I'm in Washington at the 2005 Serious Games Summit DC. The venue has WiFi and I'll be covering the sessions here, as best as I can keep up, today and tomorrow. Stay tuned below for updates throughout the day.

There are numerous tracks at once, and I can only cover one at a time. You may also want to check out live blogging from one of my Georgia Tech masters students, Matthias Shapiro. If you're also liveblogging, let me know and I'll add a link to your coverage here too.

Keynote: What's so Serious About Game Design? The Art or the Science (Dr. Peter Perla and Douglas Whatley)
Lecture: Healthcare and Forestry -- Half-Life 2: Meet Serious Games Modding (Tim Holt)
Panel: Assessment for Interactive Training Applications (Jerry Heneghan et al.)
Lecture: Get off the couch with Eye Toy and Physical Gaming (Joe Brisbois)

What's so Serious About Game Design? The Art or the Science
Peter Perla, Director, Interactive Research Center for Naval Analysis
Douglas Whatley, CEO, BreakAway Games

Perla discussed the "science of wargaming" -- arguing that wargaming is a science, not just an art or a craft. Much of what a physician does looks like an art, but you'd never trust a physician with no underlying knowledge of the science of his craft. Moreover, an art cannot teach people to produce credible games without assuming that creators would all have a "spark of genius." Perla asked if the fundamental concepts of wargames have a science, are there scientific principles to wargames. He presented an approach to this kind of thinking. Perla also argued that the results of such approaches are applicable to serious games in general.

Wargaming, argued Perla, is Not any type of modeling, exercises, campaign, analysis, or computer simulation. Instead, Perla argued that wargaming is a warfare model or simulation that does not involve the operations of actual forces, in which the flow of events affects and is affected by decisions made during the course of those events.

Wargames have several dimensions: time, space, forces, effects, information, and command. Forces, argued Perla, means not only military forces but also civilians and "other entities," as well as things like physics and external dynamics. From this position, Perla outlined the domains of real war: the physical, informational, and cognitive. These are the "actual reality" of war, and the way we can sense and report about war. The science of wargame design defines, constructs, and proves connections between game and reality. Realism = how well the player relationships with the game reflect relationships with the real war domains.

Perla cited four military references for their approach: Carl von Clausewitz (friction and chance), Mark Herman (entropy-based warfare), Martin van Creveld (command and uncertainty), and Paul Vebber (modeling networks in warfare). Based on these approaches, they defined key factors for operational warfare:

Friction, destruction, disruption, and chance
Entropy, inherent energy unavailable for carrying out missions
Entropy leads to uncertainty, which military command systems exist to overcome
Command counters friction/direction
Descreasing in one dimension increases others
Success = relatively better control of entropy

As wargame designers, the goal is to represent these key issues using only a small number of variables, the six dimensions previously described. Perla argued for three interlinked topologies: operational (the material world), command (structure of the command system), and informational (structure of information flow between the domains). These topologies are the interface between the players and the game, and between the game and the world. The players must relate to these topologies in an analogous way to how they relate to the world.

Realism thus = the degree of agreement between how the players relate to the game's universe and how real soldiers relate to real war.

Perla suggested that this approach is applicable to other domains. He argued that you must identify the basic scientific principles, determine the philosophers who have thought most widely and deeply about the subject matter, what basic concepts your game must represent, and how you can make them tangible in your game universe.

Doug Whatley then presented some ideas about how to take Perla's approach and make it more pragmatic. Wargaming has been a longstanding approach in the military, even if not on the computer. The military is thus very experienced at taking apart a problem and addressing it in game form. Military wargaming thus offers a good approach for many other domains.

Whatley defined a Serious Games as follows: a product that is not specifically entertainment, but which uses the techniques and processes of entertainment to achieve a specific purpose. We want to find a new way to do things, argued Whatley, using game-based learning techniques in interacting with the real world.

Whatley then offered some remarks on the production process for games, showing a timeline of the development process. Serious Games don't yet have a standard development process, which is negatively affecting some projects and approaches; some government contracts don't allow for more decision points or iterative design, which can very much harm the development process. It's a very bad idea to assume that an idea on paper will pan out just because it looks good.

The first phase is the Concept phase: what is the concept for the game. A vision statement for the product and a separate concept statement for the reason you're doing the project in the first place.

The second phase is the Design phase, in which a design document, technical design document, and risk mitigation document are created.

The third phase is the Prototype phase, in which the team proves the concept.

The production phase is the actual development phase, where the majority of the work is done.

The next phase is Testing, in which the product is validated. Whatley noted that some government clients don't like to pay for this phase, but that is changing.

The final phase is Support, for those who need to manage ongoing development and alterations.

Whatley then looked at the concept phases and the challenges in defining the types of products. There are a number of different games, which he borrowed from "Dr. Thompson," a speaker at a conference he attended last week [sorry I don't have the full name/citation].

Training: something akin to the military model: a metric, task, and developable skills, which are measured against some norm or baseline.

Education: problem formation and the expansion of cognitive ability, to apply learning in different, new, and non-existing contexts.

Simulation: a speculative exercise with rules , goals, and containing a disequilibrium outcome. It is used to test an adaptive skill set within a planned context.

Play: everything old is new again: the outcomes are often unknown and unexpected. We attempt to let kids make things explode.

Toys: toys are fun objects that allow one to explore the woodness of wood. It is the simplicity of studying an object by the use of great imagination. Whatley cited virtual worlds as promising environments for play and toys.

So, in a concept document, Whatley advised taking Perla's methodology and mapping into it the type of game and goals that you want to apply to it.

Tim Holt, Healthcare and Forestry -- Half-Life 2: Meet Serious Games Modding
Tim Holt, Research Assistant, Oregon State University

Holt presented two applications of Half-Life 2 modding for serious games. He discussed how to make a game relatively quickly using modding.

Exploring serious game space is difficult given the immense challenges of development. Putting the energy into development, let alone pedagogy or assessment is near impossible; researchers are the "98 pound weakling."

Modding gives you access to top-end games without having to spend a million bucks. Open source game engines are out there, but they may not be supported or may not work properly. But modding can give you a proven, commercial platform for a very low cost. Modding is not like traditional development. Normal development means large staffs and resources, but modding relies on the commercial game as a platform for modification and addition. The game companies endorse this use of the games; it is not a legal problem. Modding relies on taking what is already solved and using those solutions as bases for appropriate problems.

Half-Life 2, for example, provides sounds, 3D modeling tools, level editing, and code samples. Game companies do this so that the public will keep the game fresh and new. The game engine for Half-Life 2 provides all the code to get the stunning visuals, agent AI, physics, and so forth. You can compliment these tools with free tools, such as the limited version of SoftImage, a 3D editing tool. You also can reuse all the art, textures, models, and so forth from the original game. Holt counted around 5,000 existing hi-res textures for use with the game. Of course, Half-Life is a post-apocalyptic world, so much of that is grungy art, but you can also create your own. Valve and other developers encourage modification because it can also create new resources for them; they have hired a number of employees from the mod community.

As far as restrictions go, you can't make sell your mod. You also need to own an original copy of the game to play the mod; the mods run on top of HL-2, for example. And, you can't move content between games, for example from HL-2 to Unreal. These issues are covered in the EULA for the modding tools.

Holt then showed some samples of modes. In a health game, he took the characters from HL-2 and turn them into doctors. He also worked on a forestry mod, taking grass fields and making a virtual exploration for the real forest. The game engine has a system built in for ground cover. Holt built on this, modifying the grass for trees. In the forest mod, Holt made the player smaller and slower rather than making the trees and environment larger.

Holt recommended that would-be modders start by making simple changes to the original game. First, play the source game! Figure out what it can do. The games are cheap, so it's easy R&D. Then get the SDK and evaluate what it can do. Then go back to the game and play it again, thinking about what you can do with it given what you now know about the SDK's affordances. You can also find resources like wikis, mailing lists, and websites by and for modders. Then explore the game's content, with a tool like GCFScape which allows you to explore the internal assets for a game file. Then make a simple change to a game.

Assessment for Interactive Training Applications
Jerry Heneghan, CEO Virtual Heroes, Inc.
Elaine Raybourn, Sandia National Laboratory
Priscilla Elfry (sp?), NASA
Jeff Tateman (sp?), Duke Medical School
Frank Bozeman (sp?), 3DE Solve
(sorry if I got the names wrong; they were not on the program list) This panel confronted "The White Elephant" of serious games -- the next evolution of where Serious Games is going.

The panel looked at types of data captured during games, how After Action Reviews (AARs) work in learning games, and how students are tracked. "If you don't track it, it probably didn't happen." Heneghan argued that this goes beyond "crosshairs on targets."

Assessment needs to be quantifiable and cannot be strapped on at the end. One solution is to tack the game onto an LMS or other tool for measuring traditional systems.

What types of state data are tracked in a game?
Elaine discussed a system developed for training officers. Adaptability in this context means being able to adjust to novel situations and adapt to new interpersonal encounters, negotiating "win-win" situations effectively, and seeing new possibilities in unusual situations.

To assess these activities, we need to address the question of flexibility; there is not one "best way" to be adaptable. In their system, there are new roles in the simulation that can be created ad-hoc. Observers can evaluate, in game, the process that is taking place while new techniques are in use. Engendering conversation and awareness are the principle goals.

Jeff discussed computer-controlled mannequins used to produce patient states. The learners try to take care of the patient, and the performance is videotaped and analyzed with the same tools used to analyze basketball player performance at Duke. But high-fidelity simulators don't scale -- the mannequin is $2-3 million and can only service 2-3 people at the same time. Computer simulation can help improve these figures.

The program called Studio Code, adapted from sports, is used to mark up and discuss the performance in the game. The system is a sort of note-taking tool that can time code responses on video. Jeff then discussed using this data over time to demonstrate longitudinal performance.

Frank discussed a task training for a specific military specialty. Retention from lectures is about 5%, whereas retention from hands-on training is upwards of 75%. How close to that 75% can we come in games, and how close are we getting on the basis of individual students. Assessment and validation are two things: assessment has to do with an individual student's performance. Validation has to do with the courseware content itself, and determining if it is teaching the right things. Assessment with soldiers in this case provides very specific learning goals. An XML file defines 30 - 60 steps in a training routine, which drives the student's interaction with the virtual world, and the game tracks every action with timecodes. At the end of the lesson, they compare the student actions with the goals and generate a report, which is displayed to the student and to the Army's LMS, providing an overall view for the trainer. The instructor thus becomes less about "death by powerpoint" and more about coaching and individual guidance.

Priscilla discussed the problem of "surface services" for spacecraft management -- the operations stuff required for making spacecraft work when they are in service in space. Much more involvement up front has allowed the technicians and the floor operators to become more involved in the process of managing the equipment.

What types of things comprise a state-of-the-art AAR?

Elaine explained that the evaluators approach the process as an assessment and a training tool. While some roleplayers conduct a negotiation, the observer will note the core competencies for the skill and timestamp them to review them later. The evaluators provide a quantitative score at this dimension and present them back to the user. The users are then encouraged to consider the improvement possibilities for future runs.

Frank discussed the use of specific instructional design models for different portions of the experience. The XML file provides a basis for the portions of this assessment. These systems are used as "training wheels" which provide hints. During evaluation, everything is tracked and the student is expected to perform according to the "ideal" model... these two are compared, generating a comparison of the two paths.

What happens after the data is collected? How do you link these data to Learning Management Systems (LMS's)?

Priscilla discussed the fact that the Apollo program did not keep the records from training on that program, or the records are lost, or the people who worked on it are dead. NASA thus assumes a 30-50 year time horizon for keeping and managing data. NASA is currently saving everything, which isn't a long-term solution.

Frank explained the use of video and interactive sessions. The availability of the assessment data allowed more precise focus for the trainers. Instructors thus use their time in a different way, which changes behaviors for training over time.

Jeff discussed "objective-driven" measurement: something you can "truly measure." In Duke's curriculum, students must learn over 20 pages of objectives, and they track where each objective takes place in the curriculum. Using this method, it is easier to identify where things are going well and where they are going poorly.

What needs have been uncovered by current applications?

Jerry discussed the use of VoIP in America's Army as an example of new approaches to both allow communication and capture that communication as data packets. To be successful, it's important to involve the instructor in the process of designing new content; canned scenarios designed from above won't work; instead, you need tools that instructors can use to challenge students on the fly.

Priscilla mentioned the use of high-end 3D simulations for AARs. Using simulations accurate to the inch, they have been able to save millions of dollars in early design errors. However, this also meant that people had to go to a special lab to interact with the system. To solve this, NASA has been moving these Silicon Graphics applications to the desktop as much as possible. Two years ago it was impossible, but now it is becoming more available due to the rise of game technology. NASA is working on a small project to move some very high end simulations piped through a game engine for use on the PC desktop.

Frank emphasized the importance of instructors: it takes intense cooperation with the instructors and the subject matter experts to create reasonable simulation content that actually achieves the goals in mind.

Get off the couch with Eye Toy and Physical Gaming
Joe Brisbois, SCEA

Entertainment can help solve the obseity problem in America. 65% of US adults are overweight, 31% are obese, and 15% of children are obese. Obseity is on the rise. Lack of exercise is partly caused by community designs that discourage activity, poor diet, limited healthy food choices, lifestyles that rely on convenience foods, increased time in sedentary activities, reduction in PE classes and after-school programs. Conventional media are often accompanied by snacking. Sedentary behavior leads to more sedentary activity.

Physical games are those that require the player to do physical activity to manipulate the input device. They use devices like pressure pads, USB cameras, microphones, instruments, and fitness machines. These games encourage players to learn new physical movements that they master over time, thus taking on the same muscle development patterns as skills like physical fitness.

The design of such games need to first introduce a new physical movement. The player refines this movement over time, and becomes quicker. The game then increases the challenge and is able to perform longer periods of gaming. Eventually the player is able to control their body in ways they didn't know they were able when they started playing the game.

Beyond the obvious benefits, physical gaming increases self-esteem, bridges the gap between non-athletic, self-conscious kids and physically active ones, and give children a means to exercise in their own home with world class experiences.

Brisbois then showed numerous video demos of dance pad and Eye Toy games. These are hard to reproduce here, but the products covered were Dance Dance Revolution, Eye Toy: Play, Eye Toy: Antigrav, Eye Toy: Kinetic, SingStar, Karaoke Revolution, Donkey Konga, Guitar Hero (the last four of which require specific peripherals).

Looking toward the future of physical games, all of the devices (including next gen) require new peripherals, which bring up shelf space issues (for example, Guitar Hero will be down under the shelf, not on the prominent eye-level shelf). Demoing these games is also hard in retail... most retailers can't (or won't) support demos for these products. Working with ex

Comments (2) | Post a Comment | TrackBack (1)