Hi there! I am DiMono, creator of the currently most difficult puzzle on Picross Touch, Guaranteed Unique, seen in this guide's thumbnail at an accurate early state of progress. I enjoy making difficult puzzles, and even more than that, I enjoy solving difficult puzzles. I've seen a few puzzles in Picross Touch that are genuinely difficult, and I've seen lots that aim for difficult but end up just being tedious or boring.

In this guide I'm going to walk you through how to create a puzzle that is not only difficult, but interesting. I'll also show you what pitfalls to avoid in making puzzles, so that people hate you for the right reasons.

Let's get started.

There are certain puzzle types that everyone hates. The reason for this is a combination of the puzzles being tedious, and there simply being so many of them that solving them is genuinely boring. Some examples of those are below:


If you make levels like this, people will wish you harm for wasting their time. You may think it's interesting - even I thought there was some merit to the third one at one point - but the simple truth is that people don't like boring levels.

Not only that, but the dev has been known to actively search for and remove levels like those pictured above, because they are a huge waste of everyone's time. So keep that in mind as you create your levels: if you make ones that suck, they might disappear when you're not looking.

Another type that everyone hates is puzzles that just draw a letter. It's okay to include some text in a larger puzzle, but if your entire puzzle is just a single letter, you have wasted the solver's time. As well, if you need to add a giant border to your puzzle to fill up the space, then you really should have just used a smaller puzzle size in the first place.

QR code puzzles are also a waste of everyone's time, including yours. Nobody who solves the puzzle is going to care about your YouTube channel, or whatever else you're linking to. QR codes are random noise, and that's just not fun to solve. It's tedious and boring, and everyone hates it.


Those are the things you should avoid. So what should you put into your puzzles then?

What people want to solve is a puzzle that draws a picture of some sort. Don't give us the word TREE, give us a drawing of a tree. Don't give us a pattern you think is interesting (at these sizes, it's probably not), give us something the pattern might be used in. The "Pic" part of Picross means Picture. So rather than trying to give us something clever, give us a picture to reveal. It's that simple.

Also, give it a name that's in some way useful. If you made a picture of a trophy, don't call it blablabla, call it something like prize, or award. Ideally, the title should describe the picture without simply saying what it is.

The easiest way to describe what makes for a difficult level is to look at what makes a level easy, and do the opposite of those things. We can list those things very simply:

In Picross, a line is considered "complete" if there is only one way to fill it in. For instance, in a 5x5 puzzle, if the clues for a line are 2 2 then you know there is only one way to draw that: two on the left, and two on the right. A line is also considered complete if there is only one number in it, and that number is the same as the side length. As a solver, complete lines are the first thing I look for. Here is a small example of this in action:


This puzzle was obviously not meant to be a difficult one, but it fully illustrates just how easy having lots of complete lines makes a puzzle. This is especially true if the border contains complete lines, because the solver can usually get a lot of information very quickly from a complete border. The more complete lines your puzzle contains, the easier it will likely be to complete.

This one should be fairly self-explanatory: the bigger a number is, the easier it is to draw it on the board. In a 25x25 puzzle, it's easier to get parts of a 19 on board than a 7. The faster you can get squares filled in, the faster you get information on how to solve the puzzle, and the easier the puzzle becomes.

This is related to having complete lines, and the two frequently happen together. Here is an example of a puzzle made easier by having long segments and complete lines:


The above puzzle is a 20x20 that is meant to be hard, but there are 9 full-length segments, an 18, a 15, an 11, and an 8 on a line that quickly shortens as other squares get filled in. This results in a faster solve time, because there is less to do. The more long segments your puzzle has, the easier it will be to solve.

Related to this, if you have a line with only one segment, that line will be much easier to solve than a line with multiple segments. If you are solving a 25x25 puzzle, and one of the lines has only a 3 on it, then as soon as you get any square on that line you can immediately clear out at least 20 squares on that line, because they will be too far from the square you have for them to be used in the solution.

Symmetry is when you can flip or rotate the puzzle and get the same thing. This commonly happens with patterns, and top views of things that are symmetrical, like bugs or swords. The reason this makes a puzzle easy is that once the solver realizes the solution is symmetrical, they only actually have to solve half the puzzle, and then they can copy that to the other side.

As a solver, this is the second thing I look for. If a puzzle is symmetrical, then I can use that knowledge to immediately fill in everything that crosses the middle of the puzzle, and then for the rest of the puzzle I can duplicate anything I do on one side to the other.

Here is an example of a puzzle that was supposed to be difficult, but was actually very easy because of its symmetry:


That is a 20x20 puzzle with lots of small parts. Under most circumstances, that should be a fairly difficult puzzle. However, as you can see from the solution, it is symmetrical in every way possible. You can flip it horizontally or vertically and get the same design. Instead of taking a long time to solve, I finished it in just over 4 minutes. If your puzzle is symmetrical, it will probably not be hard.

Now, there are always exceptions to the rule, and one of those is the puzzle What25, by Daniel. It is 180º rotationally symmetrical, but still has an average solve time of almost 50 minutes. As of writing this guide, it is the second-hardest puzzle in the game.

The end result of all of this is that a hard Picross level will have a lot of whitespace, mostly short line segments, and will usually not be symmetrical. That's not to say that the entire puzzle must conform to these guidelines: it's perfectly fine to have a puzzle that starts really hard and gets easier after a certain part is finished. One hard part in a puzzle is sometimes enough to make that puzzle hard. But you will want at least a large portion of your puzzle to follow these ideas.

So you've created a level, and you think it's difficult. How do you test it without setting it free in the world? Picross Touch actually gives you a way to test this right in the editor: the validity checker.

The validity checker works by going line by line, row by row, trying to solve the puzzle. If it has a unique solution, the checker will find it. This ensures that only puzzles that can be solved make it into the game, and it is a very good idea.

Since the validity checker works procedurally, that means the longer it takes to validate a puzzle, the more complex the solution is. Basically, the longer it takes to validate, the harder it is. My puzzle "Guaranteed Unique" took over 5 minutes to validate. If the validator solves your puzzle quickly, then it is not a hard puzzle.


Now, that doesn't in any way mean it's not a good puzzle. In fact, many of the most enjoyable puzzles in the game are of medium difficulty or less. But if creating an evil masterpiece is your goal, a fast computer solve time is not what you're looking for.

When making a difficult puzzle, you should also expect to be there for a while. The puzzle in the screenshot above took about a dozen tries to get right, and it validates in about 2:20. Each time I make a change, I need to revalidate it to make sure I didn't break anything, which means another two and a half minutes or so. Every time. It took me two hours to make "Guaranteed Unique."

Now you have to ask yourself a question: "Should I publish this puzzle?" The obvious answer is "Of course I should! I did exactly what I wanted to do, and now I have a hard puzzle. Why wouldn't I publish it?" The reason you might not want to publish it goes back to one of the first things I wrote here: people don't like doing puzzles that are tedious. Your puzzle may be hard, but if it's more tedious than interesting, people are still going to hate it. If the most interesting thing about your puzzle is that it's hard or time consuming, then you probably shouldn't publish it.

Fantastic! The hard puzzle enthusiasts like me will be very pleased if you're right. Now there's one thing left to do: pick a name. A quick word of advice: don't say the puzzle is hard in the title. The reason is that that gets people's expectations up, and if it turns out you're wrong, they're going to be disappointed, and may judge you harshly for it. I encourage you to search the workshop for puzzles with things like "hard" or "difficult" and see how hard those puzzles actually are. A couple of them live up, but most are actually incredibly easy.

As I mentioned earlier, your title should be something that describes the finished product but doesn't spoil it. One of my favourite hard puzzles is called "Brick By Brick" and is perfectly titled for what it is, despite the title being relatively vague. A good title really does make a big difference.

On the other hand, if you feel like you have to lie about the puzzle in the title to make people want to solve it, then you have made a bad puzzle and you should not publish it. I have seen this in the workshop for this game, so it is a thing that happens sometimes. Don't do it.

If you'd like some practical examples, to see the things I'm talking about in action, I have been favouriting all the puzzles that take me a long time to complete given their size. The criteria I'm using is as follows:

• 5x5 - First solve time more than 10 seconds
  
• 10x10 - First solve time more than 1 minute
  
• 15x15 - First solve time more than 7 minutes
  
• 20x20 - First solve time more than 14 minutes
  
• 25x25 - First solve time more than 25 minutes

Some of the levels I've favourited are good. Some of them are boring and tedious. I will leave it to you to decide which are which, and to use that knowledge to help guide your own level creation.

Above all else, remember that this is a game, and games are meant to be fun. So have fun!

Happy level designing!