When you are designing a computer game, there is a number of constraints you need to take into account. Some of those constraints come from the technical possibilities at the time, some come from limited time and budget, some come from the desired theme and inspirations. But there is also a whole category of constraints that come from the fact that you are designing a game to be played by humans. Our brains have their limits. We orient in 2.5 dimensions, we can only hold 5±2 items at a time in our working memory, we get tired by calculations and step-by-step analysis, we try to see immediate and binary cause-and-effect relations everywhere and we use very bad heuristics for probabilities. To be enjoyable, the games have to take those limitations into account, at the same time being challenging just enough to keep us in the flow. There are several well-established techniques that game designers follow since ages to make their games easy to pick up and enjoy. But there are some consequences.
Because those techniques appear so often, and not just in video games, we have learned to expect them and take advantage of them. If they are missing, we tend to dislike the game. There are many such techniques, but let us look at just one of them as an example: instant feedback.
When we do something in a game, we need to see a reaction within 0.1s, otherwise it interferes with our amazing ability to "get into the skin" of the game's character. If we get no feedback for more than 1s, then we assume that clearly something is broken. After 10s we start doing random things to see if we get any response at all. Then we lose interest and go to do something else. That's how we are wired. There can be a lot of variation in this behavior, depending on how import the task at hand is, how prepared we are for the lag and whether we have recently eaten some sweets or not (I'm not kidding!). But the lesson is clear: a game needs to provide immediate feedback to our actions. It gets worse. If the effect is too far in time from its cause, we are never going to make the connection. At least not intuitively – we might be able to reason about it and even do some in-depth research if we have the incentive – but normally we wouldn't bother and just assume the effect was a random event, or caused by some higher power other than RNG. (Humans didn't figure out what causes pregnancy for hundreds of years!)
This forces the game designers to use certain established patters. Consider two possible designs of a skill system. In one approach, you select a skill that you would like to train, and then you go around the game and gain experience points, until you gather enough of them to get a level up in the selected skill. In a second approach, you earn experience points that you can then spend to buy a skill level. The second system will feel much better to the majority of players, because the effect (new skill being available) is not removed in time from the cause (the selection of the skill to train). And indeed, if you look at the games out there, the second system is much more popular. It just feels so much better for us to make a choice and immediately be rewarded for it, than to work hard on something without seeing results.
There are many other such patterns. Indeed, there are so many, that when we see the initial screen of the Pacman game, we immediately know what to do: we have to eat all the dots! How do we know it? The game has been designed with an intent, and because it was designed by humans, and we are humans too, we are able to pick up hundreds of small hints and guess what the designer had in mind. And once you guess it, all you have to do is to perform it – make it happen. Does that feel familiar? When you were at school, did you notice that they don't really teach you to, say, analyze poetry? They don't want you to do any analysis. You are just supposed to guess what the teacher thinks, and then write that. Get the "correct" answer. Obey. That is what the school trains you to do, and that is also what the video games train you to do. Obey.
As far as I know, this effect is inherent in some extent in practically all man-made objects. We can't help ourselves but to put messages in every little thing we make. Messages that tell about our expectations of the objects, their affordances. You can't design a video game without including that "obey" message in it. Or can you? It turns out that you actually can! But you have to let go of the design, at least in part. Make it procedurally generated. I think that this is at least in part why Minecraft has gotten so popular – it's a game that doesn't say "obey", it's a game that lets you do whatever you want. I think that roguelike games also have the opportunity to be such games. Of course, this depends on the game, but the feeling that you are fighting against a random world, and not against a particularly crafty level designer is worth a lot to me.
But is a "sandbox" game the best that we can make? We can make a game that doesn't tell you to obey, but would it be possible to make a game that actively teaches you to not obey? I'm not sure, but here is one idea for such a game:
The designed, easy to read, obvious messages are all against you. The useful, helping things are all random. You are forced to come up with your own solutions, instead of relying on guessing the intentions. Sure, guessing the intentions may still be useful, but only to subvert them. Ideally, the game would have a similar setting to the puzzle game Portal – you are a subject of an experiment forced to follow a path. Except in Portal, even the subvertive parts of the game were designed – no matter where you got, you knew that there is a designed solution to the puzzle. Here all bets are off. There might be an insanely simple shortcut that lets you finish in 3 moves, or there might be none and you will be forced to follow the marked path after all. I think that this kind of approach could be good for promoting independent thinking and problem solving. Unfortunately, although I attempted to implement it during this year's Seven Day Roguelike Challenge, I saw pretty early that I wouldn't be able to pull that off and switched to a backup plan.
Whether you like this idea or not, and whether you want to use it or not, one important lesson that I want you to learn from this is to be careful, as a game designer, what your games are teaching. Instead of blindly copying the design patterns from other games, take a step back and think what kind of behaviors you are promoting. I'm sure that it will lead to more interesting, less grindy games.