# Are [word-property] puzzles similar to poor ciphers where you must break a random code?

Take a look at this sample question:

If a word conforms to a special rule, I call it an X/Y/Z Word™.

Use the following examples below to find the rule.

Examples are then listed below, and that's it.

Take a look at another sample question:

Below is a cipher:

ABCDEFGH

Can you crack it? Good luck!

And that's it.

What is the difference between these two types of questions?

Should posts be allowed just because, as another user once said, they stood both the tests of time and votes? But then I ask, what about the test of puzzle quality?

Although it is a bit subjective, and can be improved, we all know the above cipher wouldn't pass it. Ask yourself why. Now, can't you use the same reasons for puzzles too?

• Hints for these types of puzzles are usually in the title or the X/Y/Z.
– user14478
Nov 28 '16 at 0:15
• @LukasRotter If answering the question requires the title, then why not add that into the main body, and have another title? Nov 28 '16 at 0:18
• It is also in the main body. "If a word conforms to a special rule, I call it a X/Y/Z Word™." (the X/Y/Z is the clue)
– user14478
Nov 28 '16 at 0:20
• To be fair, not all simple ciphers are bad posts, and the same way, not all [word-property] questions are good ones. Nov 30 '16 at 22:32

## No, they aren't.

Firstly, despite what you suggest, the test of time and votes does correlate with the test of puzzle quality. Even back in the spaghetti era when low-quality puzzles were often getting a lot of upvotes, a succession of puzzles of the same type would usually have strictly decreasing scores unless some of the later ones were really really good. There's been a lot of complaints here on PSE meta about how bad puzzles don't get downvoted enough, but frankly I don't think the situation is that bad, at least not nowadays. If the puzzles were really as bad as all that, they wouldn't have lasted as well as they have.

More importantly, here are some reasons why the puzzles are objectively better than the terrible cipher puzzles so eloquently discussed in Code Puzzles: What (Not) To Do?

• The title is usually a major clue as to the nature of the puzzle.

One part of what makes really bad cipher puzzles bad is that no clue is given as to what kind of cipher we should be looking at: it's just "solve this cipher, boom". If there's a title which hints at what kind of cipher it is, then the puzzle is automatically improved.

By contrast, when a puzzle looks like "What is a Supercalifragilisticexpialidocious Word™?", then whatever adjective comes in there is somehow descriptive of the answer. It shouldn't be enough to solve the puzzle on its own, but it might be enough to guess at the general ballpark of the solution, or to know when one is on the right track. It's almost like a mini-riddle or cryptic clue.

• There's a lot of pattern and context in the words themselves.

With a bad cipher puzzle, all we're given is a string of nonsense characters, and one kind of nonsense looks much like another. The setup of the puzzle doesn't give us any clues in itself. the only way of hitting on the right cipher is by trial and error or bugging the OP for hints. If there is something in those nonsense characters which gives us a serious hint towards the nature of the solution, then that's when it stops being a bad cipher puzzle and becomes a decent one.

In contrast, the list of words in a puzzle should contain the answer right there, if we just look at them the right way. What kind of words are being included in the list? Do they all contain certain letters? Are they nouns, verbs, adjectives? Are they singular, plural, conjugated forms? Which words on the left and right sides are most similar to each other, and what does this tell us about what the property must be? Asking these questions is a good way to attack this kind of puzzle, and asking enough of them often leads to the answer.

• The answer is easily self-confirming.

With a bad cipher puzzle, even once we've made a decent guess at the nature of the cipher involved, we usually need to do a lot of tedious calculation in order to confirm or refute our guess. There's no way to tell instantly whether our guess is right, except by applying it to the ciphertext and checking whether it gives something meaningful.

On the other hand, puzzles can mostly be solved without making any calculations at all. Once we have a guess as to the answer, we can easily check it by looking at whether it applies to all the left-hand words and doesn't apply to all the right-hand words. This should be doable mentally, without needing any computation, and if the guess seems valid, we can post an answer straight away.

• I wouldn't say "objectively better", that needs an objective measure of what "better" is, but let me respond to your points: Like I said before, why not add this title in the main body, and have another title? If the title is a hint to solve the puzzle, then the hint should be in the puzzle, not in the title. As for patterns and context, that depends. An ideal, perfect cryptographical text would be essentially indistinguishable from random text, and on purpose. Cipher puzzles need not have patterns in the cipher itself, that restricts the nature of the puzzles for no reason... Nov 28 '16 at 0:47
• .... And spoilered hints below the cipher are a better option. But then, the hint must not be required to solve the cipher, after all, if it was required, then it should be added to the main puzzle. Nov 28 '16 at 0:47
• @CipherRiddle The adjective that appears in the title is in the body of the puzzle as well. If the words are Supercalifragilisticexpialidocious Words™, then Supercalifragilisticexpialidocious will appear in the body as well as the title. And yes, an ideal enciphered text would be indistinguishable from random text and therefore uncrackable. Ciphers posted as puzzles here on PSE have to be crackable by design. (However, I'd rather not get into a long comment debate about it, but rather let the points in my answer stand to be voted on.) Nov 28 '16 at 0:49
• If the title also exists in the body, then perhaps another title could be chosen altogether, to avoid repetition. About ciphers, yes, a cipher puzzle posted on PSE must be uncrackable in some form or the other, but what I was saying is that a random cipher, that is uncrackable, with little explanation, say just a word or two and nothing else, that doesn't seem like an acceptable post. If I make an 'X/Y/Z word" puzzle, then essentially if no other hints are there, the 'X/Y/Z" part is basically the only hint provided. Who decides how many hints are acceptable? Setting a hard minimum isn't.... Nov 28 '16 at 0:53
• ....Isn't what I mean, but who decides how many hints are not enough, and how many hints are too much? Nov 28 '16 at 0:54
• @CipherRiddle: The voters decide. That's the entire point - people can collectively vote on the quality. In a [word-property] puzzle, the words themselves are hints as to how it is solved. As Rand said, you can look at them grammatically, alphabetically, visually, and a lot more. In a plain cipher, a random string is not a hint. You cannot gain any information from it unless you already know how to decode it. But by looking at the types of words chosen in a word-property puzzle, you can make progress.
– Deusovi Mod
Nov 28 '16 at 1:42
• @Deusovi I disagree that the words are a hint in the same sense of the word I was utilizing it in. The words are part of the puzzle, and the hint is separate from the main puzzle. Nov 30 '16 at 23:47