“No, Not that Way!”

Words: 2053 Approximate Reading Time: 15-20 minutes

Back when I was playing Tears of the Kingdom and writing up my thoughts on some of the game’s mechanics, I realized that I had never done a dissection of physics puzzles in video games. These are such an integral part of both Breath of the Wild and Tears (especially Tears, though) that it seemed at the time like an obvious thing to talk about. Then I got distracted and never really did anything with the idea. But physics puzzles will still keep getting made.

And yet, I didn’t really want to talk about physics puzzles within the context of those two games specifically. Firstly, because the physics puzzles tend to work fairly well in those games. Emphasis on “tend,” though. Secondly, because focusing on those two games wouldn’t tell us much. Even when they do run into problems, we wouldn’t be able to pull much information from those failures – they would only tell us why those particular puzzles weren’t fun to engage with.

Physics puzzles are fairly notorious for being bad in one way or another. Often because the physics are too hectic – whether working correctly or bugging out – to create any feeling of real consistency. At their best they can usually feel like exercises in frustration, performing the same action over and over again in the hope that the physics will behave the way you want them to so that you can succeed. At their worst, they might not even make sense.

And we won’t even get into the fact that you might be able to just bypass the puzzle solving process entirely – something that can be either fun or frustrating depending on a host of factors.

So I want to dig into the problem with physics puzzles and why they so often fail to be engaging, at least beyond the initial novelty. These problems exist at the root of the game’s mechanics, and while I think this probably is something that can be overcome, it is probably an uphill battle.

Puzzle Solving

To understand the issue with physics puzzles, let’s break down the puzzle solving process into two components.

Let’s call the first part “insight.” This is where the player tries to figure out what they’re supposed to do in a “big picture” sense. Insight takes on different forms depending on the type of game we’re talking about. In a game with a lot of inductive logic such as The Witness or Myst, this might be figuring out how the puzzle itself works. For a game with more clearly defined mechanics, it might be discerning the specific “trick” that gets you to the end goal. Maybe it’s understanding a new mechanic or puzzle element, maybe it requires seeing how to utilize several mechanics in tandem, maybe you need to learn how to use an old mechanic in a new fashion.

Let’s call the second part “manipulation.” Here’s where you take your knowledge about the puzzle and try to actually put the particular components into the right arrangement to achieve the win state. Perhaps your insight phase yielded that you need to put a box on a button to hold a door open, so you work on getting the box onto the button. As we start throwing in more and more objects, the more complex this process gets. Sometimes it could be a problem of order – you need to activate a lever before you put the box on the button, or else the box will get knocked off. Sometimes it could be a problem of timing – the box will disappear after ten seconds, so you need to get through the door before the box goes.

These two parts of the process do not have to take place successively. We can cut up larger puzzles into smaller iterations of this procedure. You could have multiple buttons and levers that need to be pushed and pulled, each one being their own mini-problem requiring both insight and manipulation.

By the same token, sometimes messing around with puzzle elements yields insight. If you don’t know how something works, you might experiment to see what it does when you push, prod, throw, or hit it. So manipulating these elements can provide you with the insight that you need to understand what you’re trying to do.

And of course, just because you think you have full insight into the puzzle does not mean you actually do. Maybe you see how to reach the end goal, and so you work on the manipulation and arrange the puzzle pieces…only to realize that there’s a new problem to grapple with. Sometimes your manipulation can reveal the puzzle’s real trick, which brings you back to the insight phase.

All of this to say that while we can describe puzzle solving through these two processes, they are not clearly delineated.

But one key takeaway is the relative degrees of fun in these two parts. The insight process is the one that is most interesting, the one that genuinely yields the “Aha!” moments that make us feel good and smart. The manipulation process can still be fun, but experiences diminishing returns. The more time you spend just moving pieces around without knowing what’s happening, the more you’ll get frustrated.

As an example, let’s imagine a puzzle that is entirely manipulation. Picture a wall with a dozen levers on it. You need to pull the levers in a particular order to open a door. There is no information to tell you when you’ve pulled a specific lever in the correct order, nor to tell you what the proper order is. All you can do is run through various combinations until you get the right one. For the sake of simplicity we’ll presume that you don’t have to pull any lever more than once. That means you have 12x11x10x9x8x7x6x5x4x3x2x1 possible combinations to work through, or 479,001,600. Just sitting there running through nearly half a billion possibilities.

If that seems annoying and not worth doing, that’s the issue with putting the focus of a puzzle on proper manipulation. You’re not actually “figuring out” anything. You’re just running through possible moves until it works. This is why good puzzle design needs to focus on insight – what’s your goal – and not manipulation – do things the right way.

Toying with Physics

With that setup out of the way, why do physics puzzle games so often feel annoying or dull?

The superficial answer is that their puzzles rely primarily on manipulation. You generally have some goal that is obvious, and so the “solution” is figuring out precisely how to move or create objects to achieve that goal. But that just gets us back to the wall of levers – you have tons of possibilities, and you’re just running through those possibilities until you find one that works.

But the reason why these games tend to focus on manipulation forces us to dig a bit deeper. To which we find a couple of root causes.

The first of which is the emphasis on “realistic” physics. That is, you have an almost infinite number of ways that puzzle elements can interact based off of incredibly tiny differences. You drop a ball through a set of pegs, and it falls in one way. You move it ever so slightly to the left, and it drops a completely different way. The idea is to capture a sense of randomness based on how physics can feel, removing the ability to perfectly predict what will happen. But in turn, that randomness or sense of randomness makes it difficult for the player to figure out what they’re doing. It turns the process into trial-and-error, to the point that you can do the same thing multiple times and yield different results just because of tiny differences. Even worse, you might be doing the right thing and have to do it multiple times just to succeed.

The second, and probably bigger, problem is when those physics don’t behave properly. When objects move in ways that don’t feel right. Some of this can be chalked up to a combination of poor programming and the fact that game objects don’t literally behave the same as real objects. For example, if you can create objects at will, you can do things to break the game’s physics that wouldn’t be possible in the real world. But nevertheless, when the physics don’t work correctly – when even doing the literal same thing yields wildly divergent results – then you’re not just putting the player through a bunch of tedious trial-and-error. You’re effectively gaslighting them, making them think they’ve done something wrong even if they’re correct.

In a sense, the root problem with physics puzzlers lies with the focus on physics itself.

Theoretically a physics puzzle game can be good. But to be so, it needs to fight these tendencies for the physics to feel off. In particular, it needs to shift the focus away from making objects work on minute levels, and instead give players less freedom. Maybe you can create blocks, but you can only create blocks of certain sizes. Maybe those blocks can only be spawned on a grid, so that there’s more consistency in how they’ll behave. These types of solutions make the physics more predictable. We can see an example of this in Tears of the Kingdom’s “Ultrahand” ability, which lets you levitate and rotate objects. Objects can only be rotated in a set number of orientations, which means that if a puzzle, say, required you to figure out the correct orientation for a ramp, you only need to try two or three different possibilities, rather than dozens.

In addition, to be good these games need to present physics puzzles that involve genuine complexity. Rather than “balance these objects” or “topple these blocks” or “move this ball from Point A to Point B,” give the players a goal that they also need to figure out. Provide them with an opportunity to experiment and gain insight into what the puzzle is asking for and what they’re doing. The more focus there is on just manipulating the game’s objects until you arrive at a working solution, the more the game becomes an exercise in frustration. Again, this is why something like TotK’s puzzles tend to be better – you have a limited set of tools and abilities to work with, and part of what you’re doing is trying to determine how to craft something from all of that that yields a solution.

Concluding Remarks

I try as best I can to avoid claiming that a particular game mechanic or genre is genuinely unworkable. Sometimes it may be the case for a specific instance of that mechanic or genre. Maybe horror and comedy don’t mesh well, and so you need to rethink your project and be willing to sacrifice something. Maybe fast-paced action shouldn’t be combined with slower-paced tactical combat, but you can find a way to cut back on one or the other and create a coherent experience. There’s often ways to make things work if you’re willing to be flexible.

But physics puzzlers get close to feeling like they’re just unworkable. It’s not that they can’t ever be fun. They do provide a degree of freedom and a simplicity in their approach that can allow for an enjoyable experience for a bit. But building a whole game with tons of “puzzles” for the player to “solve” just introduces tedium. At a certain point the novelty of the play turns into frustration as the game doesn’t really do what you want it to do. Something like Breath of the Wild and Tears of the Kingdom can only really succeed with these puzzles because they are a single part of a larger game. If you were doing nothing but solving these puzzles, they’d probably start to feel grating. And that’s even allowing for the fact that these puzzles are probably among the best physics puzzles in gaming.

Really, the root problem is in the physics itself. The freedom offered in some way runs counter to the intellectual processes involved in puzzling. Making a good physics puzzler requires shifting the emphasis away from the physics component. It requires treating the game first and foremost as a puzzle game, and understanding that the player’s engagement should be about solving a problem, rather than trying to brute force a solution.