The Bible code

What the Bible code is

Write the Hebrew text with no spaces. Pick a letter and a number. Read every letter that many places apart. Sometimes those letters spell a word. That is an equidistant letter sequence, or ELS, and the number is the skip.

Bible code and Torah code mean the same thing. Torah codes is the Jewish term, and it comes from the Five Books of Moses, where the classic examples and the main research were done. The Bible code became the common English name after Michael Drosnin's book. Both refer to the Hebrew text. A translation has different letters in a different order, so these sequences do not carry over into English.

Here is the classic example, which you can check with a pencil. Start at the first ת of Genesis, the last letter of בראשית. Count 50 letters and you reach a ו. Count 50 more for a ר, and 50 more for a ה. You have spelled תורה, Torah. Rabbi Michael Dov Weissmandl is remembered for this one, decades before computers.

How the pictures are made

Lay the text out in rows of equal width, often the skip itself, and a sequence stands as a straight column. Then you can look for other words nearby. That picture, called a matrix, is what most people have seen. The words in it may be other sequences at their own skips, or plain text running across a row.

Every part of that picture is a choice: the width, the rectangle, the neighboring words, the spellings. The whole debate turns on how much those choices can do. Our matrix guide walks through one.

The science came first

The serious work began before the book. In 1994 the journal Statistical Science printed a paper by Doron Witztum, Eliyahu Rips and Yoav Rosenberg. They took the names of rabbis from a Hebrew encyclopedia and the dates of their birth or death, and tested whether those pairs sat closer together as sequences in Genesis than chance allows. They reported a rank of 4 in a million for the true pairing, and flat results in control texts such as Isaiah and a Hebrew War and Peace.

The editor presented the paper as a puzzle his referees could not explain. Note what it did not do. It said nothing about future events. It was a test of whether related words cluster. Our WRR page explains every step.

The 1997 book

The Bible Code, by the American journalist Michael Drosnin, came out in 1997 and sold around the world. A sequel followed in 2002. It reprinted the 1994 paper in full, and it went much further than the paper did. The book's central claim, in its own terms, was that the Torah held an encoded record of later events.

Its best-known story concerns Yitzhak Rabin. The book laid out a matrix with Rabin's name at a skip of nearly 5,000, crossing the words רוצח אשר ירצח in Deuteronomy 4:42. Drosnin said he read these as a warning about an assassin, and that he wrote to Rabin about it more than a year before Rabin was killed in November 1995. The book presented these matrices as predictions, some of them of events still to come.

That phrase is worth looking at. It is plain text, not a sequence, and the verse is about the cities of refuge. JPS 1917 gives "that the manslayer might flee thither, that slayeth his neighbour unawares". A reading of a phrase apart from its verse is a choice the reader makes, and critics pointed that out at once.

The authors of the 1994 paper did not stand with the book. Rips issued a public statement that he did not support its work or its conclusions. Witztum and Rips both rejected the idea that sequences can be read for future events.

The Moby Dick answer

In a Newsweek interview in June 1997, Drosnin dared his critics to find a message about the assassination of a prime minister in Moby Dick. If they did, he said, he would believe them.

find a message about the assassination of a prime minister

Michael Drosnin, as quoted in Newsweek, June 9, 1997

Brendan McKay, a mathematician at the Australian National University, took him up on it. He searched the English text of Melville's novel in the same style and published matrices placing the names of assassinated figures beside words about their deaths, among them Indira Gandhi, Martin Luther King Jr., Abraham Lincoln and John F. Kennedy, and Rabin himself.

McKay's point was never that Moby Dick knows anything. His point was that with free choice of words, spellings, skips and windows, you can build this kind of picture in any long text. That is a claim about the method. It does not touch the 1994 paper, which used a formal test on a fixed list. For that, the critics needed something stronger.

The reply to the 1994 paper

That came in 1999, in the same journal. McKay, Dror Bar-Natan, Maya Bar-Hillel and Gil Kalai reran the test and got the same small number. They then argued that it came from the choices behind the list. Rabbis have many names, and they report that the paper used fewer than half of those known. Dates can be written in at least eight forms. The paper used three, and the other five, tried alone, do poorly. To show that this much freedom was enough, two of them changed some names on the list and got a similar score in War and Peace, with the very text that WRR had used as a control.

They also ran new experiments with an outside expert, and found nothing. In 2004 a Hebrew University committee that included Rips and one of his critics reported on a related claim, and both of its experiments came in at the middle of the pack. Two members dissented. The critique page has the full record, including what proponents say in reply. Witztum and others still publish on the subject.

So is the Bible code real?

It depends on what you mean. The sequences are real, in the plain sense. They are there, you can count them, and so can anyone with the same text. Whether they hold more than chance would put there is the live question. Here is what you see when you measure.

Most searches look like that. Count and chance sit almost on top of each other. A single word, found at a single skip, tells you almost nothing. The serious claims in this field rest on whole lists and rankings, and the published record on those is the argument above. We lay it out and let you weigh it.

Can it tell the future?

Torah Codex makes no such claim, and nothing we measure supports one. A search can only find what you already thought to type. The authors of the original paper rejected this use. We do not publish matrices about current events, elections or living people.

Measure it yourself

Torah Codex is a search engine for these sequences that tries to keep its users honest, including us. Type a word in English and choose its Hebrew spelling. Choose the Torah, the whole Tanakh, a book or a range of chapters. You will get an exact count, what chance alone would give, a matrix you can explore and a link from every letter to its verse.

Then press Run 25 shuffles and see what happens in a text with no order. Watch the trial counter as you try other spellings. If you have a real idea to test, lock it before you run it. A free account includes 10 searches.

Questions

Who wrote The Bible Code?

Michael Drosnin, an American journalist. The book came out in 1997, and a sequel followed in 2002.

Are Torah codes and Bible codes the same thing?

Yes. Both name equidistant letter sequences in the Hebrew text. Torah codes is the Jewish term, and Bible code became common with the 1997 book.

Do Bible codes work in English?

No. A sequence belongs to one exact string of letters. A translation has different letters in a different order, so we search the Hebrew of the Leningrad Codex.

Was the Bible code debunked?

The 1999 paper in Statistical Science and the 2004 Hebrew University committee report did not confirm it, and 45 mathematicians signed a statement calling the evidence unconvincing. Some proponents, including one of the 1994 authors, dispute those findings and keep publishing. Our critique page gives both sides.

Can I search my own name?

Yes. Type it in English and choose a Hebrew spelling, or type the Hebrew. You will see the count and the chance figure side by side. Common letters mean most names appear many times, and that is expected.