Live from the Serious Games Summit DC
October 18, 2004 - by Ian Bogost

I am back from our London show opening (pictures soon) and here in DC for the Serious Games Summit. The conference has a very "official" feel, thanks to CMP, which is an important step. Oddly, no badge lanyards tho, only pins. Aren't lanyards the sign of officialness? Anyway, I'll do my best to cover the conference here, but it's multiple track and I'm doing a talk and a panel, so you'll get my view of the event rather than a general view.

Update: coverage continues below as it happens, so to speak, from the back of the room. As usual, my coverage here are paraphrases and summaries of the speakers' presentations, not verbatim transcripts. Unlike usual, I'm a bit more fragmented and split between sessions, and I had a number of meetings during the conference, so apologies for the fractured coverage.

Update: I've recorded all the sessions I covered. I apologize that it's not complete as usual, but I had two sessions of my own, and my website crashed and lost a couple of sessions, so, hopefully at least you can get a sense of the sessions I attended. See you at the SGS in March at GDC.

Covered sessions:

Keynote, Jim Dunnigan
How Can Games Shape Future Behaviors (Jim Gee, Debra Lieberman, Elain Raybourn, David Rajeski)
Assessment and the Future of Fluid Learning Environments (Aaron Thibault)
Homeland Security: Uses and Opportunities for Simulations and Games (Julia Loughran)
Experiential Learning Assessment Strategies, David Williamson Shaffer
Real, Reel, and Surreal: How Games Impact Perceptions, Johnny Wilson

Keynote, Jim Dunnigan

Why did it take 30 years to have a conference like this? Jim started in the late 1960s producing simulation models for wargames, Wall Street, and other groups. His background in statistics helped him realize that the audience for games is much larger than those who understand them.

The problem with serious games is that it's not a hit-based business. There are thousands of games to be made, and you have to train people to create them. This is quite different from the commercial industry, in which you can have a few hits and retire on your royalties. Only 1 or 2 percent of the population is capable of grasping how games work.

Paper wargames were for geeks waiting for their PCs to arrive. Players of these games also became the first computer game designers. Games are a series of algorithms, a flow of procedures. Everything else is eye candy. The success of a Serious Game is creating addiction. But serious games are different than entertainment games. If Serious Games are something like "games that you have to play," then they have to create addiction out of the inherent elements of the job. The majority of people in this country work in the knowledge industry, and many jobs are mind-numbingly boring. But, you can make an entertaining simulation of anything.

The market for these games is huge. You're up against the industrial and corporate training industry, which has to be over $100 billion. The enormous growth of computing power is revolutionizing that industry. This includes AI, touch screens, voice recognition, photorealistic graphics, and so forth. The computing power is useful for serious games developers, because you can present the material to people at any level -- from blue collar to executives. This can create addiction.

The problem is in the engineering. You need the proper tools. Luckily, the tools are out there. Wargames don't have 100% fit with non -wargames, but there is a 40-50% fit: a lot of war is about training, leadership, management. 90% of the military don't fight; they work factory jobs in very harsh conditions. The military is a huge market for every kind of serious game. In the last 20 years, the military has been buying commercial if they can't build it themselves. They are becoming aware that they have problems with non-combat people that they need to solve. For the first time ever, in Iraq the non-combat troops are taking more casualties than the combat troops. This is a problem the military needs to solve, and an example of an opportunity.

Of the three modes of training (film, reading, and doing), film is the least effective. Reading is next. Doing is at least an order of magnitude more effective; it's on-the-job training (OJT). There is a tremendous market for OJT in dangerous situations -- where the first military games came from. Just think of mechanical simulators, which were first created in the 1930s to help pilots learn to fly on instruments. The first simulators did save lives, and serious games for dangerous occupations (police, fire, military, combat physicians, first responders, and others) will save lives. Games are a more realistic, safer, and cheaper alternative.

Operations Research and Systems Analysis are your friends. You go out, take a situation, figure out the parts, figure out how they work, and reassemble them. It's a tool, not a panacea, but it's a good tool. Turning any situation into a game can be done, but it's hard work. In the military, look for people who play manual wargames -- anybody who can go to that effort to create a successful experience have the skill to take things apart and understand the experiences. Sales people also need a lot of skills, and they work in a massively competitive field. "That crazy geek" could make you a fortune. Not all savants are idiot-savants.

When you're developing a simulation, always keep your eye on the goals. What is the point of this job? What are you supposed to be doing? Learn how to deal with that. In most cases, the goals are fuzzy (think Dilbert). Take customer support. One could do a customer support serious game engine and customize it for the specific needs of a group. Game engines and recycled game systems are important for these applications. Some commercial technology you can use -- it's done modularly and you can get a basic toolkit of the modules you need for, say, AI, NPCs, autonomous agents, 3D, and so forth. There's a lot of technology you can buy and get support for, so you don't have to reinvent the wheel.

If you're going to be doing these games for a living, and assuming you have the skills, and you've taken apart the assignment, do prototypes quickly. Use rapid application development systems (RAD), do things as immediately as possible.

Test your games against reality as you know it. Of course, your client may have a different view of "reality" than you do. What is the real goal, and how do you resolve all of them.

Most jobs are oriented toward process and procedure (they're dull). That means you can design with those processes and procedures, of course you have to figure out how to make them entertaining. With any type of job, there are laws and regulations, and you can build a game around them.

AI will convince the majority of the population (the Turing Test). Think of Eliza. It works, at least well enough. If your average player is convinced, that's enough.

There as not as many problems as you thought, but you have to solve the big ones.

How Can Games Shape Future Behaviors (Jim Gee, Debra Lieberman, Elain Raybourn, David Rajeski)

Jim Gee
The idea of "transfer" has had three stages of evolution: (1) transfer is easy. (2) Transfer is impossible, give it up. Teach if you want people to know something. (3) Transfer is possible, but it's hard, you have to work to get it. The latter, current idea doesn't give with the news media's impressions, but it's the current one in educational theory.

Cognitive idea of transfer:
- It isn't cheap: hard to get changed, new bahvior in different contexts. Gaming is quite good at it.
- It isn't quick
- It isn't easy

Learning is in the system you build, not the tool. You get transfer out of the system. Games are good in some systems and bad in others. Content is a process, not a set of facts. Performance orientations are required. You have to activate prior knowledge and you have to take into account the context of the learner.

Some principles:
- Undo Folk Theories. Create motivation by showing the learner what thebig picture is.
- Create motivation through purpose
- Create a pro-cocial network (either sociable, or socially good)
- You need to achieve mastery before you can transfer.
- It takes lots of time and practice to do it
- Teach less, get more.
- Teach for understanding not for facts or fixed procedures. If learners understand the system, the content will transfer
- Stay close to but at the outer edge of the new regime of learning you want.
- Encourage deliberate self-monitoring. "What am I learning? Am I learning"
- Give feedback for unerstanding, not facts.
- It helps to use contrasting cases. Avoid ritualization.
- Situate new language within action-images, not just words
- Use dialogue and debriefing
- Use prediction and expectation
- Give verbal information just in time and on demand. No lectures, no talking heads. Create readiness for overt telling
- Demonstrate the transfer overtly.
- Prompt for it, show the person that you want this
- Assess future learning, not initial performance. Look at how you did in a different, extended period of time.

Examples: Rootlearning.com, Full Spectruium Warrioer

Debra Lieberman
The researcher who understands the user is important. There is an art and science in creating educational media expereinces. Many principles are the same from PBS as in games (Lieberman worked with PBS in the 1970s). Lieberman is interested in changing behavior through games, which is compelling but harder than information transfer.

In an informal study of learning preference, cChildren 6 - 10 only 2% picked books or videos over videogames as a medium for learning. Videogames put you in the action and give you the expereince. Learning by doing is very effective. It's important for designers to focus on the goals of the games. Many clients will know the topic, but not specifically what the game ought to be teaching. It's sometimes easier to ask what the optimal outcome study would look like, then work back to the game.

Games are real experiences. We respond to media almost the same as to real life. Seeing and interacting with media is social, We have emotional sresponses to media that are like interactions with people.

Learning and behavior can be effected through presentation by media.

Games are especially powerful because they are participatory. They show feedback very well. Games are engaging Participation makes the player pay close attention. It demands thoughtful planning and decision-making. It demands learning in order to succesed (don't learn, then you can't succeed). Planning, decision making, and learning content to succeed are fungible in a game. Behavioral, ratehr than factual learning is a good strategy.

Games offer consequences. These are not abstract or hypothetical; they are represented in the game directly. The player is the character and identifies with him or her. Success and failure map directly to the player's actions; their ego and self-image are invested in the experience.

Games provide role models; characters who the player can learn from and understand their behavioral experiences. The game gives you rewards for behavior, points, power, rank, and so forth. This positive feedback in the game can encourage it in real life.

To those concerned about the negative effects of games: we can encourage to look at the positive potential of them. Games can be part of a larger context.

Elain Raybourn
Intercultural communication led to her interest in gaming. How do we induce change in ourselves if we are the game players?

Games can be used for Intercultural communication: Games in environments to teach culture shock, for example. Understand the realization of change, and how they can feel a phenomenon. Through gaming, we can provide a person with a sense of what to do in a dangerous situation.

Social process simulations create an unscripted environment that creates cues or contexts for an engagement. A social process simulation to recreate a situation and present it as non-threatening could present an environment in which the player challenges their assumptions. Then you challenge the player to solve problems within the environment. In these situations, we negotiate the outcomes through conflicts and through feedback mechanisms we learn about our feedback with ourselves.

But, how do we get people to execute behaviors in real life similar to those in the game? Identity and power in online environments in games provide a window onto this solution. Face to face environments do not translate directly to computer game environments.

Games are a series of patterns that give you a sense of how to perform in the future. Developing new novel behaviors to recognize new patterns is the benefit of games. We want people to be addicted to real life, to thinking and feeling, and the gaming environemnt can communicate that. It's not enough to create an addiction to the game; you want the player to be addicted to the principles in the game that they can replicate in real-life. There are culture-general skills that you can learn through games.

David Rajeski
David described himself as a "Confessed policy wonk." But why does a policy wonk care about videogames?

David told a story about a small climate model sim from the Netherlands he saw a number of years ago. The model gave hima sense of how things would happen. Within two months, David had learned more about climate change from this game than from any supercomputer and any policy panel. Stop thinking about eGovernment, and more about games. Instead of eGovernment, we want gGovernment.

We face a huge number of long-term problems. Budget, energy usage, etc. Actions and consequences are massively separated in time. These issues also present collective action problems (e.g., how do you distribute a flu vaccine). We have situations where we need policies of persuasion. We need to get people to act in certain ways to do things that are important for the collective good. The typical policy maker is running out of poptions. Games have an opportunity to fill this gap. Emergent behavior is also a challenge to model in standard ways. Games can provide a good alternative: understand how emergent behavior in that situation works is more important. MMOGs are a massive boom in experimentation of this kind. The last set of problems is called a "black swan" -- something that happens that surprises us but we can't figure out the probability. 9/11, the falling of the Berlin Wall are examples. How do we collectively prepare for black swans. In government, you see these every day.

Games bring four things to these opportunities:

(1) Topsight (from Mirror Worlds), the rarest commodity known to humankind: the ability to see an entire system at work. Zoo Tycoon and Sim City give topsight. Topsight leads to foresight. When you have a deep understanding of actions and consequences, you build foresight. Topsight is one of the basic definitions of leadership: the ability to sense the whole.

(2) Collective Wisdom. "DARPA Hard" problems are messy. We can't solve then with 10 or 150 people. If someone solves a problem -- no matter who they are -- the government should want to know about it.

(3) Soft Failure. Failure in a political sense reduces the tendency to experiment. If you can create an environment where failure is possible and encouranged, then you will learn. Games provide this possibility.

(4) Practice being Surprised. An exercise machine for your frontal cortex; you want to learn how to respond to surprises.

Assessment and the Future of Fluid Learning Environments (Aaron Thibault)

Note: I found this presentation almost completely impenetrable. As such, my notes may not represent the speaker's intent accurately, if at all.

Fluid learning environments adapt to a users behavior over time and can therefore deploy specific learning content for each user. Instead of reading a textbook, you could go through an environment with friends. This is a more immersive way to learn, but how do we assess them?

One answer is, we assess each other all the time. A subject matter expert can quickly and easily understand a lot about an interlocutor. Multiplayer online games have complex social structures, in-game economies, emergent social behavior, etc. These constraints can create opportunities to detect learning in the environment. So, you either have to build a database of all the behaviors, or you need a way to record all the actions inside the game (provided you can interpret the recording). Mentoring (formal and informal) in online games (or any organization) takes place naturally. Some things are unique to such organizations, but others are important no matter where they go. When you get to granular areas in the game, you can use other methods (behaviorist, reading, etc. ) to measure performance.

This notion of assessment implies that games can teach. Efficacy studies for games are needed, along with a process to see if people actually can learn in games; there is not yet a silver bullet that shows precisely what people are learning, and what they can best be learning in games.

But, we do know that humans need emotional reinforcement. There are different kinds of learners and different kinds of learners. Richard Bartle has a well-known methodology for why people play (primarily for online games):

Socializers - interested in other players
Player Killers - interested in disruption
Achievers - work toward goals
Explorers - go through the environment, absorb new experiences

Game designers make something fun. Instructional designers want people to learn. The latter use the well-accepted practices and scaffold learning with that approach. The imagination in instructional design is thus less of a lengthy process than that in game design. The merger of fun (or engagement) with learning constitutes a new field that demands greater contribution.

Andragogy (Knowles) is a theory of how adults learn - Games are constructivist environments where you can do stuff, that demands self-assessment. You need to be able to engage what you are doing quickly. This sounds a lot like what happens in games.

Traditionally, assessment is comprised of standardized testing, grades, and so forth. We already know these. It's likewise easy, even if not easily quantifiable, to determine if you are having fun. To assess in games, you need a really great way to model the system, which you can use as the template for assessment later.

Biometric assessment can also be used with real-time gameplay to determine what parts of the brain are used for different types of play. Eye tracking, brain scanning, biosignaturing, cardiorespiratory, etc. are possible -- but not on the console.

Assessment can be used on learners, teachers, peers, content, technology, context, data, trust, etc. Can social software assess our ability to trust our

Constructing a framework for assessing learning could include:
- Cognitive psychology
- Game design
- Business negotiation
- Complexity theory
- Social network theory
- Machine learning
- Neurobiology
- Educational theory

We learn effectively through collaboration. innovation is greater in a web of ideas. We have an ecology, a cultural organism of learners.

Learners are the game players, but they also have construction in those environments -- they are the game builders. They learn through building things in the game.

Teachers like intelligent agents in the game are useful to some extent, moreso are humans, mentors. Evaluators are like feedback systems.

Guided learning is possible in peer-to-peer environments, social networks, and so forth, but we haven't formalized the process.

So, how do we do assessment?

It has to be...
- Social (networked)
- Real-time and changing
- Engaging and fun
- Embedded

The environments and the learning can be fluid, changing, and abstract.

What can you do with these environemnts? You can create an adaptive learning ecology. An agent-based interface to this environment goes beyond the notion of content and learning objects to include engaged experience, a guided pathway. Its assessment depends on a collaborative network of people and agents.

Possible learning: Establish, improve, sustain, or change content. Integration, operations, you can change an understanding of behavior.

Intersection of sciences: System Intelligence, System Architecture, Peer to Peer, MMOGs, all come together in the right way to create the idea of fluid learning.

"ALE" is a hypothetical solution for integrated learning. (Aaron showed a series of visual graphs of the combinations that are impossible to reproduce here).

Thought communities offer assassment through engagement. Collaborative tools can be used to assess in this environment. Embedded knowledge in the environments requires an art of modeling -- capturing the essence of a subject in an environment.

Example: whyville.net, by Jim Bower is a science and citizenship-based system that assesses through group achievement in microgames. it has 900,000 registered users.

AI could be used in a learn for the learner system: the AI could learn from the human participants in an environment. An AI can scaffold with subject-specific knowledge. This can create a pathway to guided learning.

Family collaboration is a group of people the player identifies with. Members become aware of what connects them through specific knowledge.

A starting point for assessment:
- Embedded knowledge
- Intelligent collaborative tools
- Virtual tools with process maps
- Social networks
- Real-time feedback
- Human interaction
- Ranking for humans

Learning opportunities on demand still require humans in the loop.

Homeland Security: Uses and Opportunities for Simulations and Games (Julia Loughran)

Julia covered a project her company ThoughtLink is working on for the Office of Domestic Preparedness. What's out there, what's the market, and what are the requirements for homeland security exercising?

Julia avoided the sims vs. games distinction, but made the basic point that simulations and games are generally synomymous, or at least for the puropose fot his presentation.

The Office for Domestic Preparedness is a group that prepares the country for acts of terrorism, both preparation and prevention. The group has a group that supports training, but primarily they provide state-based grant money. They wantedt an assessment of the current state of models for homeland security to map their requirements to the current and future potential for such systems.

Currently in Homeland Security, training products are purchased based on momney from Washington. Each state provides needs and risk assessments. The states sometimes pass that down to locallities, and the localities pass the money on to local service bureaus. Research and technology products come from individual jurisdictions. This makes it hard to market COTS-type products.

Today, Training and exercising IT&E) is primarily face to face. This includes awareness training (in a classroom, in seminar games (table-top exercise) that test hypothetical situations, full-scale exercises that take place in the real world as mock scenarios, and top officials exerciess, where two states and different countries performa 5-day exercise. Demand for T&E far exceeds the current capacity. The problem with face-to-face instruction is that it requires instructors and facilitators, which is hard to find and to fund. Games can increase the scope and lower the cost of training for local jurisdictions.

Some examples of products include:

Emergency: Fighters for Life, a computer game that allows the user to practice strategic and tactical decision-making in responding to thirty different accident or disaster scenarios. It's designed for consumers, and so it doesn't include typical training materials.

Human Patient Simulator, an incident response system focusing on medical treatment. It's a mannequin that's used in medical teams.

AEAS, a simulation toexercise decision makers in WMD scenarios. It's an incident command response system that runs on a PC.

Civil Emergency Reaction and Responder Training System, a teachers emergency response crisis for emergency operations centers.

NOTE: I had to leave this session to attend a meeting, so my coverage is incomplete

Experiential Learning Assessment Strategies, David Williamson Shaffer

David talked about "epistemic games," and the idea of scientific management being applicable to other human activities.

Cognitive theorist Vygotsky argues that a complex reaction doesn't consist of a series of separate processes that can be arbitrarily added and subtracted. David gave the example of chess -- the best players see the board as a holistic space, not a space of individual moves. A computer can compete, but it's not competing in the same way. David argues that sophisticated skills cannot be decomposed into simple skills. So why do we keep teaching as though they can? Because we can do it, it's easy. A novice chess player could be trained to spend less time per move, but that doesn't make him a better chess player. Measuring simple variables misses the point.

Sophisticated skills are not captured by simple measures. In a school system, where educational measurement looks like a multiple-choice option, it makes sense that educational content looks like Math Blaster. Teaching simple skills and simple knowledge is easy. But, these skills don't come together into the complex forms that we really care about.

So, here's a different approach. To look at this problem, consider the process of mediation: any problem has stakeholders with legitimate conflicting interest. The mediator is trying to reconcile those interests. The mediator is trained with simulated negotiations -- the interests of each stakeholder are aligned numerically. When kids participate in a mock version of this process, the learn the map of issues in its complex interrelations. This sense is more complex AND it's more like the sense a mediator would make of it.

When kids play Soda Constructor, using the practice of design engineers, they learn about engineering concepts. mycity is a project where students redesign the downtown of Madison WI, using the practices of urban planners. When students play at being subject experts, something happens: skills and knowledge come together in the context of a complex performance. These develop skills, identitiy (a community of practice); knowledge and values (islands of expertise -- new experiences clump onto the knowledge you care about), and epistemology, which binds all those. The students can ask what questions are worth asking, how they go about answering them, and how they know that they have an answer. The "epistemic frame" is a way of describing the way of thinking of particular groups of people.

Professions, or work characterized by autonomy, require you to make decisions in uncertainty, and which therefore require judgment. People who do this exhibit a particular kind of thinking called reflection and action. Our thinking serves to reshape what we are doing while we are doing it. The professional practicum (e.g. residency for doctors) binds skills and knowledge.

In games, we can take the practica and make them accessible to kids. However, we have to do it in epistemic games, one that recreates the epistemic frame of a social practice. An epistemic game recreates the epistemic practice.

Games depend on an activity system and a game engine. The activity system matters: the things they do, the people they do this with, Dice are a game engine; a craps table is an epistemic game.

In an epistemic game, content is free. In the context of doing something they care about, in a community of practice, they get the content. The motivation is realtveily easy, becuase that's what an epistemic game does. What's hard is authentically recreating the frame of the practice.

Real, Reel, and Surreal: How Games Impact Perceptions, Johnny Wilson

Games are threating because they turn control over to the player... the trainer or the educator cannot fix the educational experience. It takes some kind of collateral materials to be able to discuss the content for definite educational goals to be met. Games are chaotic and threatening.

However, games are a very human experience. Wilson discussed Huizinga's Homo Ludens: play is a voluntary activity where the player is immersed into a proceeding of fixed limits with fixed rules to advance in a small group with secret regulations with the yield of no material gain from the outside world. Immersion means the player is taking on threating role shift. With no risk of loss, there is no game. Gaming is also about status: I passed that level, I beat you. Games maintain an idea of secret resources. We play games to entertain principaly, not for material gain.

Games can help us study a subject without the expense and equipment necessary to test the real system, to offer an experience without the risk.

Wilson discussed a large number of games that affected him personally.

Turn based games give you time to think about the consequences of each action. Today, you cannot find turn-based strategy games. For training, education, and thought-provoking games, you need that time to think. Real time games are really accelerated games. In real time, you have some time to think about what your options are.

There is a ripple effect for all decisions. There is interdependency in global community. Evaluating risk versus reward, evaluating negative consequences in spite of "success", Familiarity with geographical, cultural, economic, and political facts. many people don't want players to experience the negative consequences; but you need those mistakes to evaluate and learn. Games allow you to make mistakes before you cannot correct them.

POST A COMMENT

TRACKBACKS