Our Puzzle Philosophy

As I uploaded the latest batch of products to the store, I noticed the number of unique puzzle products we’ve made has surpassed fifty. Fifty! It made me realize that we learned a lot about how to create and present these puzzles, and how, over time, we settled on a “template” or “ruleset” for how we approach these.

In this article, I wanted to talk about that and explain it. This is hopefully interesting information for puzzle designers and enjoyers, but also for teachers/parents so they can use our puzzles in the best way possible.

What’s the idea?

At its core, a puzzle is a solo problem-solving activity. And whenever you solve a problem, you are training and exercising certain skills.

The basic skill that any puzzle tests is, of course, problem solving and logic. Most puzzles, though, have a secondary skill that they focus on and which makes that puzzle unique. Crosswords are about spelling and vocabulary. Sudoku is about spatial reasoning (rows, columns, boxes) and counting (1–9).

This description, however, is identical to how you’d describe an “exercise”, “test” or just “homework”. What makes puzzles different?

• You get rewards for doing them. The best puzzles give small rewards along the way (“I completed this part of the puzzle!”) and a big reward when done (“Hey, now that I’m done, an image appears!”)
• They have a very simple (core) rule set which is simply repeated for all the puzzles. The difference is in the map layout, or the setup, or the specific clues given—the basic rules, however, don’t change. Learn them once (in a few seconds), and you’re good for a while.
• There is freedom in how you approach it. Many puzzles have multiple first moves you could make, multiple areas where you could focus first, multiple ways in which you can get an “Eureka!” moment.
• There is creativity (and theme). Of course, the most famous puzzles are “abstract puzzles” without theme—just black numbers or letters in a white grid, usually. But that’s what we’re here to change! Puzzles can easily support a story, a theme, more color and illustration.
• Puzzles are at their best as step-wise exploration. You can’t solve a puzzle unless you understand the trick behind it. As such, every next puzzle should show a different trick. “Oh, you learned that you can do X to solve a puzzle that looks like that? Well … what if the puzzle looks like THIS instead?”

This makes them ideal teaching resources, doesn’t it?

• The rewards make them more fun and make kids more motivated to do them (a lot).
• The simple rule set means you only need a very short “teach” or “lesson”. And you might get hours of puzzling and problem solving out of that. (Something that kids can even do on their own once they understand those rules.)
• Creativity, problem solving, agency (making choices; making smart choices) are things you definitely want to encourage all the time. And things our current systems of education, unfortunately, do not.
• You can’t easily “cheat” at these puzzles. Answers can’t be looked up online or easily copied from someone else. It makes no sense to continue to the next puzzle without solving the current one, because you’ll just be extra stuck on the next one. Going “step by step” prevents overwhelming and allows people to go at their own pace.

These are the things we look for when designing our puzzles. We are not interested in making a slightly modified version of homework. We’re not interested in explaining a set of rules, and then just dumping 50 randomly generated puzzles afterwards (with no changes except for making them BIGGER and BIGGER), and that’s that.

Instead, we design puzzles in multiple “levels” or “difficulties”. The first level should always just be 1 or 2 rules. Both the rules and an example puzzle should easily fit on a single page. The same is true for later levels: they only add 1 or 2 clever simple rules, such that the explanation and a first teaching puzzle fit on one page.

An Example

Consider Sudokus. There are many people who solve a sudoku every day, usually one from the newspaper or one from any number of popular websites that give you a randomly generated Sudoku each day. I’m not saying people can’t have fun this way or that this is bad. Of course not! People love doing this, and it’s absolutely better for your brain to do a Sudoku a day than to not do it.

But such players are not “growing”. They usually settle on a difficulty they like (“3-star sudokus are fine, 4-star seems impossible”) and just keep doing that forever. My dad has been solving Sudokus for ages, but refuses to do anything above 3 stars (in difficulty). When I started doing the newspaper Sudoku too, I surpassed him in only a month or so. Not because I’m a genius, but simply because I went out of my way to find harder Sudokus to keep challenging myself. I started doing “Killer Sudokus”. They felt impossible at first; now they’re easy too. I even started writing with my left hand to up the difficulty ;)

Educational puzzles are, well, educational and meant to stimulate growth. That’s why our puzzles will only stay at the same difficulty level for so long before adding a new twist or variation to challenge you in a new way. In our experience, most rulesets need 5–10 unique puzzles to fully explore it and make sure the player understands the different aspects.

Research shows that puzzles are good for your general intelligence, but only up to a point. If you keep doing the exact same puzzle over and over, you are mostly training yourself to be very good at that specific puzzle. And so we design our puzzles with variation and diversity at their core. Keeping our players on their toes, if you will, so that you actually strengthen a wide set of skills instead of getting good at that specific puzzle.

Additionally, I remember being into Sudokus as a very young kid. But after a year or so, I completely grew out of them. Why? Because there was nothing new, no variation, nowhere to go from there! There weren’t as many Sudoku variations yet, and in any case I hadn’t discovered those yet at that age.

Kids are notorious for being easily distracted. For being obsessed with something for 30 minutes, and then completely losing interest because they want something new. In fact, in the current day and age, we might even be able to extend this to adults too, with their shortening attention span.

As such, if we were to create a Sudoku puzzle, we would do it differently.

• The first page would explain the bare minimum to start playing. Maybe we’d only explain rows and columns, and only numbers 1–4.
• Then we’d have 5 puzzles with this ruleset. Each highlights a different trick you can use or different situation to solve.
• Then we’d go to the next level and add a twist. Maybe we do numbers 1–9 now, or we add the boxes/sub-groups.
• After 5–10 puzzles, you’ll have seen all this has to offer and we go to the next level.
• This continues for as long as we can invent good, reasonable, simple twists to add for each level. (Without adding so many extra rules that the puzzle turns into something else entirely. At that point, we’d cut it off and start a new puzzle with that base ruleset.)

We’re unlikely to do this, though. Because why reinvent the wheel? There are so many places to play Sudokus (and variants), online, for free. No, when we design puzzles we try to create something that doesn’t exist yet and serves a specific (educational/curriculum) purpose.

Breaking Puzzles Into Levels

And so the question became: how do we do this? How do we break a puzzle down into its absolute smallest parts, so we can teach those one at a time? What unique things can we bring to classic puzzles?

Puzzles For Kids

The first unique aspect we recognized is our line of “pre-reader” and “pre-counter” puzzles. Puzzles for kids usually don’t extend beyond a basic word search or maze. This is somewhat understandable, as most (logic) puzzles use numbers and words and advanced concepts like addition.

This is also a shame, however, because most puzzles can quite easily be adapted to be “textless” or “numberless”. We’ve just explained how incredibly good puzzles are at training skills while its player is having fun. We wanted to prove this to the world.

Below are some of the basic steps to making a puzzle accessible to (young) kids.

• Many things can be made physical. Instead of playing by writing or drawing lines, it’s often possible to provide things to cut out and place. Or to literally play by cutting/folding.
• Letters can be restricted to the easiest or most common ones. Even better, words can be changed into recognizable images.
• Numbers can be lowered (to just, say, 1–4). Even better, they can be changed to simply showing that number of icons, or showing the exact thing/shape/whatever that needs to be placed.
• A few rules can be left out. Yes, this makes the puzzle “less interesting or challenging”. But that’s actually a feature here. We get a simpler puzzle that kids won’t struggle to solve.
• Teach the rules through images. Reinforce them through simple shapes/colors/icons, really adding a textless “guiding hand” wherever possible.
• Set the entire puzzle to a theme that kids know and understand, such that its rules/mechanics feel “intuitive” or “natural”. (Because they’re a simulation of how that thing works in the real world, as the kids know.)

Yes, it can be hard, it can be a struggle, it can be “annoying” that you have to ditch a really interesting rule because it would require reading ability. But the fact that we have nearly fifty puzzles for pre-readers and pre-counters shows it can absolutely be done. In fact, by simplifying like this, we were able to get adults to play, who otherwise never puzzle because they feel it’s too hard or taxing on their brain.

Smallest Rule Unit

The second unique aspect is our “step-by-step level structure”. It means we don’t have to overwhelm you with a wall of text when first learning a puzzle. Even complex puzzles are simple and obvious when broken into 2–5 separate teaching moments.

The next time you invent a puzzle, or learn a new puzzle, think about this. Think about every distinct rule or distinct set of information that you had to learn for this puzzle. Call them the “smallest rule units” (SRUs) if you want; rules you can’t break down any further into smaller units.

Now … simply … teach those SRUs one at a time! One per level. Not all at once.

The entire ruleset exists for a reason, of course. The puzzle is at its most interesting, and challenging, and engaging when all rules are enabled. That’s why it might feel like you’re ruining things by breaking them into bite-sized pieces.

We came to realize, however, that you’re not ruining anything at all.

• The first few tries of a new puzzle will be interesting/engaging just by virtue of being new. You can get away with overly simplistic rules, because those are new too and will carry initial excitement and challenge.
• With small tweaks, you can often make the simplified ruleset interesting too. Because you know you’ll only be doing ~5 puzzles with this, you can really do a handful of unique things and keep interest high that way.
• The player will also know that something new is coming in a few puzzles. And again, even if that new rule is only a minor addition (and not the full “best” ruleset yet), it’s still new. It will keep the interest for a bit longer again.
• And, our most important realization, you can add rules at the start that will be removed/irrelevant later.

That last point was probably the trigger that made me write this article, so I’ll give it some more explanation.

In puzzles (and games), I’ve come to learn there are “hard rules” and “easy rules”.

Hard rules are things that always hold, which means players need to remember at all times. This is hard to understand! It’s easy to make mistakes here! In games, other players can keep you accountable and check mistakes. But if you’re puzzling on your own? Hard rules will be … hard to execute well.

Soft rules are things that hold conditionally or that are optional. If you see this icon, then the rule holds. If you connect a bridge to this kind of location, then the bridge is worth double the points. These are much (much) easier. Because you don’t need to keep them in your head. There is some icon, or board state, or visual to remind you at the specific moments that it’s relevant.

It also means … that we can “deprecate” the soft rules without having to do anything. We simply don’t make that icon appear ever again!

Hopefully you see where this is going.

• We can make puzzles simpler and more interesting at the early levels by adding a rule of our own. Some twist that makes sense. (And which doesn’t exist in the puzzles or games by which our product was inspired.)
• If we make this a soft rule, we can simply … forget about it in later levels, when the full ruleset is alive and kicking.
• Without any extra explanation, or confusion, or statements like “oh, and also, that one rule you used in the previous 10 puzzles, FORGET ABOUT IT NOW!” we fixed the entire issue. From start to finish, all puzzles are engaging and challenging in their own way.

An Example

I just finished the first Brainblankets puzzle. It’s perhaps not the best puzzle we ever did, or the best example of this, but it’s in my short-term memory now and it is a good example of these ideas.

The puzzle is based on “Dominosa”.

• You get a grid of numbers (0–3 in the smallest size).
• You must “pair” the numbers (create 2x1 dominoes) …
• … such that all numbers are part of a pair, and no two pairs have the same numbers inside.

It’s a very simple puzzle that nevertheless felt really challenging the first time I played one. Moreover, it fit the topic of “grouping/pairing” perfectly, so I decided to do a variation on Dominosa for that topic.

What did I change? First, I modified it for kids.

• Firstly, the numbers became symbols (shapes + colors). Because there’s no puzzle reason for them to be numbers, and now young kids can play too.
• I added “blankets” that you can place over the puzzle to pair numbers.
• I started with an even smaller grid (than the smallest one you’ll find anywhere else) and fewer different icons.

Then, I focused on the educational angle.

• At later levels, when the puzzle rules are well-known, we switch from icons back to numbers. (And added a number line as guide/reference too.)
• You can only “pair” numbers that are the same or consecutive. This subtly teaches numbers and counting. Without any further teaching or exercises, a kid will now repeatedly check if two numbers come after one another or not. And when you know 2 follows 1, and 3 follows 2, and so forth … you can count.
• The final level adds a “wildcard” that can be whatever you want/need to finish the puzzle. This allowed keeping puzzles small (instead of blowing up their size just to make them harder near the end). It also does something very important (but oft-forgotten) when teaching something: exploring the inverse too.
  • Before, a kid would ask themselves “are number X and number Y right after one another?”
  • Now, with a wildcard, a kid must ask themselves “I have number X, what number comes right after it?”

Finally, I focused on the different levels and making it doable/understandable for 3 year olds.

• Level 2 adds the “forbidden blanket”. This may sound like a curse or an extra challenge, but it actually makes thing easier. Now you know this blanket can’t appear, which eliminates a lot of options. This allows me to make larger/harder puzzles later without them being too hard.
  • Note that this is a “soft rule”: the specific blanket(s) that are forbidden are shown next to the puzzle. There are many puzzles that just show nothing, which means none of this needs to be kept in mind. The final puzzles don’t use this, for example.
• Similarly, level 3 adds the “3x1” blanket. It simply groups 3 things instead of 2. This is, again, mostly something that helps. You can group more things at once, and it’s highly unlikely your two 3x1 blankets have the same icons inside. But it’s a kind of help that provides a different challenge from before.
  • Again, a “soft rule”. You only use the two 3x1 blankets if their icon shows next to the puzzle. Otherwise, forget about this rule entirely.
  • NOTE: In Dominosa, a puzzle always uses all possible blankets. That’s simply guaranteed by the size of the grid and the number of different numbers/icons. As such, to actually make puzzles solvable with forbidden blankets or wildcards, I had to take some “bites” out of them or add one more icon type than strictly necessary.
• Level 4 then switches to numbers. Only 1–5 at first, then 6–9 too.
• Level 5 adds the wildcard. As explained above, this is a simple rule that solidifies the teaching (by checking if you can do it in reverse), as well as twisting the challenge slightly again.

Each level only requires 2 short paragraphs + image next to it. But at the end, you’re not just playing Dominosa, you’re even playing a slightly harder and more interesting version of it.

Of course, there were many other options I could’ve chosen. Most projects end up with a file called discarded.md, which contains a long list of things that didn’t make the cut. Some rules were slightly too complicated. Some rules would add extra challenge, sure, but not really a different kind of challenge.

For example, another possible rule was an icon (in the puzzle) that shows the blanket orientation needed here. In other words, it’s a helpful hint/clue that says “this one must have a HORIZONTAL blanket” or “a VERTICAL blanket here”. I opted for the forbidden blanket instead, however, because …

• An icon inside the puzzle was slightly messier. And visual clarity is very necessary for young kids.
• It’s harder to gauge the helpfulness of such a hint. It would take me more time (or trial and error) to figure out what to hint to prevent making it too easy or basically worthless. (Example: hinting that a horizontal blanket must be here … when logically, if you just looked at the puzzle for 2 seconds, you’d already know this … is a bit useless.)
• The puzzle is in the “grouping/comparing” topic, whereas orientation and spatial stuff is handled in other topics. The forbidden blankets are basically “forbidden groups of icons”, so they focus much better on the topic at hand.

Conclusion

These are the kinds of considerations and experiments you don’t really see in the final product. But they matter a lot. When you buy a puzzle from us, we did our best to distill the idea to ~5 levels that are absolutely as simple, focused, minimalist, and interesting as they can be. And we hope kids have fun doing them.

In practice, 25–30 puzzles total seems to be the sweet spot. This is a good amount of content that will keep you busy for a bit, and it justifies a reasonable price tag. It prevents overwhelming or wasting ink/paper with near identical puzzles. (It is also, to be honest, the limit of what my current computer can handle!)

Puzzles are not homework. Dumping an entire ruleset on a kid at once, and then giving them 100 nearly identical puzzles with that, just isn’t the way to go. Instead,

• We modify puzzle ideas to be as textless, simple and accessible as possible. We try to provide a reward at the end (such as an image appearing or you filling a whole grid), or small rewards along the way (such as being sure that this part of the puzzle is now completely correct/figured out).
• We break them down further into the smallest possible distinct rules (SRUs).
• And then teach one new rule for each level, followed by 5–10 puzzles with that specific ruleset. In most cases, this means each level “resets” the puzzle size again to the smallest it can be, and we grow in size as the level continues.
• Until, by the end, you’ve usually reached both the “full ruleset” (most interesting and engaging puzzles) and we’ve sneakily taught you something like counting, or addition, or rotating shapes in a grid.
• If needed, we add “soft rules” along the way to keep the early puzzles small but interesting. And then we can just let the soft rules dissolve as we get near the end, with none being the wiser.
• Once we’ve reached our cap of ~5 levels or ~30 puzzles, we’ll break it off right there. Better to put any more advanced ideas into a separate more advanced puzzle. Better to stay minimalist and go out on a high note, and especially before the kid gets bored.

Those were my current thoughts and lessons on why and how we design our custom puzzles. Hopefully it was interesting to read!

Keep learning, keep puzzling,

Tiamo