NOVA Online | Decoding Nazi Secrets

Comments from Ed Rupp and Jim Gillogly
Back to Hall of Fame

Comments from Ed Rupp

Well, I first checked to see if the keyword for it was in my dictionary. It wasn't.

Next I tried the suggested cleartext BEWARE ICE WEASELS but nothing looked promising. Next, I tried the second phrase RED PENGUIN FRENZY, which fit only at one place in the message:

	QA DF HS FZ WN AI DS MU RU
	_r ed pe ng ui nf re nz y_

with a best partial playfair square of:

	fina_
	h_mp_
	g_zq_
	ewus_ 
	d_yr_

which can quickly be filled in as:

	FINAL
	HKMPV
	GXZQC
	EWUST
	DBYRO

If this hadn't worked, or if no suitable probable text was available, I would have applied a random improvement swap solving algorithm (a.k.a. "shotgun hill climbing"), which may have produced the answer or gotten close enough that inspection of the best results would indicate the answer.

Comments from Jim Gillogly

Ed's solution follows classical lines: place the tip, then deduce the square from the tip. He comments that he might have used a "shotgun hill climbing" algorithm if his first attempt had failed. This is also correct. The idea here is to try various squares randomly, then move the letters around in the square while "scoring" the assumed plaintext resulting from the new square. Using this method a program can solve a Playfair of this length without any crib. Interested readers can join the American Cryptogram Association (see the unofficial ACA Web Site at http://www.und.nodak.edu/org/crypto/crypto/ ) and order the November-December 1995 issue of their magazine The Cryptogram, in which I describe the procedure. Shotgun hill climbing can also be used so solve Cipher #3, the double transposition cipher, without a crib.

Crack the Ciphers | Send a Coded Message | A Simple Cipher
Are Web Transactions Safe? | Mind of a Codebreaker | How the Enigma Works
Resources | Teacher's Guide | Transcript | Site Map | Decoding Nazi Secrets Home