Misspelling-derived tableau reconstruction

Misspelling-derived tableau reconstruction from K1-K3 known errors

In plain English: Misspelling-derived tableau reconstruction from K1-K3 known errors — using a method based on non-standard or physically-inspired methods. Every possible combination was tested (5 thousand configurations) — none produced a valid solution. This approach is ruled out within the tested scope.

NOISE Tier 2

How to read this record

Verdict – NOISE (no better than random guessing), INTERESTING (slightly above random, almost certainly coincidence), SIGNAL (statistically unusual, warrants investigation), or FULL MATCH (all 24 known letters correct).
Confidence Tier – Tier 1 = mathematical proof (permanent). Tier 2 = every possibility tested. Tier 3 = partially tested. Tier 4 = not yet tested.
Configs Tested – How many different key/parameter combinations were tried.
Best Score – How many of the 24 known plaintext letters the best attempt matched (out of 24).
Keystream Consistency (Bean) – Whether the key values at different positions are mathematically consistent with each other.
Scope Limitations – What this elimination does not rule out.

Configs Tested: 5,000
Best Score: 5 / 24 known letters matched · no better than random guessing (random guess would score: 0.0)
Keystream Consistency (Bean): FAIL
Checks whether the key values at different positions are mathematically consistent with each other.
Confidence: Tier 2: Exhaustive search under stated assumptions. Does not eliminate multi-layer variants.
Date Tested: 2026-02-20
Script: scripts/tableau/e_chart_03_misspelling_tableau.py

Scope Limitations

This elimination does not rule out:

Only tests tableau modifications derived from the 4-5 known misspellings. Does not eliminate arbitrary tableau modifications.

Assumptions

Misspelling letter pairs encode tableau column shifts or substitutions

Reproduce

PYTHONPATH=src python3 -u scripts/tableau/e_chart_03_misspelling_tableau.py

Requires the kryptos repo, Python 3.11+, PYTHONPATH=src.

Found an error? Report on GitHub