The Zendian problem. Mapping the networks.

This post is about my attempt to solve the Zendian problem. It shows my initial work on mapping the communications networks used in exercise. Don’t click “Continue reading” if you want to avoid spoilers.

Mapping the communications networks essentially means figuring out who speaks with who and establishing relationships between parties. While by itself it probably will not reveal any secrets, the information might become useful later.

There was several approaches that i could take – i could write program that would group callsigns that are used.  Or i could try doing it with text editor and notepad. Looking over the messages one callsign jumped out immediately – ZZZ, so I wrote it down and started writing callsigns it spoke to around it.

To keep track of what i am doing, i made copy of the original message file and started cuting all messages involving ZZZ callsign to seperate file.

After I ran out of ZZZ’s I continued by searching the next callsign that was involved in communication, noting down any new callsigns that appeared. And so on until i ran out of callsigns.

There are few things to notice about this group (that i named group #1): ZZZ sometimes sends message to itself. Also MAL looks pretty lonely – having sent just one message to LLZ. They could have been destroyed, or maybe it was error of some sort. Hopefully i will find it out at some point.

As there was messages remaining in file I was cutting from, I began new file and started group #2. It took me a good part of evening, as most of communication and most of callsigns were in this group. However by cutting and pasting manually, I spent lot of time looking at messages and noticed  few things about them that can be used in further research.

– Stations tended to use the same discriminant (the first group of letters, as noted in wikipedia page) when sending traffic.
– When searching for OWN callsign, message number 00161 jumped out by having three instances of this combination of letters in the message body.  It turns out several messages have more than one OWN in them, interestingly most sharing EGBGE discriminant.
– In some messages there are mistakes – instead of XXX DE YYY (to XXX from YYY) there is XXX DI YYY and similar. They could be made either by missending or mishearing letter from morse or from making typo.
– The header with numbers tend to have pattern when station are sending messages out. Some numbers are increasing, indicating that it could be serial number.

I haven’t mapped group 2 out yet, but i finished rest of the smaller groups.