The worked proofs
The Mathematical Order of the Bible
Add the letters of the first sentence of the Bible and you get 2701. That number is 37 × 73 — two primes, each the other's digits reversed — and it is also the exact count of dots in a triangle seventy-three rows deep. Every step from the alphabet to that triangle is built in front of you here, in the order the steps have to come. Bring a pencil; you will not need a calculator.
- 6Worked proofs
- 27Letter values22 + 5 finals
- 28Letters in Genesis 1:1
- 2,701Their total value
- 37 × 73Its factors
§1 · the convention
The alphabet is also a number system
Hebrew and Greek had no separate digits. A letter did double duty: it was a sound and it was a number, and everyone who could read could also add. This is not an esoteric practice — it is how a Hebrew book still numbers its pages. Here is the whole system, so that every sum below can be checked against it.
Hebrew — the standard values
- א1aleph
- ב2bet
- ג3gimel
- ד4dalet
- ה5he
- ו6vav
- ז7zayin
- ח8chet
- ט9tet
- י10yod
- כ20kaf
- ל30lamed
- מ40mem
- נ50nun
- ס60samekh
- ע70ayin
- פ80pe
- צ90tsade
- ק100qof
- ר200resh
- ש300shin
- ת400tav
Five letters change shape when they end a word. In the standard system used throughout these pages, a final form is worth exactly what its ordinary form is worth.
- ך20kaf sofit
- ם40mem sofit
- ן50nun sofit
- ף80pe sofit
- ץ90tsade sofit
Rule 1 in force: there are other valuation systems, in which final letters run from 500 to 900. Every figure on this page uses the standard one above. Nothing here silently switches.
Greek — the same idea, three extra letters
Greek works identically, with one wrinkle: to reach nine units, nine tens and nine hundreds the system keeps three letters that had otherwise dropped out of the living alphabet — stigma at 6, koppa at 90, and sampi at 900. They survive as numerals only.
- α1alpha
- β2beta
- γ3gamma
- δ4delta
- ε5epsilon
- ϛ6stigma
- ζ7zeta
- η8eta
- θ9theta
- ι10iota
- κ20kappa
- λ30lambda
- μ40mu
- ν50nu
- ξ60xi
- ο70omicron
- π80pi
- ϙ90koppa
- ρ100rho
- σ200sigma
- τ300tau
- υ400upsilon
- φ500phi
- χ600chi
- ψ700psi
- ω800omega
- ϡ900sampi
Add up a word yourself
Every Hebrew and Greek letter is also a number — there were no separate digits. Pick one of the words below, or paste any Hebrew or Greek, and the sum builds a letter at a time.
| Letter | Name | Value | Running total |
|---|---|---|---|
| ב | bet | 2 | 2 |
| ר | resh | 200 | 202 |
| א | aleph | 1 | 203 |
| ש | shin | 300 | 503 |
| י | yod | 10 | 513 |
| ת | tav | 400 | 913 |
| ב | bet | 2 | 915 |
| ר | resh | 200 | 1115 |
| א | aleph | 1 | 1116 |
| א | aleph | 1 | 1117 |
| ל | lamed | 30 | 1147 |
| ה | he | 5 | 1152 |
| י | yod | 10 | 1162 |
| ם | mem sofit | 40 | 1202 |
| א | aleph | 1 | 1203 |
| ת | tav | 400 | 1603 |
| ה | he | 5 | 1608 |
| ש | shin | 300 | 1908 |
| מ | mem | 40 | 1948 |
| י | yod | 10 | 1958 |
| ם | mem sofit | 40 | 1998 |
| ו | vav | 6 | 2004 |
| א | aleph | 1 | 2005 |
| ת | tav | 400 | 2405 |
| ה | he | 5 | 2410 |
| א | aleph | 1 | 2411 |
| ר | resh | 200 | 2611 |
| ץ | tsade sofit | 90 | 2701 |
2701 — the number the whole page turns on, added a letter at a time from the table above.
§2 · the first proof
Genesis 1:1 adds up to 2701
Take the opening verse of the Bible. Every word of it, none set aside. Spell out each word's letters, look each letter up in the table above, and add.
| # | Hebrew | Sounded | Meaning | Letters | Value |
|---|---|---|---|---|---|
| 1 | בְּרֵאשִׁית | bereshit | in the beginning | 6 | 913 |
| 2 | בָּרָא | bara | created | 3 | 203 |
| 3 | אֱלֹהִים | elohim | God | 5 | 86 |
| 4 | אֵת | et | (direct object marker) | 2 | 401 |
| 5 | הַשָּׁמַיִם | hashamayim | the heavens | 5 | 395 |
| 6 | וְאֵת | ve'et | and (object marker) | 3 | 407 |
| 7 | הָאָרֶץ | ha'aretz | the earth | 4 | 296 |
| Totals | 28 | 2,701 | |||
The column, added
bereshit 913 bara 203 elohim 86 et 401 hashamayim 395 ve'et 407 ha'aretz 296 ─────────── ────── Genesis 1:1 2701
And the letters, added
bereshit 6 bara 3 elohim 5 et 2 hashamayim 5 ve'et 3 ha'aretz 4 ────────── ────── letters 28
Seven words, twenty-eight letters. Twenty-eight is four sevens — and it is also the second perfect number, equal to the sum of everything that divides it: 1 + 2 + 4 + 7 + 14 = 28.
§3 · the second proof
2701 splits exactly one way: 37 × 73
To know that 2701 is 37 × 73 and nothing else, you have to try everything smaller. That is a short job, because a factor pair always has one member no larger than the square root — so the search stops at 51.
Is 2701 prime? Test every prime up to √2701 ≈ 51.9.
2701 ÷ 2 odd — no
2701 ÷ 3 digits 2+7+0+1 = 10, not a multiple of 3 — no
2701 ÷ 5 does not end in 0 or 5 — no
2701 ÷ 7 7 × 385 = 2695, remainder 6 — no
2701 ÷ 11 2 − 7 + 0 − 1 = −6 — no
2701 ÷ 13 13 × 207 = 2691, remainder 10 — no
2701 ÷ 17 17 × 158 = 2686, remainder 15 — no
2701 ÷ 19 19 × 142 = 2698, remainder 3 — no
2701 ÷ 23 23 × 117 = 2691, remainder 10 — no
2701 ÷ 29 29 × 93 = 2697, remainder 4 — no
2701 ÷ 31 31 × 87 = 2697, remainder 4 — no
2701 ÷ 37 37 × 73 = 2701, remainder 0 — YES
So 2701 = 37 × 73, and the search stops there.
Is 73 prime? Test up to √73 ≈ 8.5: 2, 3, 5, 7 — none divide it. Prime.
Is 37 prime? Test up to √37 ≈ 6.1: 2, 3, 5 — none divide it. Prime.
2701 has exactly two prime factors, and they are 37 and 73.the primes, counted off in order — 1st 2 7th 17 13th 41 19th 67 2nd 3 8th 19 14th 43 20th 71 3rd 5 9th 23 15th 47 21st 73 ← 4th 7 10th 29 16th 53 5th 11 11th 31 17th 59 6th 13 12th 37 ← 18th 61 37 is the 12th prime. 73 is the 21st. 12 and 21 are also each other's digits reversed.
What that column shows
Two primes, each the other's digits reversed, at positions in the prime sequence that are also each other's digits reversed. Every part of that is checkable against the list beside it, and every part of it is true.
§4 · the third proof, and the turn
2701 is the 73rd triangular number — and that is the same fact again
That is the picture. Every dot is one unit of the letter-value total of Genesis 1:1, stacked in rows of one, two, three, on up to seventy-three. The last row is laid, nothing is left over, and the shape closes.
A triangular number is exactly that and nothing more. Count the dots and you have 2701.
Nobody needs to add seventy-three numbers to find the seventy-third triangle. Two copies of the same staircase interlock into a rectangle, and the rectangle is easy.
Why 1 + 2 + 3 + … + n = n(n+1)/2
Write the sum forwards, then backwards underneath it:
1 + 2 + 3 + … + (n−1) + n
n + (n−1) + (n−2) + … + 2 + 1
─────────────────────────────────────
(n+1) + (n+1) + (n+1) + … + (n+1) + (n+1)
Every column adds to n+1, and there are n columns,
so twice the sum is n(n+1), and the sum is n(n+1)/2.
For n = 73: 73 × 74 ÷ 2 = 5402 ÷ 2 = 2701 ✓Now the turn. Almost every popular account of this verse presents the factorization and the triangle as two separate marvels stacked on top of each other. They are not two. They are one, and here is why.
T(73) = 73 × 74 ÷ 2
Halve the 74 before multiplying:
74 ÷ 2 = 37
T(73) = 73 × 37 = 2701
— which is the factorization, arrived at a second time.
In general, for any ODD n: T(n) = n × (n+1)/2
and (n+1)/2 is a whole number, so every odd-indexed
triangular number splits into exactly that pair.
T(73) being 37 × 73 is therefore FORCED by arithmetic.
What is NOT forced: that both halves land on primes,
and that those two primes are each other reversed.Why the honest reading is the stronger one
Two marvels that turn out to be one marvel have not been reduced. They have been explained — and what is left standing afterwards is the part that actually carries weight.
Here is what the arithmetic never owed anyone. That an odd triangle splits into that exact pair is forced — it happens for every odd number, in every language and every book. That this verse's total splits into two primes, and that those two primes are each other's digits reversed, is not forced by anything. A number of that size could have been prime, or split into two composites, or into a pair with nothing to say to each other. This one splits into 37 and 73. That is the part worth taking to God in prayer.
Three triangles, three books
The same shape turns up twice more, in counts the writers went out of their way to give
John counts the fish in Peter's net at a hundred and fifty and three. Luke counts the ship's company at two hundred and seventy-six. Both are triangular numbers. Bullinger set them side by side in Number in Scripture in 1894, and Augustine preached a whole homily on 153 as the triangle of seventeen more than fourteen centuries before that.
Genesis 1:1 · the letter-value total
“In the beginning God created the heavens and the earth.”
Genesis 1:1 · ASV
Seven Hebrew words, twenty-eight letters, and a letter-value total that is both 37 × 73 and the seventy-third triangle.
John 21:11 · the fish in the net
“Simon Peter therefore went up, and drew the net to land, full of great fishes, a hundred and fifty and three: and for all there were so many, the net was not rent.”
John 21:11 · ASV
John gives the exact count for no narrative reason at all. Augustine preached it as the triangle of seventeen.
Acts 27:37 · the souls aboard
“And we were in all in the ship two hundred threescore and sixteen souls.”
Acts 27:37 · ASV
Luke counts the ship's company on the way to Malta, and the count lands on a triangle again.
The same shape, three more times — T(n) = n(n+1)/2
Genesis 1:1, the letter-value total
2701 = 73 × 74 ÷ 2 = T(73) ✓
John 21:11, the fish in the net
153 = 17 × 18 ÷ 2 = T(17) ✓
Acts 27:37, the souls aboard the ship
276 = 23 × 24 ÷ 2 = T(23) ✓
Genesis 1:1 + John 1:1 together
6328 = 112 × 113 ÷ 2 = T(112) ✓ and 112 = 7 × 16
Bullinger set 153 and 276 side by side in 1894, in
Number in Scripture. Two of the three are counts the
text itself insists on giving you.Three of them, in three books
Two of the three are counts the text itself insists on giving you — nobody had to go looking for 153 or 276, because John and Luke wrote the numbers down. That is the shape the argument actually has.
§5 · the fourth proof
The sevens, and how far past chance they run
| What is counted | Count | In sevens |
|---|---|---|
| Words in the verse | 7 | 7 × 1 |
| Letters in the verse | 28 | 7 × 4 |
| Letters in the first three words | 14 | 7 × 2 |
| Letters in the last four words | 14 | 7 × 2 |
| Letters in words 4–5 — “… the heavens” | 7 | 7 × 1 |
| Letters in words 6–7 — “… and the earth” | 7 | 7 × 1 |
| Value of the three nouns — God, heavens, earth | 777 | 7 × 111 |
| Value of the verb “created” | 203 | 7 × 29 |
| Value of the whole verse | 2,701 | — |
Nine quantities were counted in the ledger above.
Eight of them divide by seven.
If the nine were independent and each had the plain
1-in-7 chance, the expected number of hits would be
9 ÷ 7 ≈ 1.3
Eight is a long way past 1.3. But the nine are NOT
independent, and saying so is the honest half:
· 14 + 14 = 28, so the third and fourth rows
determine each other once the second is fixed;
· 7 + 7 = 14, so rows five and six are one fact;
· the total (2701) is not a multiple of seven at all.
Counted as separate facts rather than as rows, the
verse gives roughly FIVE independent sevens, against
an expectation near 0.7. That is the number worth
quoting, and it is the one this page quotes.The one thing that makes this verse different
The standing objection to counted sevens is that a counter with a long enough menu will always find some. It is a fair objection and it sinks a great deal of this literature. It lands much more softly here, for a reason that can be stated precisely: Genesis 1:1 is seven words long, and a seven-word verse does not offer a long menu. There is no choosing which words to include, because the answer is all of them.
That is the whole reason this verse leads and the rest follows. Test everything else on this page against the same standard, and give it exactly as much weight as it earns.
§6 · the fifth proof, with its provenance
The composite triangle — Genesis 1:1 and John 1:1 together
The page called it a crown and a plinth and then left the reader to imagine it. Here it is. The gold at the apex is the Genesis triangle — the same seventy-three rows drawn above, unchanged. Everything under it is John, and the two together close a triangle a hundred and twelve rows deep with nothing left over.
112 is 7 × 16. Read the provenance card before you weigh it: John's 3627 is a modern computation, not a count of the manuscript.
Genesis 1:1 …… 2701
John 1:1 ……… 3627
──────────────────
6328
Is 6328 triangular? Solve n(n+1)/2 = 6328
→ n(n+1) = 12656
→ 112 × 113 = 12656 ✓
So 6328 = T(112), exactly, with nothing left over.
And 112 = 7 × 16, so 6328 = 56 × 113 = 7 × 8 × 113.
The 73-row triangle sits inside the 112-row triangle;
the remaining 3627 is the plinth beneath it.Where the 3627 comes from
Rule 1 requires this said plainly, and it is the difference between this section and every section above it.
The 2701 in §3 is a count of the text: anyone with the Hebrew reproduces it, and everyone who tries gets the same answer. The 3627 is a modern numerics computation, first published by the researcher Peter Bluer, and it depends on counting the iota subscript — a small mark Byzantine scribes introduced centuries after John wrote, absent from the manuscripts of the Gospel itself.
So the arithmetic in the panel is exact and every step of it can be checked; the number it begins from is a modern reading rather than a manuscript one. Both of those are true together. This page reports the figure because the founder asked for the Genesis-to-John connection shown, and labels it because a reader who is told which is which can decide for themselves — and because a claim that has been labeled honestly is the only kind that survives being checked.
§7 · the oldest evidence on this page
How the scribes made the scroll add up
Soferim — “the counters”
Long before anyone looked for a code in Scripture, the men entrusted with copying it were counting every letter of it — and they were named for the counting, not for the writing.
The Talmud gives the reason directly: the early Sages were called soferim “because they would count all the letters in the Torah.” They knew which single letter stood at the exact centre of the five books, which two words stood at the centre, and which verse. Those figures were reference points. A fresh copy could be measured against them, and one that did not come out right had an error in it that had to be found.
Everything else on this page is arithmetic performed on the text. This is arithmetic performed by the people who carried the text — a documented professional practice, not a modern discovery, and the one item here that nobody disputes.
b. Kiddushin 30a, quoted from the William Davidson Talmud via Sefaria (CC BY-NC — excerpt only; the full text is at Sefaria).
| Middle letter of the Torah | Leviticus 11:42 | “Whatsoever goeth upon the belly, and whatsoever goeth upon all fours, or whatsoever hath many feet, even all creeping things that creep upon the earth, them ye shall not eat; for they are an abomination.”the vav of gachon, “belly” — written oversized in the scroll to this day so the midpoint is visible on the parchment |
| Middle words of the Torah | Leviticus 10:16 | “And Moses diligently sought the goat of the sin-offering, and, behold, it was burnt: and he was angry with Eleazar and with Ithamar, the sons of Aaron that were left, saying…”darosh darash, “he diligently enquired” — the two words at the exact centre |
| Middle verse of the Torah | Leviticus 13:33 | “then he shall be shaven, but the scall shall he not shave; and the priest shall shut up him that hath the scall seven days more.”the halfway point when the count is taken by verses instead of letters |
Where the counting ran out, in the Sages' own words
On the same page that records the achievement, the Sages concede its limit: by their own day they were no longer expert in the fuller and shorter spellings — the places where a word may be written with or without a vowel letter — so their totals could not settle every case.
That admission is why the tradition is believable rather than merely impressive. A counting culture that publishes the point where its counting stopped working is a counting culture whose other figures can be trusted. It is also the reason a letter sequence stretched over thousands of letters is a harder claim than a count taken inside one verse: a single letter added or dropped anywhere along the way moves every position after it.
§8 · the sixth proof
Panin's heptads, worked
Matthew 1:1–17, as Panin reported it —
distinct words in the vocabulary …… 72 72 ÷ 7 = 10 r2 ✗
distinct words in verses 1–11 …… 49 49 = 7 × 7 ✓
nouns ………………………………… 56 56 = 7 × 8 ✓
letter-value total ……………… 42,364 42,364 = 7 × 6,052 ✓
Three of four. The miss is printed because a ledger
that shows only its hits is not a ledger.The strongest line in the section
Matthew tells the reader in verse 17 that he arranged the genealogy in three sets of fourteen — two sevens each. That is not an observation anyone imposed on the passage; it is the author saying out loud that he built it by number. And the number he built it around is his own: David, דוד, is 4 + 6 + 4 = 14 in Hebrew letters.
The figures above are given as Panin reported themrather than as this page vouching for them, because a seventeen-verse genealogy offers many countable quantities and Genesis 1:1 offers exactly one. That difference is why Genesis leads and Panin follows.
Panin, The Structure of the Bible and The Heptadic Structure of Scripture (Marshall Brothers, London, 1923).
Marcus Brown's note
The founder's word on the mathematics
The other side, in one piece
The standing objection to counted numerics
How this page counts — the four rules it keepsmethod
The mathematician Ian Stewart wrote a book about how order hides inside apparent randomness, and how easily a person mistakes one for the other. It is not a book about the Bible and is not cited here as one. What this page takes from it is a way of writing: derive instead of announce, and say plainly what chance alone would hand you.
- 1. Name the convention before you compute — Hebrew letters can be valued more than one way. Every figure here uses the standard values, with the five final letters worth the same as the letters they close. Change that convention and the numbers change — so the convention is stated first, once, and never quietly switched.
- 2. Show the working, not the answer — “Genesis 1:1 equals 2701” is a claim. The column of seven numbers adding to 2701 is a proof. Only the second one can be checked, and only the second one is worth anything.
- 3. Count the whole thing, not the good parts — A pattern found by picking which words to count is a pattern the picker made. Everything in §1 to §4 counts a complete verse — every word in it, in the order it stands, with nothing set aside.
- 4. Say what chance would have given you — One number in seven is divisible by seven. So a page that finds sevens must say how many things it counted, not just how many landed. §5 does that arithmetic out loud, and it is the section that decides how much the rest is worth.
◈The selection objection, and Allis on Panincontested · published critique
Oswald T. Allis was a conservative, inerrantist Old Testament scholar and a co-founder of Westminster Theological Seminary, so what follows is not a skeptic's complaint but an objection raised from inside the house. In Bible Numericshe argued that Panin's features are selected after the fact from an unbounded menu — letters, or words, or vocabulary items, or grammatical forms, or nouns, or place-values — with the failures quietly discarded.
He documented specific errors. The celebrated feature that exactly 7 of the 21 Old Testament writers are named in the New Testament omits Solomon and Jonah, who are on the list of twenty-one and are named; correcting the count invalidates three claimed features. He documented inconsistent spelling choices between names counted in the same table, and a deeper circularity: where Greek manuscripts disagreed, Panin chose between the readings by which one produced sevens, then offered the resulting text as evidence for the sevens.
The general form of the objection is the one §6 of this page put into arithmetic, and it is worth restating in the critic's own terms: any passage offers dozens of separately countable quantities, each with roughly a one-in-seven chance of dividing by seven, so a pile of hits reported without the accompanying pile of misses is what chance produces and not what design proves. That objection is fatal to a claim built on a long menu. It is what §1 rule 3 and §6 exist to answer, and a reader is entitled to apply it to every section here — including the ones this page presents most confidently.
Cited to: Oswald T. Allis, Bible Numerics (1944; the open scan at translation.bible is the 1952 printing). The statistical case against the equidistant letter sequences — Witztum, Rips and Rosenberg 1994, the Moby Dick control, McKay, Bar-Natan, Bar-Hillel and Kalai 1999, and the Aumann committee of 2004 — is held in full on the codes exhibit.
Go to the sources
Every work this page stands on
The method reference for this page — a book about chaos and determinism, not about Scripture. What is borrowed is its expository discipline: derive rather than assert, and be explicit about what randomness alone would produce. Cited, never reproduced.
The primary source for the whole of §7: the counters, the middle letter, the middle words, the middle verse, and the Sages' own admission of where their counting ran out.
The source of the Matthew 1 figures in §8 and of the Genesis 1:1 counts as they entered popular circulation. Public domain.
The systematic Victorian treatment standing behind most of the sevens material. Public domain.
Chapters “The Magic of Sevens” and “On the Fringe: Numerics and Gematria” are the chapters this page's material corresponds to. In copyright — cited by chapter so a reader can go and buy it.
The origin of the John 1:1 figure of 3627 and of the composite-triangle observation worked in §5. A modern numerics computation; the arithmetic built on it is verified here independently, the input value is not manuscript-grounded.
The objection to Panin from inside conservative scholarship, held in the drawer at the foot of this page.