Granularity, Symbol Or Representation?
This is one of my favorite subjects. When designing the details of a game, where do you stop? Should we see the characters eyes? Should we see the leaves of the trees? Etc…
This is about how the level of granularity that the game should go into. There are two parts to this, on one hand the fact that the granularity should be the same at all levels in the game, the second is that the granularity of the simulation or AI, graphics and user controls are all linked.
Same granularity through out the game
It is believed that Neanderthals wiped out the Cro-Magnon from the face of the earth because Neanderthals had a new skill, the ability to use symbols. That is good news for the designers of games,
it means that we do not need to describe things exactly like they are but we can conceptualize and use symbols effectively with our users. In the book “Understanding Comics” one point is made very clearly with the figure below
Figure From “Understanding Comics”
We have the happy face on the Right, and a realistic picture on the left. One is a symbol, the other a representation. It is easier to identify yourself to the happy face than to “Clark Gable”, so if you use a symbol in the game rather than a representation, it is easier for the user to bring in his own set of expectation into the game.
A great example of this in another media is the recent “Blair Witch Project” movie. I suggest you look at that movie again if you have not seen it, and carefully take notes (mentally) on what information you are really getting. I have done that exercise after Will mentioned it, and there is very little information. Most of the movie experience is about the fears the viewer projects himself into the movie story. The granularity of the movie is such that most of the movie scenes, are created little vacuums that suck the story from the viewers mind.
It is even better to watch that movie with a few other people and get them to describe what happened. They will all describe different experience, because they did not get as much from the movie as they got from their own mind.
One question I ask the designers regularly is what rule do they use in the game to decide to what level of granularity they need to go? It is very complex to figure out.
Clearly it is better to have the same rule all over the game, because then the user will have the right set of expectations when getting into a new corner of the game. An example of rule is one we use in the Sims: the granularity is such as to maximize the story creation in the head of the player. Enough details to suggest that something is going on, not enough details to go against the story that the user believe is driving the behavior of the characters, breaking it.
Here are some examples of issues that arise from poor granularity: If you show finger, they should hold objects properly. When doing Madden on the 3DO, for the first time the graphics were letting you see the arms of the players, and everyone was bothered because the players were not raising their arms when catching the ball.
Most of the difference in granularity gets caught by the time the game is finished, but the aware designer will understand which rules he wants to use to define it and will save development time.
Granularity relation between “User Interface”, “Simulation AI”, “Graphics”, and “User mental model or Story”.
How do the different level of details relate to one another in the categories listed above?
What we want is the user to give simple direction, and we want all his commands to be acknowledge by the simulation, this means that we have “User Interface” < ”Simulation”.
We can in general display at low cost more details in the graphics that we have in the simulation, and that is a good payoff. As we mentioned above we do not want to have a lot of difference between the two, but in general we still have “Simulation” < ”Graphics”.
Finally it is clear that users always build more in their mind than there is really in the game. We all have countless stories about that, here is one where in Road Rash, a user saw a pedestrian next to a car stopped on the side of the road and concluded that someone was getting into a cab, a concept that never was in the game. From this we can say that “Graphics” < ”User mental model or story”
Regrouping all of those we have:
“User Interface” < ”Simulation” < ”Graphics” < ”User mental model”
Again this means that the user has a level of control and that the simulation will drive at a higher level of detail,. The graphics will reflect a slightly higher level of details and finally we can expect that the user in this head will build a story that has yet even more details. This is like magic for the user, he gives simple actions, and life like complexity comes out of it on the other side.
Let me give an example from SimCity: The user lay down roads, the simulation generate traffic trips on those roads, the graphic shows a police car on the road and the user think that the car in front was speeding and is about to be stopped by the cops.