THREAT MODEL AND MOTIVATION
OTP is the only cipher with a mathematical proof of perfect secrecy. The catch: "used correctly" is a very high bar.
0.1 What OTP Solves
Modern cryptography (AES, RSA, ECC) is computationally secure. This means an adversary cannot break it given current technology and time constraints. This is not the same as saying they cannot break it. Computational hardness assumptions can fail. Algorithms have been deprecated. Quantum computing threatens asymmetric systems.
OTP is different. Shannon proved in 1949 that a one-time pad, used correctly, is perfectly secret. An adversary who intercepts every ciphertext group and has unlimited compute learns nothing about the plaintext. There is no attack surface. The security is mathematical, not computational.
The catch: "used correctly" is a very high bar.
0.2 The Soviet Experience
The VENONA project (US SIGINT, 1943–1980) decrypted Soviet OTP traffic not because the OTP was mathematically broken, but because:
- Key pages were duplicated under wartime production pressure (1942)
- Operators reused key material on different messages
- Messages were sent twice (once in OTP, once in a weaker system)
- Some operators used non-random key material
A single page reused across two messages allows the analyst to XOR (or in this case, subtract mod 10) the two ciphertexts together, eliminating the key entirely. The resulting "depth" is two superimposed plaintexts, which a skilled analyst can recover using linguistic cribs and statistical methods.
The cipher is not your weak point. You are your weak point.
0.3 Why Straddling Checkerboard
Raw OTP requires your plaintext to already be numeric. English text is not numeric. Simple A=01, B=02 substitution works but is inefficient — every letter becomes two digits, so a 26-letter message requires 52 key digits.
The Straddling Checkerboard converts text to digits using variable-length encoding. High-frequency letters (E, T, A, O, N, …) get one digit. All others get two. A typical English message encodes to roughly 1.3 digits per letter instead of 2.0. This matters because key material is finite and precious.
0.4 Why Mod-10
Binary XOR works on bits. Our key material is decimal digits (from dice, from printed pads). Mod-10 addition is the decimal equivalent of XOR: adding without carrying, wrapping at 10. Like XOR, it is its own inverse complement with a slight modification: to decrypt, you subtract mod 10, which is the same as adding the tens complement.
Mod-10 has one advantage over XOR for hand operations: it uses familiar decimal digits. No base conversion required.
0.5 System Overview
The complete pipeline:
PLAINTEXT (letters/digits/special)
|
v
[STRADDLING CHECKERBOARD ENCODE]
|
v
INTERMEDIATE DIGIT STRING
|
v
[MOD-10 ADD against KEY STREAM]
|
v
CIPHERTEXT DIGIT GROUPS
---[transmitted]---
CIPHERTEXT DIGIT GROUPS
|
v
[MOD-10 SUBTRACT key stream (or add complement)]
|
v
INTERMEDIATE DIGIT STRING
|
v
[STRADDLING CHECKERBOARD DECODE]
|
v
PLAINTEXTEach step is reversible and independent. Master each before combining them.