Cold War Secrets
Secret Writing: A History of Codes, Ciphers and Codebreaking
Course 1 · Chapter 10 · Cold War Secrets
Chapter 9 ended with machines and secrecy at their wartime peak. After 1945 the contest changed shape. One cipher was proved to be unbreakable in principle, and the fight moved to people: spies, couriers, listening posts and the small mistakes that even a perfect system cannot prevent.
The Cipher That Cannot Be Broken
The one-time pad is the simplest strong cipher there is. The sender and receiver each hold a copy of a pad of random letters or digits. To encipher, each message letter is combined with the next pad letter, as in the shift ciphers of Chapter 2, except that the shift changes every time and is never repeated. The used part of the pad is then destroyed.
Because every pad letter is random, any cipher text could mean any message of the same length: with a different pad, the same letters decode to something else entirely. An enemy who intercepts the message has nothing to grip. This property, perfect secrecy, was later shown mathematically, but the idea of a random key used once is older than the proof.
Perfect secrecy: the cipher text gives the enemy no clue about the message, whatever their computing power.
VENONA: a US programme (from 1943) that studied Soviet messages enciphered with reused pad pages.
Number stations: radio broadcasts of spoken digits, widely believed to carry pad-enciphered messages to agents.
Kryptos (1990): a sculpture at a US intelligence agency's headquarters carrying four enciphered passages.
The Price of Perfection
A perfect cipher sounds like the end of the story, but the pad moves the difficulty somewhere else. It must be truly random, as long as every message you will ever send, delivered safely to both ends, kept secret, and never reused. For a spy in a foreign city, that means a physical pad hidden in a book or a shoe, and digits written out by hand. For an embassy handling hundreds of messages a day, it means a mountain of pads.
| Requirement | Why it matters | What goes wrong |
|---|---|---|
| Truly random | Patterns in the key leak patterns in the message | Operators invent "random" digits that are not |
| As long as the message | A short key must repeat, as in Chapter 5 | Pad runs out, so pages are copied |
| Delivered securely | The key has the same secrecy problem as the message | Couriers can be caught or pads copied |
| Used once only | Two messages on one key cancel the key out | Reuse under pressure or shortage |
The same problem returns at the end of this course: sending a key safely is the very thing Chapter 11 sets out to solve.
VENONA: When the Pad Was Reused
The weakness was human. Wartime shortages and the sheer volume of traffic are widely reported to have led to some Soviet pad pages being issued twice. That is the reuse failure of Chapter 7 and Chapter 9 once more, now with the strongest cipher in history. Two messages enciphered with the same pad pages can be combined so the pad drops out, leaving the difference between two plain texts, which codebreakers attack with guessed words.
The work was slow, partial and spread over decades. Most messages were never read, but the fragments that were read are reported to have helped identify Soviet agents and gave a glimpse of how much was being taken. The existence of the programme was revealed in the 1990s. It is a good example of the pattern running through this whole course: the mathematics held, and the habits of the people using it did not.
Listening Posts and Number Stations
Both sides built enormous networks to listen. Signals intelligence meant intercepting radio, telephone and cable traffic, recording it, and sorting it. Even when a message could not be read, traffic analysis still told you something: who talked to whom, how often, and when the volume suddenly jumped before an event. Hiding the content is not the same as hiding the pattern of use.
Number stations are the most visible leftover. For decades, shortwave radio listeners have heard flat voices, sometimes in many languages, reading out groups of digits at fixed times, often with a musical introduction. The usual explanation is that they send one-time pad messages to agents, who need only a radio receiver and a pad. A shortwave receiver gives no sign that anyone is listening, which is the point. Governments have rarely confirmed them, so most of what is said about them is informed guesswork.
Kryptos: A Puzzle in Copper
In 1990 a sculpture called Kryptos was installed in a courtyard at an American intelligence agency's headquarters. It is a curved copper screen cut with thousands of letters in four sections. Three of the four were solved by outside enthusiasts and by agency staff, and the first three are reported to use ciphers of the kind met earlier in this course, such as keyword alphabets and transposition. The fourth part was still unsolved at the time this chapter was written. Chapter 12 returns to it with the other open puzzles.
What the Cold War Teaches
First, a perfect cipher still needs a perfect process, and processes fail under pressure. Second, metadata is a message too: traffic analysis works when content does not. Third, delivering the key is the hard part, and that is exactly the problem the next chapter solves.
Show the worked answer
(a) E(4)-X(23) = -19, which is 7 = H. Q(16)-M(12) = 4 = E. N(13)-C(2) = 11 = L. V(21)-K(10) = 11 = L. Z(25)-L(11) = 14 = O. The message is HELLO.
(b) E(4)-I(8) = -4, which is 22 = W. Q(16)-C(2) = 14 = O. N(13)-W(22) = -9, which is 17 = R. V(21)-K(10) = 11 = L. Z(25)-W(22) = 3 = D. The "message" is WORLD. Both read as perfectly good English, and nothing in the cipher text says which pad is the real one. That is perfect secrecy: without the true pad, every message of that length is equally likely.
Where This Leaves the Secret Writer
By the 1970s the situation was strange. Perfect security existed on paper, and the world's governments relied on the next best thing: huge, carefully managed key distribution systems. Banks and businesses had no such resources, and the new computer networks needed a way for strangers to agree on a secret. The next chapter tells how that was solved, in public and in secret.
Exercises
A number station broadcasts the group 3816. Your pad page for the day reads 7592. Subtract the pad digits from the group digits one place at a time, with no carrying (for example 3 - 7 is 6, because you wrap around from 10). What is the message?
Use A=0 up to Z=25. Two messages were enciphered with the same pad, giving the cipher texts JE and RD. You are told the second plain text is ON. Subtract the cipher texts to find the difference, add the second plain text back to recover the first plain text, and then recover the pad letters.
Use A=0 up to Z=25. (a) Encipher SAFE with the pad KTRB. (b) Give two reasons why a one-time pad is hard to use in practice for an organisation sending thousands of messages a day, and say which of the four requirements in the table each reason breaks.
Quick Reference
- One-time pad: random key, as long as the message, used once
- Perfect secrecy: any message of the right length is equally likely
- VENONA: reused pad pages gave codebreakers a way in
- Traffic analysis: who, when and how much, even without content
- Number stations: spoken digits, probably pad messages to agents
- Check: dates, names, and the status of Kryptos
Next chapter: Public Keys, with the paint-mixing and padlock ideas, and the secret discovery that changed the Internet.