The title says it all. Is it harder to decrypt a secret message that has some Letters added to it or one that has some removed , like:

Heerlollo, ibtq'is meen becomes the same as Hlo i' m:

Hello, it's me In one case you have to remove every second letter (but you have to find that out first), in the second one you have to find every second char.

Let's assume you have only one message at the time.

I really don't know which type is harder, I googled for it, but didn't get any usable results.


closed as too broad by boboquack, Glorfindel, APrough, Rand al'Thor, Beastly Gerbil Nov 29 '17 at 19:15

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ What is the actual question here? Is this an actual puzzle, or merely a question about which cipher is harder to crack? $\endgroup$ – APrough Nov 28 '17 at 21:35
  • 2
    $\begingroup$ I think this question is too broad - it depends heavily on exactly how the added or removed letters are chosen, and you don't specify. $\endgroup$ – Deusovi Nov 28 '17 at 21:41
  • $\begingroup$ @Deusovi I think your comment is partly an answer and the question is „broad“ but not in a bad sense. I think it is a legitimate post on site, although it falls out of the ‚typical‘ ones. There might not be a fully satisfactory answer to it, who knows, but it is the type of question people might google and come to SE for. I would vote against ‚closure‘ and let it live here. $\endgroup$ – BmyGuest Nov 30 '17 at 7:53

Well, for the first type, you have the entire message, you just need to remove some noise. For the second type, you're missing half the message. There are near infinite possibilities to fill it up, especially if you don't know whether there's a pattern to the removal. In the given case, it's just every second letter, which can be easy to fill up, but if random letters are removed you can't even determine where you should look for words.

So long story short, I'd say it's almost objectively possible to say that the second type is significantly harder.

  • $\begingroup$ This is a very good answer, definitely worth an upvote. Actually there should be a pattern in the missing letters, else it would be impossible to decrypt the message! $\endgroup$ – monamona Nov 28 '17 at 21:28

The problem with removing letters is that natural languages are very predictable, that is, the amount of entropy in a single letter of English is surprisingly low.

Shannon estimated that nearly 50% of the letters in English text are redundant.

This is roughly (almost criminally roughly) equivalent to saying that you can often recover the initial message even if half of the letters are removed.

"Se, t s't ey ad o ed hs."

"See, it isn't very hard to read this"

This can be made a lot more difficult by including non-letter character as eligible to dropping. However, this starts to lose information, since the 50% estimate depends on word boundaries and knowing the previous words in the sentence (IIRC).

"Nts ayayoe si?

"Not so easy anymore, is it?"

Now on the other hand, adding noise to the original plaintext leaves the whole message "hiding in plain sight". If you are systematic in adding the noise (like, adding one random letter between any two letters", your message can be trivially found by reading only every other character. You can get sneaky though, and add purposefully misleading letters as your noise. This will, of course, only confuse humans with their puny language processing neural networks (brains), which will latch onto any familiar looking letter sequence and try to interpret that as a word.

"Mounceht miocrael dripfefeincrublate"

"Mounceht miocrael dripfefeincrublate"

This is, however, almost trivial to read by using a computer, so you'd be better off adding the noise in a pattern only known to you. Keeping things fair, you should only use the same amount of noise, so this would leave longer portions of the original words visible. Again, you can select the noise to disguise things, so you should be pretty well off.

"Allumposente imaspiostastibarlies"

"Allumposente imaspiostastibarlies", original message split into parts of 2 and 1 letters, each part then followed by the same amount of handcrafted "noise".

So in summary, for a human reader, the added letters approach can be at least as difficult as the dropped one, while it maintains the original message better.

Neither approach will stand up to any cryptographic challenge though: dropping letters in a manner that can be recovered is basically just compression, and adding noise is, well, steganography, which will work given arbitrarily much noise, but only as long as the method is not known to the attacker.

And since this PSE we are talking about here, either approach can be made into an enjoyable puzzle. Well, not the "whitespace can also be dropped" case though, I suppose.

  • $\begingroup$ I see your point as weel, but I never wanted to make the whitespaces droppable as well. I have now another point of view to think about $\endgroup$ – monamona Nov 30 '17 at 13:33

Not the answer you're looking for? Browse other questions tagged or ask your own question.