8.2 Part II: Designing, Implementing and Testing the EmerGEnT System
8.2.2 Emergence as an Approach to Game Design
Various issues need to be considered when developing emergent game systems, including level of creative control for game developers, effort in designing, implementing and testing, effort in modifying and extending, issues related to uncertainty and quality assurance (knowing the system won’t break) and ease of
feedback and direction to players. The EmerGEnT system studies (Chapters 4 to 6) allowed these issues to be examined in more detail, adding insight and depth to the current literature.
8.2.2.1 Level of Creative Control
Scripted systems give the game developer complete creative control, as the game developer decides what will happen and when. However, emergent systems allow a more approximate control, in that the game developer guides the player, providing boundaries for gameplay rather than dictating specifically what will happen. In an emergent system, the developer can set goals, but cannot specify how the player will get there. The EmerGEnT system studies did not investigate the level of creative control as the system did not include creative content. However, the method of developing rules and properties to approximate the desired behaviour provided insight into the issues that would be present when trying to control creative content. Getting a simple environment to behave in a desired way was difficult and involved significant tuning. With a full scale world, there would be sufficient problems in getting the world to behave reasonably, even without the added complication of controlling the narrative flow. Considerable future work is required to determine how narrative can be used effectively in emergent systems.
8.2.2.2 Effort in Designing, Implementing and Testing
As discussed throughout Chapters 4 to 6, both scripted and emergent systems have considerations for effort in development. The EmerGEnT system involved substantial initial effort in planning the rules and properties that would govern the behaviour of the system. It was difficult to decide how certain behaviour should be modelled and what rules and properties best capture the behaviour. Subsequently, the system required substantial testing and tuning to get the rules to generate behaviour that was desirable or acceptable.
Scripted systems also require considerable effort in planning, as well as implementing and testing. As discussed in Chapters 4 to 6, scripted systems involve every game element to be set up manually. Not only the goals need to be set, but the ways to
achieve the goals must also be specified. In an emergent system, the game problems can be determined and the player can find their own solution. However, in a scripted system, the problem and solution must both be set and the player must find the developer’s preset solution.
Both types of systems have considerations for the effort required in development, but the emergent approach would definitely be favourable as the size of the game grows and it becomes impossible to predict, plan and code everything. Games are now very large, in terms of the size of the worlds and the amount of content, and they will continue to grow. Scripting everything is already infeasible and some game developers have found that the initial outlay of effort to get an emergent game world working is a superior solution to creating the entire world manually (e.g. Valve’s Half-Life 2). But so far, these games have been limited to game companies with extensive resources, experience and time, and only the objects (rather than the environments) have been emergent.
8.2.2.3 Effort in Modifying and Extending
The effort in modifying and extending was one of the major benefits of the EmerGEnT system, as well as emergent systems in general. The use of properties for different materials and objects made it simple to add new materials and objects to the system. The system dynamically retrieves the properties of the material or object and feeds them into the rules. Consequently, the system was data driven and adding new materials or objects and making changes to game scenarios did not involve changing any internal code or rules. Ease of modification and extension is a very important benefit as it allows developers to easily add more content, create expansion packs for their games (currently a big source of revenue for games), quickly release patches to fix bugs in the games and allow players to make modifications and create additional content. On the other hand, modifying and extending a scripted system is difficult and time-consuming. As mentioned previously, once the initial work was done in the EmerGEnT system, setting up new scenarios was a simple process that involved dropping in types of objects, agents and terrain and setting any desired events and letting the system run.
8.2.2.4 Uncertainty and Quality Assurance
Uncertainty is the predominant purpose of emergent systems, as it gives rise to new and unexpected behaviour. Ironically, it is also the main drawback of emergence in games. When human players are introduced into an emergent system, they have the ability to use the system in ways it was not meant to be used, change things in the game that the developer had not expected and play the game in ways that could not be foreseen. Furthermore, human players seem to have a perverse drive to intentionally push the game to its limits, exploit its weaknesses and to make it break. Consequently, game developers’ fears of using emergent systems are justified, in that if a game has loopholes, exceptions or problems then the players will not only find them, but will actively seek them out and exploit them.
No unwanted uncertainty arose in the EmerGEnT system. However, it was a relatively small, controlled system and there were no humans interacting in the environment.
The larger an emergent system (more entities, rules etc), the more complex it will become and the more variations in behaviour that will be exhibited. With the size of current games (e.g. role-playing games), the use of an emergent approach would require a different approach to game development. It would not be possible to predict every way the players will interact with the game and ensure that it works.
Conducting extensive player testing would catch most of the problems and it would be necessary to focus on problems that will break the game or make it less fun.
8.2.2.5 Ease of Feedback and Direction to Players
Giving feedback and direction to players is simple in scripted systems. The player is simply moving through pieces of the game in a prescribed way. They are given goals, narrative and objectives and then play for a while. The player is led through a prescribed story, with pre-planned intervals of gameplay with very specific objectives.
It is easy to give the players feedback and direction because it will be mostly the same for every player and the required feedback is easily identified as the possible and correct actions are known.
Giving feedback and direction is a much bigger problem in emergent games. In emergent game systems, the range of possible interactions and actions by the player is
far more extensive and there is a lot more uncertainty. Consequently, the problem in giving feedback and direction in emergent games arises because players require a lot more feedback in these games, while at the same time feedback is much more difficult to give. This issue was not investigated in the EmerGEnT system as there were no players interacting and no goals or objectives.