PLATFORM12

ADFGVX

CIPHER

WHEN

FROM

STATUS

In the spring of 1918 the German army staked the war on a series of great offensives, and it needed a field cipher that its enemies could not read in time. The answer was two old tricks stacked on top of each other: a square that splits every letter in two, then a shuffle that scatters the halves.

The machine

A square, a keyword and a Morse key

First each letter or digit is looked up in the six-by-six square and replaced by its row and column: only the letters A, D, F, G, V and X ever appear. Then those letters are written in rows under a keyword, and the columns are taken out in the alphabetical order of the keyword's letters. Type a message and a keyword, sort the columns, and tap the result out on the key.

Break it yourself

Six columns from the front

You have captured the day's square, and the intercept has been cut into its six columns, in the order they were sent. Put them back in the keyword's order: tap a column's tab, then another, to swap them (or focus a tab and use the arrow keys). The letters on the right show what each row reads through the square.

ARRIVED · CIPHER CRACKED

The story

The cipher was designed by Fritz Nebel, a young German signals officer. The first version, with only five letters, A D F G X, and a five-by-five square, came into use in early March 1918, shortly before the first of the spring offensives. At the start of June a sixth letter, V, was added, and the square grew to six by six so that it could carry the ten digits as well.

The letters were not picked at random. In Morse, A D F G V X are easy to tell apart, so a tired operator on a crackling field radio was less likely to send or hear the wrong one.

Why it was strong

The square alone would be a simple substitution of letter pairs, easy to break by counting. The transposition alone could be undone by trying column orders. Together they are much harder: the two halves of each letter end up in different columns, far apart in the message, so neither counting nor anagramming works directly. Codebreakers call this fractionation.

The radiogram of victory

On the French side, Lieutenant Georges Painvin of the army's cipher bureau attacked the new system from the first weeks. In early June 1918, with German troops close enough to Paris to shell it, he broke a message that the French later called the radiogramme de la victoire. In our words, it told a unit to hurry up the supply of ammunition, by day too if they could not be seen. Where it was sent from helped show where the next blow would fall, and when the attack came near Compiègne on 9 June the French were ready for it. Painvin is said to have lost about fifteen kilos during those months of work.

How it was broken

Painvin did not have a general method. He needed luck in the traffic: several messages sent on the same day with the same keys, especially ones that began alike, so that the column lengths and the patterns of letters could be compared. From those he could rebuild the order of the columns, and then the square fell to frequency counting, because once the columns are in place it is only a substitution of pairs.

The puzzle above skips the hardest part by handing you the square. Even so, you have to rebuild the column order, which was always the first step. Painvin and Nebel met in 1968, half a century after the war.