Game AI Roundup Week #31 2008: 11 Stories, 3 Video, 2 Quotes

Novack on August 4, 2008

Weekends at are dedicated to rounding up smart links from the web relating to artificial intelligence and game development. This week, we have more a few videos, and one is dedicated to the memory of Randy Pausch; as always, there are some good articles and blog posts for you to read. Remember, there’s also lots of great content to be found in the forums here! (All you have to do is introduce yourself.)

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

Fixing Pathfinding Once and For All

Paul Tozour and another very good article (video included) in the Game/AI blog, this time about the long standing problem of finding a proper pathfinding solution.

I need to talk about some problems we face with pathfinding. In order to prove that these problems still exist, I felt the need to make this video … which will hopefully be taken in the humorous and lighthearted spirit in which it was intended

Chess Programs Are Not Smart

In the Zenpawn’s chessblog the author introduces an interesting paper by Thomas Hall, treating the mechanisms inside the common chess playing programs.

my personal opinion, from the perspective of a software engineer with an interest in philosophy and neurobiology, is chess programs—or maybe the game itself, or more likely commercialization—have been absolute flops at addressing the original question to which they were tasked. Of course, I’m speaking of artificial intelligence. And brute force algorithms just don’t get us any closer to an answer. Anyway, allow me to step aside and present, with much gratitude to the author…

Why chess programs find good moves, but barely understand chess after all.

AI guru lays out future plans

Damian Isla, AI Programmer at Bungie studio, talked at the Develop Conference and Expo 2008, and got covered by two of the most important videogame media.

The next big step for artificial intelligence in games will be making a character walk up to a table and pick up a pencil.

That’s according to Bungie AI expert Damian Isla, who told Eurogamer that in his line of work, problems we assume to be easy usually turn out to be “impossibly difficult”.


The man behind the artificial intelligence in Master Chief’s last two outings talks to developer conference about where Halo AI is going next.


A new kind of a Game

On his blog Sanjeev and his Green Destiny, Sanjeev Chandran comented on a game called Nero developed to use real time NeuroEvolving Augmenting Technologies, by Kenneth O. Stanley at University of Texas at Austin.

Now, I wasn’t expecting to write a post on a game, any time soon. Not even on DotA. But this is just too good a game for me to resist. I have been quite intrigued by the concept of Evolutionary Computation. One lazy Sunday, I started looking for more stuff on Evolutionary Computation. I found a lot. Particularly, a post-doctoral researcher’s blog. Julian Togelius. He’s published a paper on solving Sudoku using Genetic Algorithms. Serious stuff that! One of his other papers have been judged as one of the top ten AI game innovations of the year. This dude was one of the organisers of a Simulated Car Racing Championship atWCCI 2008. It seems there are not a lot of games available that actually use the latest Learning Techniques.

Foundations of Digital Games

Previously known as the Conference on Game Development and Computer Science Education (GDCSE), this year’s conference broadens its scope to cover the breadth of game research and education. The conference is targeted at researchers making contributions that promote new game capabilities, designs, applications and modes of play. FDG 2009 is calling for papers, explicitely including Game AI works:

“…this year’s conference broadens its scope to cover the breadth of game research and education. The conference is targeted at researchers making contributions that promote new game capabilities, designs, applications and modes of play.

Artificial Intelligence:

We solicit papers on artificial intelligence research that provides novel solutions to traditional game AI problems (e.g. path planning, camera control, terrain analysis, user modeling, tactical/strategic decision making, etc.), supports novel game concepts or gameplay elements (e.g. interactive drama, narrative/character development, NPC belief/attitude/emotion modeling, etc.), provides automated or semi-automated solutions to game production challenges (e.g. game design, content creation, testing, prodcedural animation, etc.), or describes the integration of AI technologies (e.g. machine learning, logical inference, planning, etc.) into game AI architectures.”

New game AI using Alice

The blog Kryotech, published a second article on making a simple AI using Alice. Although below is the link to the first article, I recommend to navigate all the previous entries on the blog, as there are plenty of related material on Alice.

We first published a post that gave a very basic idea of how to make AI for games using Alice. We now have designed a smarter AI. In real life, men would retreat upon knowing they will not live longer or that there forces are not sufficient enough. We only had a crude system for this, where if the Player was within five meters of the AI, it would just turn and run. Now, we have an actual retreat system which depends on the kill count of the AI’s forces.

Hell’s Highway

Tom’s Games published a preview after the E3′08 presentation of the Brothers in Arms sequel, and here are some words about the game’s AI:

The player also gets control of up to three A.I. squads, with most missions calling only for two of them. The Assault squad, the Bazooka squad, and the Machine Gunner squad all have various strengths and weaknesses, and can be deployed in a variety of ways, including suppressive fire, straight-up point attack, or flanking duty. Plus, judging from the demo, these AI controlled squads and their respective enemies aren’t completely incompetent like in some titles.

Game Maker Robot AI Programming Competition

In the Marty Blog, the last post announced an AI competition! An arena (a room in Game Maker) with a basic “robot”, and a set of rules about how it is operated. Then people who want to enter the competition will construct AI routines for the robot to follow.

The idea is that I will create a base application which everyone can download and play with - essentially just a room from top-down view, and a simple “robot” object - and anyone wanting to participate in the contest will write an AI for the robot that will compete against others’ creations. Movement and attack variables / objects will be the same across all entries (I’ll be coding those in scripts), and once everyone’s AI script is finished, we’ll draw up a tournament tree, I’ll have the bots compete, and then record the results and broadcast them, via either YouTube or some kind of live Podcast.

Making the A.I. fortify bottlenecks

Red Brick Games, and another blog post regarding the upcoming indie RTS Orc vs Martians AI. Phillip continues his works to make the AI able to detect and fortify bottlenecks in the map.

The work on making the A.I. place towers at natural bottlenecks in the terrain, is coming along pretty well. I implemented a quick, “dumb” algorithm to try it out. Its effect on gameplay is pretty favorable

Do players stop and smell the … AI?

The blog Artificial Intelligence and You (AIandU) posted an article with a bit of a rant, about the repercusions in the game-players side, of the Paul Tozour’s pathfinding article on the game/AI blog (which we are also commenting on the current roundup).

Paul Tozour’s pathfinding thread was linked on kotaku and some people’s comments really seemed…strange to me. Some were actually saying that games like zelda and mario (for snes) had just as good AI as games do now. Other people weren’t quite that extreme, but how could anyone not notice improvements, does animation overshadow AI that much?

Wii MotionPlus and LiveMove 2

LiveMove 2 is a brand new product that fully supports the Nintendo Wii MotionPlus accessory. AiLive collaborated with Nintendo to design the MotionPlus hardware and is offering LiveMove 2 to help game developers take full advantage of its capabilities.

Dynamic, Intelligent Camera

Kotaku commented on an Dennis Dyack’s article, where he talks about the dichotomy in the concept of cut-scenes in videogames, and the approach will be following in his current project, Too Human.

“Too Human will have cut scenes, but I think that we’ve managed to blur the line between what people would consider a cut scene and what people consider in-game. See, part of the reason we as designers want to use cut scenes is because it allows us to be cinematographers, and that’s fine. But in-game, Too Human will use a dynamic, intelligent camera system that presents the in-game in a more cinematic light, at the same time being conducive to good gameplay.”

Indie Roundtable

Gamasutra arranged a roundtable discussion between three of the major innovators in the shoot-em-up genre: Jonathan Mak (Everyday Shooter), Kenta Cho (ABA Games), and the pseudonymous Japanese creator Omega (Every Extend). Here is trascription of part of the debate, about scripting events against randomized events.

“JM: What do you think about… I have this theory that - okay, so a lot of modern games now have set piece gaming, you know, like Heavenly Sword or something. Scripted events happen, so my theory is that you can have scripted events, but you inject… some sort of randomness, and then every time you play that scripted event again it’s always new. So that’s, I don’t know, have you ever thought of creating sort of randomized scripted events? Cause I notice all your games are completely random.

KC: I think randomness is very important. Especially since I write all the games by myself and I test play my games continuously, I like randomness. Also because the developer can’t know how my game… the behavior of the game changes dynamically every time.

JM: That’s good, because I really like random, but have you thought of a scripted event that is random? You know what I mean? So like, on the last level it’s like there were those spinning things and they go away and the boss comes, but it’s still random. What do you think about that?

KC: I tend to write the pattern or script in my games myself, so I try to have minimum sets of script in the minimum sets of algorithm in my game, and another sequence generated from my program. Also, I’m not good at writing scripts or events in my games, so I try to write simple games with random dynamic patterns, but also can be enjoyed by the player; those kinds of random sequences.”

Redefining Game Narrative

Interviewed by Gamasutra, the Ubisoft’s narrative designer Patrick Redding tells about his work in Far Cry 2 where the usual narrative mechanisms are limitated due to the dinamycs of an open world game.

The argument that cutscenes are dead as a narrative form in games has been spread far and wide. The sense that we must push the medium toward a form of interactive narrative that is as strong and vital as the innovations in other areas of gameplay and technology has taken hold with many creators.

Here, Patrick Redding details for Gamasutra the work that he has been doing as narrative designer on Ubisoft Montreal’s Far Cry 2, which is due out later this year for Xbox 360, PlayStation 3, and PC, describing how he and other members of the team have chosen to design the game for fully interactive, player-influenced narrative.

The Last Lecture

I would like to make a simple tribute to the memory of Randy Pausch, who passed away on July 25th, at age 47. Randy was a Professor of computer science, human-computer interaction and design at Carnegie Mellon University, and although he became a best-selling author due to his famed speech “The Last Lecture”, Pausch was the author or co-author of five books and over 70 articles and the founder of the Alice software project.

On his last lecture at the university Sept. 18, 2007 “Really Achieving Your Childhood Dreams”, before a packed McConomy Auditorium, Pausch talked about his lessons learned and gave advice to students on how to achieve their own career and personal goals.

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

Discussion 1 Comments

kofman on August 4th, 2008

Listening to Randy Pausch's Last Lecture was inspirational.

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!