page, and the set of pairs it produces are the intellectual property of Incredible Answers Org, © 2026. That copyright is asserted and retained.
It is nevertheless licensed to everyone, freely and irrevocably. You may use, quote, republish, adapt, build upon, teach from and redistribute it for any purpose, commercial or otherwise, without asking permission and without payment. A reference to Incredible Answers is requested and appreciated. It is not a condition of use, and nothing here is withheld from anyone who omits it.
The underlying counts are arithmetic over the OpenBible.info dataset, which is CC BY 4.0 and carries its own attribution terms. Those apply to that data whatever is done with the method described here.
Inferred from context. A verse-level dataset has a verse-level blind spot: scripture argues in passages, and a reference recorded at one verse does not attach to its neighbors. Genesis 1:28 and Exodus 20:9 share nothing, though Genesis 1:28 and Exodus 20:11 are directly cross-referenced. Where a comparison is thin, Pair-a-Verse reports recorded references within four verses either way, never leaving the chapter. The citation behind such a result is real; what is inferred is that it reaches past the verse it was filed under — and four verses is a blunt stand-in for a passage, which in narrative or proverbs may be shorter or longer. On a sample of random cross-book pairs showing nothing at verse level, 2.3% had a recorded reference that close; for deliberately chosen passages the proportion is far higher.
The share figure — what proportion of everything the two verses carry between them is held in common — is withheld when either verse has fewer than five references of its own. 10.7% of verses carrying any references are in that position, and for them a single shared verse swings the figure too far to mean anything.
To test whether the clustering is real rather than a by-product of some verses simply carrying more references, the same graph was rewired at random — every verse keeping both the number of references it makes and the number it receives, only the destinations reshuffled. Keeping both matters: a model that let heavily referenced verses stop being heavily referenced would assume away the very thing being tested. The comparison is stark:
| Shared references | In scripture | Expected by chance | One pair in | Times more often than chance |
|---|---|---|---|---|
| 1 or more | 6,733,882 | — | — | — |
| 3 or more | 291,080 | 12,400 | 38,900 | 23 |
| 4 or more | 118,882 | 744 | 650,000 | 160 |
| 5 or more | 55,914 | 57.6 | 8.4 million | 971 |
| 6 or more | 28,115 | 5.41 | 89.5 million | 5,200 |
| 8 or more | 8,391 | 0.0697 | 6.9 billion | 120,000 |
| 10 or more | 2,808 | 0.00117 | 413 billion | 2.4 million |
| 20 or more | 38 | 1.04 × 10−12 | 4.66 × 1020 | 3.7 × 1013 |
The expected column is computed rather than simulated, because beyond about eight shared references a rewiring of this graph produces nothing at all to count — twenty rewirings returned zero every time. Where the two can be compared the computed figure runs high: 1.3 times the simulated count at three shared references, 1.4 at four, 1.6 at five, 1.7 at six. That error runs in the safe direction. Every expectation in the table is an upper bound on chance, and every ratio a lower bound.
The scale is easier to feel at the bottom of the table than the top. At ten or more shared references, chance yields one such pair per 850 Bibles; scripture holds 2,808. At twenty or more, you would need on the order of a trillion Bibles before expecting a single one; scripture holds 38. The whole canon offers 483,682,753 verse pairs, which is twelve orders of magnitude short of the 4.66 × 1020 that row asks for.
What that comparison does and does not show is worth stating plainly. It measures whether the recorded connections converge on the same passages rather than spreading evenly. It does not measure how the texts were written, and it cannot separate deliberate quotation from the compilers’ own diligence. The rewired graph contains exactly the same links the compilers made; only their destinations changed.
For scale: 31,103 verses make 483,682,753 possible pairs. The compilation records 344,799 of them — one pair in every 1,403.
Versification. The cross-reference data uses ESV versification, which differs from the King James in one place. This site counts 31,103 verses to match the dataset; the traditional King James count is 31,102. The difference is 3 John, which the King James prints as fourteen verses and the ESV divides through fifteen. Both 3 John 1:14 and 1:15 show the King James verse entire, with the point of division marked, so every verse on this site matches a printed King James. Scripture links open in the King James; the link for 3 John 1:15 opens verse 14, the verse that contains it.
The atlas follows the visual form of Chris Harrison's Bible cross-reference visualization. The traffic page follows the interaction pattern of Carlo Zapponi's peoplemovin. Both implementations here are original.
The four pages under Jesus is do not use the cross-reference dataset. They count words: how often a wording appears in books (Google Books, English, 1800–2022) and how often people search for it (Google Trends, monthly since 2004). Search figures are rescaled by Google over time, so every one carries its export date.
Every phrase, the exact query sent, what came back for each source, and what none of it can tell you are set out in full on their own page.
The Chiastic Galaxy draws three independent bodies of work side by side, each credited here by name.
Arc Galaxy. The Arc of the Covenant chiastic research dataset. Every structure carries its evidence, its status (verified, supported or candidate) and its review.
Murai Galaxy. Literary structure proposals by Hajime Murai, from LiteraryStructure.info, gathered with KJV numbering in the Arc of the Covenant Murai dataset. Hebrew and Greek words are given as Murai marks them.
Xchange Galaxy. Chiasm proposals compiled from the Biblical Chiasm Exchange, Chiasmus Xchange, gathered in the Arc of the Covenant Xchange dataset.
Murai and Xchange structures are shown as their sources give them, as proposals, not graded here. Verse numbering follows the KJV, with Hebrew and Greek numbering in brackets where they differ. Story mode's steps and threads in the Chapter Atlas are written for this site, and every link they draw is checked against the cross-reference dataset before it is drawn.
The songs that play in the Chapter Atlas's story mode, the Chiastic Galaxy's guided mode and the tour on the Overview page are by WhatGodDoeth.org.