The most famous — and most fiercely disputed — claim about hidden codes in the Torah, told from all sides.
In 1994 a paper in a respected statistics journal appeared to show something startling: that the names of famous rabbis, paired with their birth and death dates, sit unusually close together as hidden letter-skip sequences in the Book of Genesis — centuries before those rabbis were born. It set off one of the sharpest methodological fights in modern statistics, which is still argued today. Here is what happened, in three layers.
An Equidistant Letter Sequence (ELS, Hebrew dilug, “skipping”) is found by starting at some letter in the text and reading every n-th letter — every 7th, every 50th, and so on. Skip far enough through a long text and words will inevitably appear this way; that much is not in dispute. The WRR claim was subtler and statistical: that for a curated list of rabbis, each rabbi’s name-ELS and his date-ELS tend to appear closer together in Genesis than chance would predict.
Doron Witztum, the mathematician Eliyahu Rips, and Yoav Rosenberg published “Equidistant Letter Sequences in the Book of Genesis” in Statistical Science (1994). The journal’s editors were openly skeptical — computers can “mine” any large text for surprising-looking patterns — so they published it not as an endorsement but as a “challenging puzzle” for readers to try to solve. (That framing was widely misread by outsiders as a stamp of approval; it was not.)
The key detail — and the one that makes the experiment interesting rather than trivial — is the second list. The paper actually reported two experiments. After a referee raised the natural worry that the first list of rabbis might have been hand-tuned to succeed, the authors compiled a second list of 32 less-famous rabbis, defined by a fixed rule (those whose entries in the Margaliot rabbinic encyclopedia ran 1.5 to 3 columns). The names and spellings were prepared by Professor Shlomo Havlin of Bar-Ilan University. Because the selection rule was fixed in advance, the second list was meant to work like a pre-registered replication — a guard against cherry-picking. It, too, showed the effect. That is the result our software reproduces below.
Our reproduction of the published computation
Using WRR’s own second list, our independent implementation of their “corrected distance” method computes a combined score of ρ₀ = 4.0 × 10⁻⁶, with two of the four statistics ranking 1st out of 1,000,000 random permutations (N = 163 word-pairs, k = 62). In other words, we reproduce the numbers WRR published.
What this does and does not mean. It means our arithmetic is faithful to their method on their data — the computation is real and repeatable. It does not mean the codes are real. The entire dispute (below) is about whether that particular list of names and spellings was itself tuned to produce this result — and when independent scholars rebuilt the list from scratch, the effect disappeared.
Five years later the same journal published “Solving the Bible Code Puzzle” by Brendan McKay, Dror Bar-Natan, Maya Bar-Hillel, and Gil Kalai (“MBBK”). Their central finding: the result is “fatally defective,” and “merely reflects on the choices made in designing their experiment and collecting the data for it.” The ELS score, they showed, is extraordinarily sensitive to tiny changes in how each rabbi’s appellations are spelled — and a rabbi typically has several defensible spellings.
Crucially, MBBK did not merely argue — they ran their own independent replication. They hired Dr. Simcha Emanuel, a specialist in rabbinic history, to compile the appellations and dates by his own professional judgment, and applied WRR’s selection rule cleanly. The effect vanished. They further demonstrated that with the same latitude in word choice, an equally “significant” result could be conjured from a Hebrew translation of Tolstoy’s War and Peace — a text no one claims is encoded.
The journal’s editor, Robert Kass — who had originally called WRR a “challenging puzzle” — wrote in his introduction that “considering the work of McKay, Bar-Natan, Kalai and Bar-Hillel as a whole it indeed appears, as they conclude, that the puzzle has been solved.” No rebuttal of MBBK has since appeared in a peer-reviewed journal.
The code’s proponents did not concede, and their responses deserve a fair hearing. When MBBK suggested that WRR had quietly avoided certain “alternate” spellings, WRR reran the experiment using MBBK’s own suggested alternates and reported results as strong or stronger — which, they argued, undercuts the “wiggle-room” theory.
Harold Gans, a former senior cryptologic mathematician at the U.S. National Security Agency, made two arguments. First, that MBBK’s “tuning” theory really implies a conspiracy — among WRR, Havlin, the rabbis who later attested to Havlin’s list, and the participants in Gans’s own independent experiment — and that “the number of people necessarily involved… will stretch the credulity of any reasonable person.” Second, he ran a separate experiment matching rabbis to the cities of their birth and death (rather than dates), and reported a positive result there too. Robert Haralick, a computer-science professor, added a “redundancy” experiment reporting that WRR’s list kept succeeding at deeper, non-minimal skips where a control list did not.
The most careful voice belongs to Robert Aumann, the Nobel laureate in economics, who followed the controversy for years and has no stake in mysticism. His verdict is the honest one to end on:
“Though the basic thesis of the research seems wildly improbable, for many years I thought that an ironclad case had been made for the codes… Then came the work of the ‘opponents.’… it did convince me that [manipulation] could have been [done]; that manipulation was technically possible… Research conducted under my own supervision failed to confirm the existence of the codes — though it also did not establish their non-existence. So I must return to my a priori estimate, that the Codes phenomenon is improbable.”
The mainstream scientific verdict is that the Great Rabbis Experiment does not demonstrate a real phenomenon: the published effect is inseparable from the freedom the experimenters had in choosing and spelling names, and it does not survive independent replication. The proponents’ counter-arguments about conspiracy-implausibility and independent experiments are real and worth reading, but they have not overturned that verdict in the peer-reviewed literature. A fair summary is Aumann’s: possible, but improbable.
That is exactly why TorahExplorer.com treats every code-style result as exploratory. Our tools let you reproduce and inspect the computation yourself — the arithmetic is real and repeatable — while making no claim about hidden meaning. Seeing precisely where a “finding” comes from is the best inoculation against over-reading one.
Build up to eight terms — click a rabbi from WRR’s two lists, type a Hebrew word, or transliterate an English name — and draw them together on one Genesis letter grid. The first term forms the column; the others are highlighted wherever their letter-skip sequence crosses the same window. Keep the debate above in mind while you play: proximity like this turns up in any long text once you’re free to choose the words and their spellings. That freedom is the finding and the flaw at once.
Add one or more terms and press Run the codes to draw the grid.
Type a Hebrew word
Transliterate an English word
Or click one of the 66 rabbis from WRR’s lists
Loading the rabbi list…
Honesty notes: clicking a rabbi fills the same term boxes typing would, so a click and a manual entry give the identical result. This tool searches the Koren letter-text (304,805 letters) — the same edition the Rips/Witztum/Rosenberg research used — so what you find here corresponds to the published computations. (The site’s reader uses the Masoretic MAM text, which differs from Koren by a handful of spelling variants.) Finding a word as an ELS proves nothing on its own — that’s the whole lesson above.
Witztum, Rips & Rosenberg, “Equidistant Letter Sequences in the Book of Genesis,” Statistical Science 9(3):429–438 (1994) — PDF. · McKay, Bar-Natan, Bar-Hillel & Kalai, “Solving the Bible Code Puzzle,” Statistical Science 14(2):150–173 (1999) — PDF. · Overview and full controversy timeline — Wikipedia: Bible code. · Proponent and critic archives — McKay’s codes archive (ANU).
Presented for educational and exploratory purposes. TorahExplorer.com makes no claim as to the significance of any code-style result. See our Terms & Disclaimer.