Abstract
Strategy games are a popular genre of games with a long history, originating
from games like Chess or Go. The first strategy games were published as
“Kriegspiele” (engl. wargames) in the late 18th century, intended for the education of young cadets. Since then strategy games were refined and transformed over two centuries into a medium of entertainment. Today’s computer strategy games have their roots in the board- and roleplaying games of the 20th century and enjoy great popularity. We use strategy games as an application for the procedural generation of game content.
Procedural game content generation has regained some interest in recent
years, both among the academic community and game developers alike. When
the first commercial computer games were published in the early 1980s,
technical limitations prevented game developers from shipping their titles with
a large amount of pre-designed game content. Instead game content such as
levels, worlds, stories, weapons, or enemies needed to be generated at program runtime to save storage space and memory resources. The generation of game content “on the fly” has gained popularity again for two reasons: first, game production budgets have risen rapidly in the past ten years due to the necessary labour to create the amount of game content players expect in games today. Secondly: the potential audience for games grew in the past years, with a
diversification of player types at the same time. Thus game developers look for
a way to tailor their games to individual players’ preferences by creating game
content adaptively to how the player plays (and likes) a game.
W we extend the notion of “procedural game content generation” by “game
mechanics”. Game mechanics herein refer to the way that objects in a game may
interact, what the goal of the game is, how players may manipulate the game
world, etc. We present the Strategy Games Description Language (SGDL), a
tree-based approach to model the game mechanics of strategy games. SGDL
allows game designers to rapid prototype their game ideas with the help of our
customisable game engine. We present several example games to demonstrate
the capabilities of the language and how to model common strategy game
elements.
Furthermore, we present methods to procedurally generate and evaluate game
mechanics modelled in SGDL in terms of enjoyability. We argue that an
evolutionary process can be used to evolve the mechanics of strategy games
using techniques from the field of machine learning. Our results show that
automated gameplay combined with expert knowledge can be used to determine the quality of gameplay emerging from game mechanics modelled in SGDL, and that algorithms can augment the creativity of human game designers.
from games like Chess or Go. The first strategy games were published as
“Kriegspiele” (engl. wargames) in the late 18th century, intended for the education of young cadets. Since then strategy games were refined and transformed over two centuries into a medium of entertainment. Today’s computer strategy games have their roots in the board- and roleplaying games of the 20th century and enjoy great popularity. We use strategy games as an application for the procedural generation of game content.
Procedural game content generation has regained some interest in recent
years, both among the academic community and game developers alike. When
the first commercial computer games were published in the early 1980s,
technical limitations prevented game developers from shipping their titles with
a large amount of pre-designed game content. Instead game content such as
levels, worlds, stories, weapons, or enemies needed to be generated at program runtime to save storage space and memory resources. The generation of game content “on the fly” has gained popularity again for two reasons: first, game production budgets have risen rapidly in the past ten years due to the necessary labour to create the amount of game content players expect in games today. Secondly: the potential audience for games grew in the past years, with a
diversification of player types at the same time. Thus game developers look for
a way to tailor their games to individual players’ preferences by creating game
content adaptively to how the player plays (and likes) a game.
W we extend the notion of “procedural game content generation” by “game
mechanics”. Game mechanics herein refer to the way that objects in a game may
interact, what the goal of the game is, how players may manipulate the game
world, etc. We present the Strategy Games Description Language (SGDL), a
tree-based approach to model the game mechanics of strategy games. SGDL
allows game designers to rapid prototype their game ideas with the help of our
customisable game engine. We present several example games to demonstrate
the capabilities of the language and how to model common strategy game
elements.
Furthermore, we present methods to procedurally generate and evaluate game
mechanics modelled in SGDL in terms of enjoyability. We argue that an
evolutionary process can be used to evolve the mechanics of strategy games
using techniques from the field of machine learning. Our results show that
automated gameplay combined with expert knowledge can be used to determine the quality of gameplay emerging from game mechanics modelled in SGDL, and that algorithms can augment the creativity of human game designers.
Original language | English |
---|
Publisher | IT-Universitetet i København |
---|---|
Number of pages | 194 |
ISBN (Print) | 978-87-7949-286-8 |
Publication status | Published - 2013 |
Series | ITU-DS |
---|---|
Number | 91 |
ISSN | 1602-3536 |