One of my areas of research is in the entertainment factor of games; in particular, in which way artificial-intelligence techniques can be used to evoke and maintain an entertaining game experience. Some of our work is based on the theory of incongruity. In game research, incongruity is the difference between the complexity of a game, and the complexity of the mental model that a player has of the game. For example, if the game is chess with Deep Blue as opponent, and I am the player, there is a positive incongruity as Deep Blue would wipe me off the board. However, if the game is chess with my nine-year-old daughter as opponent, and I am the player, there is a negative incongruity, as my daughter has just learned the rules. The theory of incongruity states that if incongruity is highly negative, the player will get bored, while if it is highly positive, the player will get frustrated. However, if incongruity is balanced, the player will be interested in the game.
The problem with games in this respect is that they usually are highly multimodal environments, and it would be too simplistic to just assign a one-dimensional complexity measure to them. We must distinguish many different kinds of complexities. For example, for a specific game one could distinguish tactical complexity, strategic complexity, story complexity, visual complexity, timing complexity, and interface complexity. Presumably, incongruity should be balanced on all those complexity features (what makes a complexity feature ‘balanced’ naturally differs from player to player). One might assume that a player will get frustrated even if only one of them is highly positive.
I can talk a great deal about this, and I will probably do that in future posts, but for now I just want to focus on interface complexity. In this case, with interface complexity I am only referring to the user inputs: the buttons that must be pressed, the movements that must be made, to control the avatar on the screen.
Until recently, I assumed that game developers would aim at making incongruity for interface complexity very low. Not so low that the game would get boring (e.g., if whole game would be controlled with only quick-time-event-like button mashing, it would probably get boring after a while), but rather erring on the side of negative incongruity than positive incongruity. To put it another way: a game that is hard to control loses entertainment points.
Nowadays, most games have introductory levels, which exist mainly to make the player familiar with the interface. After the introduction, the player is let loose in the game world, and the interface should be second nature at that point. In some games the interface complexity is set deliberately in a positive range: many platform games have the player perform all kinds of tough interface stunts. In terms of the game world, such an interface complexity actually makes sense: if Lara Croft has to cross a hallway with dangerously swinging blades, she has to execute her movements with precision and tight timing, and thus the player has to control her under similar restrictions. A problem is that while Lara must run, jump, dive, and roll, the player must press A, B, X, and Y. Making the mental translation from Lara’s movements to thumb controls is not for everyone. That is one of the reasons that the Wii is so well-received by sections of the population for whom gaming until now was off limits.
Currently, I am playing the game Eternal Sonata. This an originally Japanese RPG, which is set in a world built around music (up to the point that it gets a little annoying). It has a rather intricate story, told mainly in cut scenes, which are interspersed with repetitive fights and an occasional interesting boss fight. The game deals with the interface complexity in a weird way. The interface basics are explained during introductory levels. But the deeper you get into the game, the more complex the interface becomes. For instance, during the first few gaming hours, the fights are basically turn-based, allowing the player practically infinite thinking time. But a bit later the game throws out the thinking time, making the fights more and more real-time events. Nearing the end of the game, the developers really went over the top: for each fight, the buttons get randomly re-assigned to new commands. Yes, you are reading this right: the player has to learn new controls for each fight. Frankly, this is rather ridiculous and highly irritating. And it is completely wrong too, as within the confines of the game world there is no good reason why interface buttons would get re-assigned.
For me personally, interface complexity is not something I derive pleasure from. For me, a balanced interface complexity would mean: standard button assignments, linked to the neural pathways that I have trained to translate my thoughts to actions on the screen, and preferably some time to consider my movements. I am definitely not alone in this, but I understand that others find their pleasure in higher interface complexities. Still, I cannot fathom why anyone would enjoy what Eternal Sonata is doing during the final stages of the game. I assume it has been playtested and found acceptable. However, for me it feels like the developers feared that their game would become boring as the fights are very repetitive, and rather than making the fights more diverse, decided to throw the player back to the learning stages where fights are hard for the simple reason that the player cannot find the right buttons.
Well, randomly re-assigning buttons is definitely cheaper than creating extra interesting game content.