Integrating Manual and Automated Game Balancing

Game balancing is the process of systematically modifying parameters of game components and operational rules in order to determine satisfactory configurations regarding predefined goals. Such goals can for instance be to maintain a proper of level of challenge for the player without making the game so hard that they quit, or preventing the emergence of any dominating strategy in multiplayer games so that a unique, de-facto standard for playing materializes.

While game balancing is typically considered a manual activity carried out by game designers and developers following intuition and simple heuristics and decision rules, striving for semi-automated balancing approaches holds considerable potential benefits. For instance, it may help designers to uncover multiple different sets of valid parameter configurations that enable different ways of playing a game, or it may even uncover unconventional settings that a human individual would rarely come up with. Automating parts of the balancing process is an emerging research topic and initial approaches have demonstrated the use of evolutionar algorithms for that purpose.

For this thesis, the author will apply an existing, integrated game balancing process to a given video game. To that extent, the goal of the game shall first be specified in a formal, measurable way. Then, both manual and automated balancing shall be performed. For the latter, an evolutionary algorithm must be properly designed and configured. For this purpose, the author may use and adapt pre-existing outcomes of a project seminar. Finally, results from both manual and automated balancing must be combined, possibly in multiple iterations, to create a balanced version of the game. The results of this process are to be documented in the thesis.