Pigeon Code Almost Certainly Not Broken

Sometime on Sunday evening, as I was wending my way home on the Metro, I started to get messages from people asking if I'd heard that the Pigeon Code had been cracked. As the cell reception along that particular stretch of line is not great, I had to wait until I'd got home to find out what all the excitement was about.

Sadly, as it turns out, the excitement seems to have been about not very much. The story seems to have originated in the Dorset Echo ("It's a real coo as 'unbreakable' war code found on pigeon in Portland is cracked") on Saturday 15 December 2012, and was subsequently picked up by the Daily Mail ("Hit Jerry's panzers here'... code on dead wartime pigeon is cracked") and then by the BBC ("Has World War II carrier pigeon message been cracked?") on Sunday 16 December .

GCHQ issued the following statement, quoted by the BBC

We stand by our statement of 22 November 2012 that without access to the relevant codebooks and details of any additional encryption used, the message will remain impossible to decrypt. Similarly it is also impossible to verify any proposed solutions, but those put forward without reference to the original cryptographic material are unlikely to be correct.

That's very terse. My response is going to be somewhat longer, it could in fact be summed up in words which are quite short, but that is both impolite and fails to show any working. Let's start with what I believe to be the primary source, the story in the Dorset Echo.

Quoth the Echo :

Last month the Echo reported that Mr Walbridge persuaded Government Communication Headquarters (GCHQ) to delve into the mystery of the bird’s message, which was found in a blocked chimney in Surrey.

The intelligence agency recruited Canadian researchers the Lakefield Heritage Research to decode it.

What was originally thought to be an unbreakable code has now been partially cracked using a First World War artillery code book.

And indeed, if we do a quick search for 'Walbridge' on the Echo site, we find the very reportage they refer to.

Retired government worker Neville Walbridge, 74, used his knowledge and links to ensure the historic document reached the Government Communications Headquarters (GCHQ) – but it has now been announced that the code is unbreakable.

Weston resident Mr Walbridge said: “At this stage GCHQ say that without the original codebook they can’t decode it.

"That will make a lot of people like myself laugh. They’ve had it all this time – they would’ve known that the first day they had it."

I'm going to let most of that stand without comment. Further into the article of Sat 15, we get :

Weston resident Mr Walbridge said: “I thought someone would come up with it sooner or later. After the war all the code books at Bletchley Park and the computers had to be destroyed. We accepted that. But I thought there would be a way to do it"

I'm not going to let that slip through though. It is indeed true that a lot of cryptographic material was said to have been destroyed by GCCS at Bletchley Park at the end of World War II. The veracity of this is, to put it politely, uncertain.

It is however, profoundly irrelevant. BP was engaged in providing intelligence by breaking Axis codes and ciphers, it was not, as many people commenting on the Pigeon Code seem to believe, some kind of government cipher communications hub. Any relevant code books would be the ones held in common by the sender and recipient of the message. THere are many things that are unknown about the pigeon code, but one thing tyhat we can be very sure of indeed was that the message was in no way intended for or on its way to, Bletchley Park. And no, Dorset Echo, GCHQ did not recruit "Canadian researchers the Lakefield Heritage Research". Sigh. Breathe deeply, move on.

Leaving aside some of the more hyperbolic statements there, lets concentrate on the code. The basic claim is that the message has been deciphered using a "First World War artillery code book." Given that this is a WWII cryptogram, this pretty much immediately fails the sniff test. However, let us take it at face value for a moment and have a look at the claimed decipherment, which reads:

Artillery observer at ‘K’ Sector, Normandy. Requested headquarters supplement report. Panzer attack – blitz. West Artillery Observer Tracking Attack. Lt Knows extra guns are here. Know where local dispatch station is. Determined where Jerry’s headquarters front posts. Right battery headquarters right here. Found headquarters infantry right here. Final note, confirming, found Jerry’s whereabouts. Go over field notes. Counter measures against Panzers not working. Jerry’s right battery central headquarters here. Artillery observer at ‘K’ sector Normandy. Mortar, infantry attack panzers. Hit Jerry’s Right or Reserve Battery Here. Already know electrical engineers headquarters. Troops, panzers, batteries, engineers, here. Final note known to headquarters.

Does this message pass a sniff test ? I have never been in a war, so I could be mistaken, but I'm fairly sure that calling in artillery fire is something you tend to want to do pretty quickly, without the delay of, for example, handwriting a message and then strapping it onto a couple of pigeons that will undoubtedly take several hours to reach their destination, in a different country, where the artillery batteries aren't. That of course assumes that the pigeon in question, good old NURP 37, wasn't just very badly lost.

Leaving that aside, assuming that avian carrier is a good way to call in the big guns, does this look like the sort of message an artillery spotter would send ? Again, I have no direct knowledge of this as an activity, but I would imagine that one would would need to be somewhat more precise than 'here' which along with "right" seems to be the only location information contained in the message. Let us charitably assume the message sender was suicidally heroic and his position was somehow known to the recipient, who could simply note that he was reporting "there is a shed load of enemy stuff here!" and subsequently drop several tons of ordnance in the general area of K sector.

Certainly research by Nick Pelling of the excellent Cipher Mysteries site suggests that the presumed sender, one Lance Serjeant Stout, was mortally wounded on the same day he sent the message. Research which seems to have been quoted but not attributed by many of the press stories this weekend, to the eternal shame of the 'journalists' responsible. Breathe deep, move along.

There are other oddities. Why is it important that the "Lt Knows extra guns are here" ? How is it helpful, given the inevitable comms lag of a pigeon flying through mortar fire, that the sender "Know where local dispatch station is" ? Will the recipient be sending a reply pigeon to ask for more specific locations ? Possibly, as "Already know electrical engineers headquarters" falls into much the same category. Or are those locations already known to the recipient, hence the stern insistence to "Go over field notes" ? Why are there not one but two "final notes" ? Why does the WWI code book supposedly used to decipher the message have a code letter for "Panzer", a variety of tank that only went into service in 1934 ?

There are some clues in the answer to the question : why is "Artillery observer at ‘K’ Sector, Normandy" repeated in an odd place ? I've already ranted a long while, so I'm just going to dump a big load of text here which illustrates the precise method of 'decoding' that the Canadian 'researchers' have used.


Artillery observer at 'K' Sector, Normandy.

Requested headquarters supplement report.

PABUZ [Mistranscription ?, I read U not I]
Panzer attack - blitz.

NLXKG [no decipherment]

UAOTA [ mistranscription ? I read U, not W]
West Artillery Observer Tracking Attack.

"Lt Knows extra guns are here.

Know where local dispatch station is.

HVPKD [No decipherment]
WYYNP [No decipherment]
MEMKK [No decipherment]

Determined where Jerry's headquarters front posts.

Right battery headquarters right here.

GQIRU [Mistranscription ?, this does not read HQIRH as far as I can tell]
Found headquarters infantry right here.

FNFJU [Mistranscription ?, decryptor (wrongly?) reads FNFJW]
Final note, confirming, found Jerry's whereabouts.

Go over field notes.

Counter measures against Panzers not working.

ONOIB [no decipherment]

Jerry’s right battery central headquarters here

Artillery observer at ‘K’ sector Normandy

YIDCC [No decipherment]

Mortar, infantry attack panzers.

Hit Jerry’s Right or Reserve Battery Here

Already know electrical engineers headquarters

Troops, panzers, batteries, engineers, here

Final note known to headquarters

Briefly then : (1) Conventionally, we would interpret ciphertext in a cryptogram such as this from left to right, rather than in columns as appears to have been the case here. (2) These code groups are not even internally consistent, sometimes headquarters is represented by H, sometimes by Q, whereas sometimes H apparently decodes to here and sometimes to hit, and so on. (3) Sometimes a single letter is used, sometimes groups. P for Panzer, A for 'attack' but (mistranscribed) BIZ for 'blitz', and so on.

Frankly, given that all the groups seem to have been (mostly) interpreted as simple acronyms, albeit with some notable exceptions which I'm quite sure are due to artistic licence, I would be surprised to find that the supposed code book even exists. Put bluntly, I think this is nonsense on stilts, though I will happily eat my humble pie if I turn out to be wrong. Perhaps that still isn't blunt enough. Someone simply made this up. They made it up and the Dorset Echo ran with it. Then the Mail and the BBC, who ought to be able to do at least few basic fact checks picked it up and ran with it. It remains to be seen whether the authors of this idiocy are just having the media on or if they are genuinely deluded. Sigh, breathe deeply, move along.

As I was typing this, I note that Nick over Cipher Mysteries has written a much politer rebuttal, proving himself to be much more of a gentleman and scholar than I.

UPDATE Mon 17 Dec 1720
I got some traffic today referred from a post at the Royal Pigeon Racing Association upon which there is a comment by the Mr Gord Young referred to in the press coverage. From the comments it is very clear that Mr Young himself never claimed to have broken the pigeon code. Nor does he appear to have made any of the other claims attributed to him. I will write something a bit longer later on.

UPDATE Mon 17 Dec 21:30
Paul in the comments points out that 'Panzer' is not just a type of tank, but a much more generic word, a bit like, er 'tank', in fact

There weren't many German tanks in WW1, but there are a couple that could have been called Panzers:
http://en.wikipedia.org/wiki/K... ("K" Panzerkampfwagen)
http://en.wikipedia.org/wiki/A... (Sturmpanzer-Kraftwagen)

Additionally, Captured British tanks during WW1 were called Beutepanzers.

I happily stand corrected. If anyone has read, or is reading, this, or any other post, and spots anything I have wrong, I am very, very happy to be put right. The more correct info we have, the more the chance of mounting an attack, or even just an analysis on the cipher system. It's probably a very small chance, but that makes getting good info lined up all the more important.

Codemonkey, iOS developer, neophyte entrepreneur, frequent hat wearer. Founder of Enigmatic Ape (@EnigmaticApe). Recidivist tea drinker. Sound tennis noob with NEVITC. Now available in Twitter, and ADN

  • MJ Simpson

    I know nothing about WW2 codes beyond the very, very basics and even I can see that this is nonsense on a stick. Is there any evidence at all of WW1 codes, more than 20 years old, being used during WW2? The transcription is obviously wrong. There is a clear W on line 2 which is utterly different to the other Ws which are plainly U (or interpreted in one instance as L I, creating the only six-letter group). Also wondering why a WW1 code would include WW2 things like Panzer, Blitz or indeed Normandy. There is a transcription of a radio interview with this Canadian saying that it’s not an encryption message because under the pressure of fire a signaller wouldn’t have time to encrypt and would use easily memorable codes – yeah, these are easily memorable all right.
    So basically if you take a message, ignore random chunks of it, change some of the letters, add a few and remove a few, you can then make it stand for anything. In this case for a long, garbled message that would have had no value whatsoever. Yet every paper is carrying this story unchallenged. Another triumph for British journalism.

  • Paul

    You’re wrong about a Panzer being “a variety of tank that only went into service in 1934″. Panzer is a shortened form of the word Panzerkampfwagen, which essentially means “tank” in German. It’s not a variety of tank, it’s practically a catch-all word used to describe all tanks! Germany built Panzers/Panzerkampfwagens before the end of WW1.

    • http://www.enigmaticape.com/blog Steve Trewick

      I stand corrected. If you can supply me with a link I will happily post an update with a correction.

  • Orbitni

    Totally agree with you. A request for artillery needs a swifter means of communication and also a method for quick decoding i.e. coordinates/bearings/distances etc. That kind of info is secret for only a few hours so an easier system would do.
    This message looks more like a one time pad type. Or perhaps a substitution followed by swapping row and columns? A statistical attack would need more material – or a lucky guess of a word or two. Some russian WWII ciphers used the latter, British WWII learned from experience that OTP was the way to go.
    And, of course, the suggested decoded text is just gibberish:)

    • jameslyons

      The M-209 has 6 wheels, which means the indicator would be 6 characters, whereas the pigeon message seems to have an 5 character indicator. The british were also using double transposition at the time, however the indicator was a sequence of 4 digits, not 5 letters.

  • Barry Traish

    Although that post from Gord Young doesn’t match the claims in the Dorset Echo (and subsequent stories), the post was made 25th Nov and the Echo story was 15th Dec. In that time, Gord Young has clearly done more work on the message and supplied additional details to the media (such as that he deciphered it in 17 mins), so we cannot actually tell what claims he has made, or, more likely, have been twisted by the press.

    Really good blog, btw. Thanks.

Built With Bootstrap
Powered By Wordpress
Coded By Enigmatic Ape