# Design of enigmatic puzzles

I've been trying to practice my enigmatic puzzle design skills, since these are the kinds of puzzles you find in many video games. (That is, puzzles in games don't often come with explicit instructions, unless the game is built around that specific puzzle.) However, it's clear my ability to assess the difficulty of my puzzles is a bit off.

Let's look at two of my attempts. In the first puzzle there was a piece of paper with some letters and clues on it, but no instructions. I knew it'd be too difficult, but couldn't figure out a good way to hint at what needed to be done without mostly giving away the enigmatic part. And indeed, when I posted a similar puzzle with explicit instructions, it was solved in about 30 minutes.

In my second attempt, I kept it much simpler, and left a lot more hints up front, but it was only solved after I left a very explicit hint at what needed to be done.

So, my question is, how can I better lead the solver through an enigmatic puzzle without just giving away the logical leaps that make it enigmatic? Answers from things to avoid, to better ways to put yourself in solver's shoes, to better ways to make use of the tools on this site are all valid.

• I think Stephen King's advice for writers is key here too: If you want to be a better puzzle creator, you must do two things above all others: solve a lot of puzzles and create a lot of puzzles. – Bass Aug 23 '20 at 23:06

It's hard to judge the difficulty of any puzzle you know the solution to. This means that

1. it's understandable that you're having trouble with this, and
2. my analyses here will be with the benefit of hindsight, so take them with a grain of salt.

That being said, I did attempt both of those puzzles shortly after they were posted, and didn't get anywhere at all in either. In general, you should indicate arbitrary decisions, and rule out incorrect paths. Here's where I got stuck, and how I think you could've fixed this in the design of the puzzles.

Mysterious Paper

When I first tried your first puzzle, with the torn sheet of paper, I had the right idea that the sequences were indicating relative positions! But I tried to arrange them in a linear order rather than a circle. I tried merging various combinations of the trigrams this way; when the first column didn't work, I tried it allowing trigrams to go both ways.

One of the things that stopped me from continuing and eventually stumbling upon the cyclic ordering was the arbitrariness of the trigrams. Generally, in puzzles, it's important to think about what the puzzle creator could have done, and what constraints they were under. The most common place this pops up is in orderings: if a list of things (words, clues, band albums, whatever) is given after being sorted somehow, that usually means that the order is unimportant. If there does not appear to be an external order to them currently, then the order is there for a reason.

From Introduction to Puzzlehunts, by betaveros:

If the clues are in alphabetical order, or if the answers to the clues are in alphabetical order, that is usually a signal that the order the clues were provided in isn’t important, so you need to find the correct ordering of answers or the correct way to assign answers to something else. Otherwise, you probably should expect to use the ordering eventually to extract something.

Here, both the letters in each trigram and the trigrams in the list were not obviously ordered somehow. This signals to the solver (or at least, signalled to me) that there's some reason that both of those are ordered as they are. It discourages people from interpreting things cyclically, or trying interpretations that ignore the current ordering of the trigrams.

If you had cycled each trigram so its first letter alphabetically was at the first position, and then sorted all the resulting rows by their first column, that would be a huge improvement in communication. It would send the messages:

• The order of letters in the word is unimportant, but the cyclic order is (because some trigrams would have letters ordered like "132" rather than "123").
• The order of the trigrams in a column is unimportant: what is important is the columns.
• Each trigram in columns 2 and 3 should be associated with the trigrams to its left for some purpose.

This would help rule out many of the wrong ideas, while still keeping the aha moment of "these letters need to be placed in a circle to satisfy all these conditions", and the subsequent replacements for columns 2 and 3.

(Note: I see that the puzzle image was changed since I first tried it. I hadn't looked at it since that change, so there's a chance that seeing that might've prompted me to retry what I had attempted earlier.)

Equally Shaped Words

Once again, we run into the problem of the first row not being alphabetized. There's no reason that, upon looking at this puzzle, someone would believe that the second row is correctly ordered while the first needs to be reordered.

"Two adjacent words share at least one letter" is pretty difficult to get out of such a small sample size. Two words chosen randomly are pretty likely to share a letter, and it's not a pattern that really jumps out at you. This means that there's not a lot pointing to that being right - if a solver thinks of this option, they may very easily discard it.

Besides that, there's another arbitrary decision: the use of . and - to space out the rows. I was absolutely convinced that this meant it had something to do with Morse code. I tried all sorts of manipulations on both dots and dashes, and inverting/rearranging the code units (where a dot is one unit and a dash is three), all to no avail.

So, I think the main problem here is that both of these puzzles didn't give the solver confidence that the right path was actually correct. There seemed to be things pointing away from the correct answers in both of them, and as a result they ended up being a bit more "read-the-creator's-mind"-y than you had intended.

It may help to consider a few things:

• How do you expect people to find the correct solution?
• What could point them away from that path? What else could those clues indicate, besides the correct thing?
• Once people have the right idea, what have you done to indicate that it's correct?

You'll never be able to eliminate all false paths without directly telling your solver what to do. People will always be able to find ways to get the wrong idea; what's important is that you rule out as many of those as possible, and make sure that when people do get the right idea they'll be confident in it.

For more on "obscurity" vs "difficulty", I'd recommend Introduction to Writing Good Puzzle Hunt Puzzles , parts 3a and 3b. Some other parts of that guide could be helpful as well.

• Good advice. I'll have to think on how that could be applied. The dots and dashes were meant to be a clue that the second row was connected (as in lines in a graph) in a way the first wasn't. Also, do you know of any way to suggest that the order might be important later? In that puzzle, the order of both rows is important for different reasons. – Joshua Taylor Aug 24 '20 at 0:59
• @JoshuaTaylor None that don't add anything extra to the puzzle. The first option that comes to mind would be to present them alphabetized, but also label them with a number from 1 to 6. If I saw that I would probably think something like "there would be no point in the numbers if they had to be applied instantly, so they must be applied later". (It would probably make it even clearer if you placed blanks numbered 1 to 6 at the bottom of the puzzle, but that may or may not be too much for your tastes.) – Deusovi Aug 24 '20 at 1:06
• But I would've just gone with the answer being spelled out left to right -- the constraints aren't that severe that you should need that reordering step. – Deusovi Aug 24 '20 at 1:06
• A great answer. I too was COMPLETELY convinced the solution to the latter puzzle would involve Morse Code - I spent a LOT of time looking at it this way and that, and in the end had absolutely nothing to show for it... Remove red herrings and it is much easier to encourage somebody to find the intended path. – Stiv Aug 24 '20 at 8:15

@Deusovi has already provided an excellent answer to this question, emphasising that you should (i) indicate arbitrary decisions, and (ii) rule out incorrect paths. I agree entirely with both of these but would like to add two more helpful suggestions - a couple of things which often help when trying to navigate through a multi-step enigmatic puzzle:

(iii) Leave breadcrumbs that show the solver they are definitely on the right path

One of the puzzle types I find most frustrating to solve is what I would call a 'multiple encryption' puzzle - when a puzzle setter says, "Here's something I've encrypted. To make it harder I encrypted it five times with five different methods - can you work out what they are?!" To which my answer usually is, "No I can't, and I don't really want to try because it isn't going to be any fun at all..."

However, a setter can make a 'multiple encryption' puzzle much more fun by making sure at each completed step of the decryption process there is a little hint that shows the solver they're on the right track - e.g. once the user worked out they were supposed to use rot-13 on the first ciphertext, although most of it appears to be nonsensical the first few letters of the new ciphertext is "VIG", at least giving the solver a fair chance by suggesting that what follows after this might be a Vigenère cipher.

In my experience, breadcrumbs (clues) like this have exactly the same effect in a multi-step enigmatic puzzle: they give you little moments of confidence that the last thing you did was correct, allowing you to focus on what comes next.

To use one of my own recent enigmatic puzzles as an example (please excuse this moment of vanity - I do not intend this as self-publicity, just that I find an example which I fully understand might be most helpful in explaining this point to you...), this mechanism is employed throughout "Which country did my sister visit?" (WARNING: SPOILERS INCOMING!)

All the information needed to solve this particular puzzle is present in the diagram provided in the puzzle (see the link), and there are small prompts throughout the solving process to tell a solver that they are on the right track with their solution:

Firstly, all of the names listed in the seating diagram are those of international airports, either directly (using the name of the person they are named after) or indirectly (breaking down a geographical name into something which sounds like a name). Once a solver noticed that one or two were international airport names - either through recognition of some of the more 'obvious ones' (George Best, Marco Polo, ...) depending on exactly where they live in the world, or using a search engine to check up on a few - they would begin to realise that every name in the diagram shared this property. There's your hook - immediately, you're on the solution path, and you know it.

The next part involved a little knowledge and a little deduction/guesswork. Notably it relied on the solver realising that all airports have a 3-character IATA abbreviation. Knowing that If the solver translated each name into its associated code, and looked at the resulting grid, they might hopefully spot a couple of letter sequences they recognise as real (and related) words:

 A N M N A S P E W
C E W T L V T S R
H T V S J U O S R
G E O P A N A M A
D T I B H D P V D
M L L E B A N O N
E S L U N A O P S
F N G O U A P V E
B I K O T P P S O
I H R M W C C U C
L I M T I U N E U
J C K V C E W D H
K E B S O L D G F

Try and spot them yourself before you check the next spoiler! Can you see them? The two hidden words are:

country names - PANAMA in row 4 and LEBANON in row 6. Surely a benevolent puzzle setter has deliberately included these - it would be too big a coincidence to find both of these words present by accident, surely...? And after all, the whole theme of this puzzle is that we are trying to identify a particular country - this must be the right track...

But now the puzzler is thinking, "Okay, I've found two country names - what do I need to do next?" Maybe they think they need to find more country names hidden in the grid - but try as they might these remain the only 2 they can find, until they spot breadcrumb 3...

At this point many regular puzzlers might start to think, "Right, perhaps I need to do something to this grid before I can find more countries..." They might consider some kind of cipher - but then PANAMA and LEBANON were spelled out without any encryption, so rule that idea out... Maybe they need to rearrange the grid in some way to make it more easily readable... and that's when they spot the breadcrumb...

Every row in this puzzle begins with a different letter of the alphabet (A-M)! Perhaps then we need to rearrange the rows in alphabetical order! And sure enough, when we do so, we get the following grid, in which three more country names can be found:

 A N M N A S P E W
B I K O T P P S O
C E W T L V T S R
D T I B H D P V D
E S L U N A O P S
F N G O U A P V E
G E O P A N A M A
H T V S J U O S R
I H R M W C C U C
J C K V C E W D H
K E B S O L D G F
L I M T I U N E U
M L L E B A N O N

Can you see them?

LIECHTENSTEIN can be seen going up column 2, MOLDOVA goes diagonally down to the right from the top row, and SWEDEN goes diagonally down to the bottom right corner.

But how does this help us to identify just one country for the final answer?

Consider the shape that these words make in the puzzle - this is particularly distinctive. And a geography buff who has made it to this point of the puzzle would hopefully notice that this shape is uniquely the shape of the flag of NEPAL! But is there a way to confirm that this is our final answer?

Well, the flag of Nepal contains two symbols: a crescent moon in the top triangle and a sun in the bottom triangle. Look closer at the grid and you might spot the words LUNA and SOL in these positions - and these words are Spanish for 'moon' and 'sun'! This breadcrumb acts as the final confirmation that the answer has been found, and the solve can rest happy in that fact and race to write up their results!

Breadcrumbs 5, 6, 7, 8 and 9:

But hang on a moment, because maybe a puzzler might not make the necessary connections between each step here - after all, I'm the OP, so I understand the process because I devised it... but would somebody else, coming in blind? Let's make sure and scatter some breadcrumbs to lead the solver to the other breadcrumbs!! This is the purpose of the second part of the puzzle - the note from 'my sister'...

5. I got to know all of my passengers really well - a hint to look up the passengers' names, which would lead to the realisation that they all shared names with airports.

6. I even remembered to take their in-flight meal orders at the right point of the journey - the key word here is orders. If they spot this, the solver will now be primed to think about reordering some component of the puzzle at the right point.

7. We flew through the airspace of five different countries... - at some point in the puzzle we will come across five (other) countries. These are the 5 in the wordsearch.

8. ...(each of which shared the same specific connection with the country of my final destination) - this is an extra clue to help verify the identity of the final country. What do the 5 countries in the wordsearch have in common? They all have exactly two international borders with other countries - this gives us a clue that our final answer also will...

9. and just before we landed I even got time to practise a little of my Spanish (even though they don't speak it there) - here's a little nod to note the two Spanish words hidden in the wordsearch too (LUNA and SOL), and as a bonus an extra clue to the identity of the country: Spanish is not an official language there.

There was also a pair of bonus breadcrumbs, much more obscure, and only really noticeable (if at all!) by PSE regulars.

"Cryptic Family Reunion" is one recent puzzle series on PSE by @JeremyDover. As a small homage I concealed two clues in the vein of this series (where a set of cryptic crossword clues have their definition part replaced by the name of a generic family member, thereby concealing the theme of the puzzle). These can be found in the very first line of the puzzle - {My sister} is always talking about pilots - and in the sister's dialogue - In short, I appreciate the attempt, {brother}. Both clue the initials IATA (via 'pilots' and 'in short', implying first letters - possibly a little obliquely), which would help the solver to realise (if they haven't already) that airport codes are involved.

Finally, if in any doubt that the solver had found the correct answer, they might notice that:

NEPAL is an anagram of PLANE, which ties in with the whole theme!

So all in all, this puzzle contained eleven hidden breadcrumbs to help guide the user to the answer. It perhaps did not require that many (and indeed, the solver did not use all of them in finding the answer) but in that case any left over at the end become fun little 'Easter eggs' for people to find.

The overall message here though is that what could have been a very difficult and obscure enigmatic puzzle was made much more doable by scattering breadcrumbs throughout it!

Finally (and briefly), one last piece of advice:

(iv) If possible, use other PSE question tags in addition to to help point out useful features of the puzzle.

Note that in the example I gave above I also provided the , and tags. The first of these was perhaps obvious, seeing as we were looking for a country as the solution, but these other two might have proved very helpful to solvers...

• The description for is "A puzzle in which certain letters, e.g. of each word, line, or paragraph, form a meaningful word or words when put together." This was used to emphasise the first letters of each line of the wordsearch grid and prompt the solver to spot the pattern amongst them.

• Meanwhile, the use of would have remained in the back of the solver's mind throughout, and would have proved useful right at the end of the puzzle by noting it had not yet been used. This tag was, of course, pretty crucial to finding the final answer!

## Take-home message:

Breadcrumbs and tags are useful tools for any creator of enigmatic puzzles. They enable you to hide confirmatory clues to your solvers to help them realise they are on the correct path, keep them moving forward, and - most importantly - keep making it fun! Those little moments of recognition that things are moving in the right direction can give a solver quite the adrenaline buzz - and as a puzzle setter, bringing those moments to your solvers is really your ultimate goal...

• Also very good advice! I mentioned the "breadcrumbs" in my last section, but didn't focus on them as much as the other two. This is a great example of how confirmation for the solver can be done well. – Deusovi Aug 24 '20 at 15:56
• @Deusovi Thanks :) And perhaps it's unfair of me to summarise your answer as just the two points I mentioned - you definitely covered more than those in an overarching way. Either way, hopefully between us we've now given the OP plenty of ideas...! – Stiv Aug 24 '20 at 16:15