Non-linearity, multiple solutions and interactive dialog

A widely discussed problem with adventure games is that they're pretty restrictive in the number of paths that can be taken through a story, or towards solving a certain puzzle. A lot of adventure games are just about taking object A to location B, which gives access to location C and object D, ad infinitum. Such linearity has the player conform to a set path — making puzzles all the more unnatural and frustrating.

Adventure games need to loosen up and drop the tunnel-vision narrative. It may be hard for a game designer to keep a good overview of the gameplay flow of a highly non-linear game, but that problem isn't unsurmountable. Non-linearity isn't like some holy grail that only exists in the fantasies of game designers. It's actually been properly done before and some new games need to do it again.

The Last Express and Discworld Noir are perfect examples. For instance, Discworld Noir gives a large amount of freedom to investigate whatever you want, whenever you want and in which order. Discworld Noir is one of those games that flows very naturally. When you compare your playing experience with that of others you discover how different they can be. I've even heard of players who never got to use the third CD because of a specific path they took through the game. In games such as Discworld Noir that often offer different solutions to puzzles, players are less likely to get stuck, and more likely to enjoy the story.

On the other hand, I don't think Indiana Jones and the Fate of Atlantis or Blade Runner are particularly good examples, despite how often they're referenced in discussions about non-linearity. Fate of Atlantis is merely multi-linear. There are three paths through the game that run parallel to each other. Once you follow one of these paths, you are locked into it until the game's conclusion, where the three paths converge. The design of Fate of Atlantis was remarkable at the time, and it had a very strong replay value, but it's not the kind of non-linearity that I find interesting. In the case of Blade Runner, one of its selling points was its wide array of different endings, but this too isn't very exciting. Who cares about seven different endings if the middle part of the game seems the same every time?

If we take a closer look at non-linearity, we can say that it can be achieved in roughly four ways.


Most adventure games are either completely linear, or they provide a limited amount of freedom through what Roberta Williams used to call the “string of pearls” model. Players can move around freely inside the pearl, until the story converges back to its main path, leading to the next pearl, and so on. However, the linear narrative and the string of pearls model are not the only viable options.

In this article on the IGDA site, Discworld Noir developer Chris Bateman discusses some alternative narrative structures that greatly enhance the possibilities for “order” and “selection”. To dive into game design theory would be beyond the scope of this article, so I'll quickly summarize the gist of it here. However, if you're really interested in this topic, I encourage following the link and reading the full article.

Discworld Noir used what Bateman calls a “threaded model”. Unlike most games, the course of the plot does not follow a single path. Instead the story is made up out of different threads that develop largely independently. Bateman writes that “this parallels the methods that quality novels and films employ to develop the plot, with multiple narrative threads being woven together to produce a satisfying story.”

After the initial introduction is out of the way, Discworld Noir has the player investigating a number of different cases. These are all different threads that can be explored in any random order. Some threads, however, have to be concluded before the player can proceed to the next act. For example, the player has to wrap up his initial investigation in one of two ways, after which he can pursue one of two (seemingly unrelated) cases. Some threads can be sidestepped; some can be ignored for a while and gotten back to later, while others aren't possible to conclude until the player has advanced to a certain phase in the story. This model gives a greater freedom of movement within the story and is vastly different from linear narrative.

Bateman also suggests a new theoretical narrative model which he believes will be the future of interactive plots. What he proposes is using an object-oriented approach, in which scenes are grouped together along any desired model (linear, branching, parallel or threaded). These collections of scenes, or episodes, are then in turn linked to other episodes to form a web. In order to realize such a complex object-oriented structure, Bateman suggests the re-use of set pieces and other assets in a way similar to TV shows. The player may see a location only once or twice, but there could actually be a multitude of scenarios for what may happen on that location.

This model has not yet been used in practice, but the point in telling this is that there are enough ways to make a game non-linear. (Bateman and his colleagues have put their theories into an engine called FreeSpeak, but it has not been used in a game thus far.)