[WalterBright]

You're right, but it was still scant information with which to bet the fleet on. The Japanese might have suspected that their code was broken, and so used disinformation to mislead the US Navy.

Hell, it's what I would have done whether I thought the code was broken or not.

The Germans had plenty of evidence that Enigma was broken. The High Command refused to believe it. I would have used the broken Enigma to send the Allies into a trap.

The way to play the code breaking game is to assume the enemy has broken it, and act accordingly to your own advantage.

[mannykannot]

Even if you know one of your widely-used codes or cyphers has been broken, I don't think it is that easy to make use of that fact, except perhaps briefly and in a limited way.

To conceal the fact that you know that it is broken, you would need to maintain use of that code at similar levels as before, without approximately doubling the signal traffic by sending the real communication under a new code. Furthermore, the fake traffic under the original code must be realistic to the degree the enemy can verify it, as they can read it, and if a major code has been broken for a period of a few weeks or so, the enemy presumably has plenty of information to use in verifying new messages, at least for a while (the verification need not be explicitly performed, at first; if new messages seem to be inconsistent with what is already known, questions are likely to be raised.)

Compromised minor cyphers and codes are another matter, and that is exactly how the Midway ruse worked.

[sidewndr46]

For Nazi Germany the "fake traffic" would not be needed for all the services. Key change happened at midnight Berlin time by all operators. The radio operators stayed up late into the night sending the personal correspondence of the various officers to their families. The codebreaking process used this huge volume of messages to feed into the "cribbing" process which aided in recovering the traffic. By the time they had extracted enough of the key to decrypt traffic, normal military communications had started

[mannykannot]

Thanks - I was unaware of that until now. It appears to be a major operational security lapse.

[mannykannot]

Correction: I wrote ‘without approximately doubling…’ where I meant ‘while approximately doubling…’ - and then one must take into account sidewndr46’s interesting point.

[yowzadave]

How do you keep your allies from believing your fake encoded messages and taking the same action that they would have taken, had you not suspected the code was broken?

[WalterBright]

I'm not sure what you mean.

[sidewndr46]

It sounds like they are suggesting there is some sort of unusual 3rd order deception going on, but I don't know really.