FP-JonathanChung
From CS294-10 Visualization Fa07
Contents |
[edit] Visualizing Strategy in a Turn-Based Video Game
[edit] Proposal
[edit] Group Members
- Jonathan Chung
[edit] Description
Turn-based video games represent an increasingly popular segment of the video game market, and some of the best titles deeply emphasize the ability to develop both low level and high level strategies in order to obtain victory. With an increase in sophistication of these kinds of games has come a higher barrier to entry for new players. Games like these are dominated primarily by seasoned experts while novices to these games are confounded and find it difficult to grasp the higher level strategies required to succeed.
This project aims to take a popular turn-based game and offer visualizations that educate the player about what moves to make and visualize the abstract nature of "strategy" in a form that can be easily grasped by anybody. It is this latter point that keeps this project very relevant to the field of visualization.
[edit] Details
I propose the following features of this visualization as possible examples of what could be done to visualize strategy in a meaningful way.
- Influence Maps - Show which spaces on the playing field are "owned" by certain teams through the use of color. Influence is primarily based upon units and their placement on the map. The influence amount is encoded using opacity, so that higher influence means a more opaque color while weak influence is shown using a more transparent color. Additionally, when these spaces are highlighted, show why these spaces are valued as such.
- Game Trees - Recommending moves for the player and visualizing this thought process. In effect, visualizing a game tree. On the flip side, also visualize the computer AI and its thought process. (It's identical)
- More View Types - One example is a medic view as mentioned in the first paper reference. Many games feature a "medic" unit that can heal units. I will encode the health of a player visually and visualize medics and their influence. Namely, units that are heavily wounded will appear as a deep red. Additionally, units with a high influence (more powerful) have a larger spread. Similarly, medic units are encoded in green, and their area of influence depends both on how far he/she can move and how much healing power he/she has. In this way, a player can quickly visualize whether he or she is properly using medic units or not.
- Visualizing Uncertainty - Finally, introduce the element of uncertainty using fog of war and adjust the influence map accordingly based on what is known now, and what was known in the past. Actions that happened long in the past will have less of an influence than actions that recently happened. This then encodes a temporal element in this visualization.
[edit] Anticipated References
- Hoobler, Humphreys, and Agrawala. "Visualizing Competitive Behaviors in Multi-User Virtual Environments." IEEE Visualization, 2004
- Niederauer, Houston, Agrawala, and Humphreys. Non-invasive interactive visualization of dynamic architectural environments. ACM Press, 2003.
- Millington, Ian. Artificial Intelligence for Games. San Francisco:Morgan Kaufmann, 2006.
- Wattenburg, Martin. "Thinking Machine 4." 2005
[edit] Implementation
This project will be implemented in Java. The game that this project is based on is Advance Wars, a turn-based strategy game series developed by Nintendo. I am working off an existing open source project called Custom Wars, which is a clone of the original game. I will write the visualization code and AI on top of this project.
[edit] Motivation
I have chosen to base this project on a video game rather than a traditional board game such as chess for two reasons besides the fact that I enjoy playing video games. First, turn-based video games are more complex than chess and present more unique challenges in the visualization realm. While I did not elect to apply visualization techniques to a real-time strategy game, these same techniques are highly applicable to that realm. Secondly, there is less published literature covering video games and visualizations, and I found that this would be a more fun and unique project to pursue.
