About Bible Codes
Search for equidistant letter sequences (ELS) in the Torah. Multi-term clustering, 2D/3D matrix visualization, and statistical significance testing.
* in display and exports to avoid genizah obligations.
Please use electronic viewing when possible. If you print any output, treat the materials with appropriate reverence.
Real-time ELS Scan
Search Terms
Alternate spellings: put variants on the same line separated by spaces.
All alternates are searched but only the best-matching one is used per cluster/matrix.
Example:
דוד דויד ← two spellings of David
References & Texts
BDB, Strong's, 260K words
ActiveRashi, Ibn Ezra, Ramban, etc.
PlannedBabylonian Talmud full text
PlannedSix orders, full text
PlannedMystical commentary
PlannedMidrash Rabbah, Tanchuma
PlannedCross-reference and search across classical Hebrew texts. Data sourced from Sefaria and open-access repositories.
WRR 1994 Experiment Academic Replication
Replication of the famous experiment by Witztum, Rips & Rosenberg published in Statistical Science (1994). 32 rabbis from the Encyclopedia of Great Men in Israel, each paired with Hebrew death dates. The proximity of name-date ELS pairs in Genesis was found to be statistically significant at p = 0.000016 (1 in 62,500).
| # | Rabbi | Appellations | Date | Pairs | D(w) | Δ |
|---|
Methodology Details
Source: Witztum, Rips & Rosenberg, “Equidistant Letter Sequences in the Book of Genesis”, Statistical Science, Vol. 9, No. 3, August 1994, pp. 429–438.
Text: Genesis (Koren edition), 78,064 consonantal Hebrew letters. Final letter forms (sofiot) normalized to standard forms for matching.
ELS Definition: For skip value d (|d| ≥ 2), letters at positions h, h+d, h+2d, …, h+(n−1)d. Skip ±1 (open text) is excluded per the WRR protocol.
Skip Range D(w): Dynamic per term. For word w of length k with letter probabilities pi, expected ELS count at skip d: E(w,d) = (L − (k−1)d) ⋅ ∏pi. D(w) is the smallest d where ∑d=2..D E(w,d) ≥ 10, capped at user maximum.
Distance Δ: Text arranged as 2D array with |d| columns. Position p maps to row = ⌊p/|d|⌋, col = p mod |d|. Δ = min Euclidean distance between any name-ELS letter and any date-ELS letter. Both array widths (|dname| and |ddate|) tested; smaller distance reported.
Original significance: 999,999 random permutations of name-date assignments tested. Actual assignment ranked 4th of 1,000,000, yielding p < 1/62,500.
This replication: Two modes available: Quick Run uses the 2D cylindrical proximity measure with dynamic D(w) and both forward and backward ELS (|d| ≥ 2). Full WRR additionally computes the c(w,w′) perturbation statistic: 125 spatial perturbations (triples (x,y,z) ∈ {−2..2}³) shift the last 3 ELS letter positions by cumulative (x, x+y, x+y+z). Proximity ω = max(1/δ) across 10 row lengths hi = round(|d|/i). Then P₁ (binomial tail on c<0.2 count) and P₂ (Gamma CDF on ∏c) yield the overall statistic P = 2·min(P₁,P₂).
Data: Appellations and dates from WRR Table 3 (Appendix). Rabbis #4 and #8 excluded—no recorded death dates.
Hebrew Date ELS Density Map Pre-computed
ELS density heatmap for all Hebrew calendar dates and years across the Koren Edition Torah (304,805 letters). Each cell aggregates equidistant letter sequences encoding that date across all skip values.
Date ELS index not loaded yet.
Generate it with: python3 tools/build-date-els-index.py
Or compute in-browser (slower):