Game AI Roundup Week #14 2008: 11 Stories, 1 Video, 1 Paper, 1 Job

Novack on April 6, 2008

After a quiet week last week, there’s been an avalanche of news and great content in the world of game AI. In this roundup of smart links on, you’ll find book reviews and tips about machine learning, white papers about pathfinding and insights into a few games.

This post is brought to you by Novack and Alex Champandard. If you have any news or tips for next week, be sure to email them in to editors at Also Remember there’s a mini-blog over at (RSS) with game AI news from the web as it happens.

Shortest paths

On his LiveJournal, David Eppstein, a Professor of Computer Science at UC Irvine, published an article entitled about pathfinding algorithms in general. Here’s the introduction:

When I teach shortest paths in my undergraduate algorithms classes, the methods I consider most important (in roughly that order) are breadth first search, Dijkstra’s algorithm, A*, the linear time DAG shortest path algorithm (section 24.2 of CLRS), and Bellman-Ford. Every self-respecting computer scientist knows the first two, but it came as a bit of a shock to me to discover that very few of our incoming graduate students had even heard of A*.

Project Overview: perfectstorm

Perfectstorm is a RTS game study written in common lisp using OpenGL for graphics display and cairo for texture generation. This is also what the official blog says about this indie enterprise:

“I started perfectstorm because i was annoyed with the Artificial Intelligence offered in most commercial RTS games. Although significant academic research is going on, it seldom finds its way into released games due to publishers’ wishes for absolute stability of the “gameplay experience” and the notion of AI as just being responsible for pathfinding plus something. Naturally, AI development can only begin late in the development cycle when the deadline is nearing. The most pressing feature the single “AI guy” most teams employ has to do is implementing pathfinding, which is not even part of AI if you want to see it that way. After that, only a poor makeshift enemy AI can be implemented before the product is shipped. Aside from that, writing good AI is hard. Or so i heard.”

Lost Empire’s AI Too Realistic?

Are reviewers too picky, or is good AI so difficult to achieve? In this case,’s review of the turn-based strategy game Lost Empire: Immortals, developed by Pollux Game Labs, mentions that realistic AI, in this case is… not a good thing!

“The game AI is somewhat unpredictable, sometimes first contact will be very peaceful, other times it’ll be all out of war and you’ll have to concentrate more on defense than resource gathering. Sure, it’s realistic in its way because why would any opposition in reality be predictable? But realism doesn’t always make a good game, especially not here. Instead, you have a huge balancing act on your hands which will test even the most hardened of strategy gamers.”

Concurrency Talk and Ants Demo in Clojure

Rich Hickey published slides, audio and source code for the demo, for the talk he gave on march 20th, about Clojure and concurrency at a meeting of the Western Mass Developer’s Group.

“Clojure is a dynamic programming language that targets the Java Virtual Machine. It is designed to be a general-purpose language, combining the approachability and interactive development of a scripting language with an efficient and robust infrastructure for multithreaded programming.”

Rethinking A.I.

In the second round of this week’s eternal battle between AI developers and game reviewers, here’s an article written by Alex Jordan. It might be a little cynical, but there’s some truth to it:

“Artificial Intelligence. Remember how every game preview ever written used to tout advanced and challenging A.I.? And remember how, once released and reviewed, the game’s A.I. would generally be a point of derision? How did we get there, and where do we go from here?

My primary concern is the A.I. in shooters. In this genre, whether the bad guys feature stellar intelligence or barely enough wit to outsmart a discarded Kleenex, enemy A.I. is often treated as an also-ran. It usually doesn’t affect the highlights of the game itself, typically: guns, blood, interesting levels, a plot (sometimes), and things that go boom. Expectations are lowered and a self-fulfilling prophecy about unimpressive A.I. is established. The result? Shooters with particularly good A.I. become very, very rare.”

Rockstar acquires Mad Doc Software

In typical Rockstar style, this seems like a very politically loaded acquisition. After a controversial Bully port for many Xbox360 owners, along with rumored financial problems at Mad Doc, it makes you wonder what was really behind this deal. Here’s the official blurb:

“Studio that developed 360 Bully port boasts key AI tech and becomes Rockstar New England […]

Mad Doc has a particular focus on AI and networking technologies and its proprietary Mad3D Game Engine and Mad AI middleware haas been available for licensing. In a statement about the acquisition, Rockstar singled out the firm’s technical expertise as a key motivation for the purchase.”

What will happen to MadDoc’s AI technology is anyone’s guess!

Learning Patterns and the Game of Flip

Christer Ericson wrote a great article about an old game called Flip which uses machine learning to predict the player’s decisions:

“Models such as these are known and studied under the terms N-grams models and Markov chains. I won’t attempt to go into them here, but it should be pretty clear that what Flip uses is just a very simple statistical model and that we could form much more sophisticated models along similar principles.”

There’s also a Java demo applet. Be sure to check it out!

Paper: The Semiotic Immersion of Video Games

Here’s some reading to occupy the designers among you:

“The paper analyzes the effect of immersion in digital games using the theoretical apparatus of game theory. The paper illustrates interactive operations and the cause and effect relationship between player and designer, explaining the importance of strategic decision-making and pathing in player immersion. It considers the game function of creating a virtual world and proposes the idea that digital games are not just computer-mediated communication to the player. These games are games of “the moment”, like the game Chicken, and played with apparently great emotion, intelligence, and physical dexterity, although represented in software form. The relationship between the player and the computer is one of sign exchange, precisely the one that semiotics calls semiosis. The paper concludes that the personal achievement of individual players (end-users) accounts for the phenomenon of deep immersion in digital games. Not virtuality, but virtuosity is the strong force in digital game playing.”

The Semiotic Immersion of Video Games, Gaming Technology and Interactive Strategies
Eduardo Neiva, Carlo Romano
The Public Journal of Semiotics I, 2007.
Download PDF

Here’s the original blog post:

Job: AI Engineer at Red Storm

Those of you looking for a change of scenery with a little industry experience may want to check out this opportunity:

“Red Storm Entertainment, Inc. is looking for an individual to work on AAA games for the 360, PS3 and PC. Working closely with the development team, you will work with an established and high-selling franchise to create realistic and entertaining AI experiences. Although experience with artificial intelligence is highly preferred, we are interested in candidates with all levels of experience!”

(Old)Book (Re-)Reviewed

John Pile posted a review of a book published back in 1996 entitled Developing Games That Learn:

“I’ve been reading a book published back in 1996 (eons ago in the timeline of computer games) called Developing Developing Games That Learn (by Len Dorfman, Narendra K. Ghosh) in the hopes that it will have something I can use in my PS2 Dominoes Game.

There are few books around that actually discuss the topic of game AI with specific examples, and I do recommend this book to add to your collection. But, be aware that like many AI books, although it solves a specific problem, it may not be a solution for your particular situation.

One of the things that I really liked about the book was the authors’ attitude toward making both the games and the experience of developing the game enjoyable.”

Reviews of the book on Amazon are not great. Anyone care to share their opinion?


Now for something a little more entertaining:

“PiroFobia is an Agent-Based AI game. Their are litle robot firefighters. The user puts them on the scenary. Adds some water suplies and afterwards, puts fire to the set :) See them work and put out the fire. The really manage themself. It’s fun.”

Situationist Game AI

On his blog My Digital Neuron, Ben Hutchison posted an interesting summary of Adam Russell’s work in Game AI Wisdom 4.

“I have just become aware of Game-Developer/Philosopher Adam Russell through his thought provoking article on Situationist Game AI in Game AI Wisdom 4. I look forward to reading more of his work in the near future.”

The very same Adam Russell that stirred up a “controversy” at the GDC roundtables :-)

Cellphone Game AI

Revival is a Symbian turn-based military strategy game in the tradition of the world-famous game title, “Civilization”. Better AI is slowly making onto portable platforms:

“The game’s artificial intelligence (AI) deserves special mentioning. The program opponents follow the same rules as the player, share his conditions and compete with each other (up to six opponents). And, in spite of that, even the most experienced players will have to do their best to win.”

Interactive Tales

asotir wrote a nice little editorial about “new media,” mentioning AI too:

“Today, the most advanced interactive fiction comes to us in the form of video games. The reader or ‘player’ is invited to become the hero of the conflict, to solve the puzzle, to fight the battles, to gain the victory for himself. He is not merely getting choices, he is actively participating, and the AI is responding according to the rules the game designers have set forth as to what is possible and likely. The player’s own skills help decide the outcome of each step, and the interactivity is a constant flow or mesh between player and AI. In multiplayer games, the many players increase the ‘noise’ or randomness of the outcomes, and add to the illusion of reality.”

Stay tuned next week for more smart links from around the web!

Discussion 0 Comments

If you'd like to add a comment or question on this page, simply log-in to the site. You can create an account from the sign-up page if necessary... It takes less than a minute!