“If you feel lost or discouraged, it's possible that this piece of advice will help you through your troubles: "As long as you have questions, continue."”
-- Starseed Pilgrim (PC, 2013)
“Explore. Take risks. You won't always know what to do next. Keep experimenting, and you'll master it.”
-- Cultist Simulator (PC, 2018)
Every game contains within itself a learning meta-game, in which the goal is to understand how the game works. Throughout this meta-game, the player learns how to deal with the game's most basic elements, i.e. its controls, rules, and goals. This task can take from seconds (e.g. when you're playing a game that you've already played before) to hours (e.g. when the game is particularly confusing).
Most of the time, this meta-game is seen as a necessary evil, as an obstacle lacking intrinsic value that only takes away from the game's "main fun". Because of this vision, many games are designed to make this meta-game as easy as possible, resulting in well-known design decisions such as: the UI displays the direction of the next goal, a menu displays the current state of every quest, a map pinpoints the exact position of important places, traditional control schemes are adhered to as much as possible, etc.
However, this vision of the learning meta-game as an "obstacle" is not necessarily true. For example, take the games Dark Souls (PS3, 2011), FEZ (PC, 2012), The Witness (PC, 2016), and Outer Wilds (PC, 2019); although they seem to have little in common, these games share a same design principle: understanding how the game works is seen not as a chore, but as a challenge -- a puzzle to be solved. In other words, they incorporate the learning meta-game into the game's intended experience, and therefore confer intrinsic value to this task which is normally abhorred by Good Game Design(TM). As a consequence of this shared principle, these games share too a divisive reception: many players deem them "too confusing" and leave unfavorable reviews, while other, more curiosity-oriented players, welcome and enjoy the challenge. We the authors fall into this second category, and through this essay we aim to both (1) analyze what exactly these games have in common, and (2) propose practical design principles that can be followed in order to create a similar game.
As far as we know, there is no specific term for this class of games, therefore we coin the term "mystified games". We think this term is apt for two main reasons.
The first reason is that, if the learning meta-game must be interesting, then the game's basic elements need to be initially unknown to the player. This means that the game design intentionally obfuscates its inner workings, making it hard for the player to understand what can be done (the mechanics) and what needs to be done (the goals). Such intentional obfuscation fits well with the dictionary definition of "mystify" as "perplex purposely".
The second reason for this name is to pay homage to Myst (Macintosh, 1993), one of the most emblematic examples of this type of experience, and a true trendsetter in game history overall.
In the next section, we detail the main mechanisms used by the player to understand how a game works, and in the following sections, we will see how the game designer can avoid these mechanisms to mystify both the mechanics (section 3) and the goals (section 4). After that, we present the difference between mystified games and two other proposed genres in the literature: "puzzlebox detective games"  and "information games" , concluding that there is a significant difference between the definitions. Finally, we conclude the essay in section 5 with a series of practical design principles that can be applied to mystify a game design.
[... come back tomorrow ...]