Challenge Help and Advice

If you are having problems on the site for some reason then you should Mail me and I'll try and sort them out. If you don't understand how the site works after looking around and messing with the scripts, etc then mail me and tell me how it could be improved. The rest of this page talks about general challenge advice.

I thought I may as well put up some help and advice and notes on where to start with everything here. If you crack substitution ciphers in a few minutes, solve mathematical problems and disassemble files by hand then you will find no help here.

So you just had a look at a couple of simple ciphers, used some program to crack a couple of them but don't really know what you're doing ? Read on.

Originally this site was quite specialised, focussing on cryptograms only, but I soon came to realise that my previous site C&C held some fascinations of its own with its difficult programming problems and reverse engineering problems and it had some great puzzles on it. This site aims to try and make a wider range of problems available to people in such a way that being stuck in one place won't thwart your advances elsewhere as it does on so many sites. In terms of help you will only find general advice here. Knowing just one specialised subject won't help as much on this site anymore and the top players will have a range of skills and knowledge.

First of all you should know your subjects, and research and some reading will get you a long way towards completing most of the puzzles. The harder puzzles just need more experience, intuition, intelligence and hard work. Trying to solve ciphers without knowing how they are constructed, what the weaknesses are and what the best approaches are is not the best way to work. Subjects covered in the problems on this site are cryptography, mathematics, programming, reverse engineering, steganography and logic. Some problems can be identified to specific areas and others require a more general understanding of all of the above subjects. There is plenty of research material available on the internet for these subjects should you think yourself to be stuck on a problem.

So what I am going to do here is to explain some of the ways that I would attack a cipher problem or crypt, and give people a few ideas, without just looking around for other peoples programs to run against them. Whilst it is nice to obtain someones program and use it to crack a few of the crypts you will soon find yourself up against something that the program won't just solve for you, and so I recommend learning to solve the simpler problems by hand. In this way you will obtain invaluable experience for cracking some of the problems later on.

OK. The first step I always take is simply to observe. Look at the crypt and examine it's features. How is it presented, what are the characters used in it, do you notice anything strange about it ? If you suspect it is a caesar then I would try all possibilities and carefully observe what comes out of it, be wary of missing a simple variation in the crypt through passing over the solution to it (and I know people do this quite often from the stats here, and from talking to people on IRC). Second step is almost certainly a frequency analysis. This gains us valuable information. What are the most frequent characters, what are the characters in use, what are the characters not being used at all ? Do you notice anything about the layout, should characters be paired together rather than singly ? Simple substitutions will dissolve at this point. You do know the letter frequencies of the language, don't you ? No ? Research it. Next it is always an idea to look at a contact count - which letters each letter appears next to, and whether it is before or after. Knowing the characteristics of Languages will enable you to deduce a lot from contact counts.

More advanced techniques, to give you a few ideas, could now be used where the problem seems intractable. Look for characteristics in the letters and where they appear, and words if word divisions are present. Can you guess a word in entirety, or is the number of possibilities very small - you do have a good dictionary file don't you, and some programming knowledge to pattern search it, etc ? OK, still stuck ? There is a very nice method of starting to solve difficult cryptograms called the consonant line method, there is digram analysis and trigam analysis, there is probable word (hint for polyalphabetics like vigenere - try encrypting your crypt with a probable word). I am not going to go into details on any of these, but I simply want to provide you with some terms so that you can go away and research them on your own.

If you want an excellent book which covers all of these subjects then I can heartily recommend CRYPTANALYSIS by Helen Fouché Gaines which is quite cheap and you can get it from online bookstores quite easily. In addition I always recommend people to read Lanakis tutorials which are available online, although they aren't as kind to the beginner as the aforementioned book.

Hopefully this gives you some ideas, but I can only repeat that you should observe and analyse. Without these two basic skills you will soon be left behind. Ensure you have some good frequency charts and lists of common words and digrams, and take heed of the title and comment on the crypt when you download it.

Hopefully this gives a few ideas of just how many ways there are to solve problems. If you get really stuck on a problem then leave it for a few days and come back to it later, you will often find a new insight and solve it quickly. Good luck in your problem solving and see you at the top of the board ;P


Site design © Cronos Sept 2001.
All scripts and programs © Cronos Sept 2001.