Press Kit Contact Buy Clark Aldrich Designs Bio Books and Articles Blog, Facebook, and Twitter

Thursday, September 15, 2011

Clark Aldrich interview of Will Wright, Warren Spector, and Jane Boston on Games and Simulations

While doing book research on the limitations of traditional media and the need to "learn to do" as well as "learn to know" I talked to three computer game designers:

  • Will Wright (WW)
  • Jane Boston (JB)
  • Warren Spector (WS)

This helped informed some of the Unschooling Rules, including:

  • #unrules01: . Learn to be; learn to do; learn to know.
  • #unrules26: Biologically, the necessary order of learning is: explore, then play, then add rigor.

(See Eliane Alhadeff's Unschooling Rules – Serious Games As Microcosms For Learning for more details.)

Here is what they said:

Clark: What ideas (and types of ideas) are best taught through simulations? What ideas (and types of ideas) should never be taught through simulations?

JB: From my perspective, simulations are best used in four ways.

First, they are ideal for developing an understanding of big ideas and concepts – those things for which experience alone can deepen understanding. It is one thing to memorize a definition of nationalism or to read a passage describing the brittleness or fragility of an ecosystem; it is quite another to enter into an environment where those ideas play themselves out based on your own actions and ability to identify and solve problems.

Secondly, I believe simulations are great for dealing with time and scale. The computer gives us an opportunity to speed up results of an action that might actually take several lifetimes to play out. This allows players to see the potential impact of decisions made now on the future. Dire warnings about issues like the finite amount of drinkable water on the earth rarely impact people’s behavior, yet a simulation has the potential of creating an emotional connection to the information that may have at least some small influence on both understanding and behavior.

I think simulations are good for situations where it is important to give people practice in decision-making before it is faced in a dangerous or critical, real-life situation. Some of the simulations used for emergency personnel provide an opportunity to experience “life-like” situations and react to unexpected and challenging problems.

Finally, simulations are wonderful resources for taking us to a time or place that we are unable or unlikely to experience directly.

Simulations are not appropriate for teaching discreet bits of information (“facts”) or for rote drill and practice. Because they are an immersive experience, they are better suited for those things that need to be learned in context and require active problem solving. There are some topics that are controversial and until we understand better what is taken away from the experience, we need to be cautious in implementing. Is a simulation of a WWII battle an immersive history lesson or a lesson in combat techniques? A simulation should never be treated lightly as “only a game”. Good simulations have the capacity to generate very strong thoughts and feelings in their participants and anyone using a simulation should be prepared for that possibility.

WS: I'm no expert when it comes to training and/or education but common sense tells me that simulations are best suited to dealing with matters of the mind rather than matters of the body. I think there are two reasons for this. First, our simulations are still pretty rudimentary -- we typically simulate only a few forms of sensory input, making genuine immersion in the sim a hit-or-miss proposition. Also, we interface with current simulations in ways that are radically different than the ways in which we interface with the real world -- I mean, there's no real world analogue of a mouse and keyboard or a gamepad! Simulations that utilize more realistic interface elements (cockpit mockups, bobsleds, skis, etc.) are few and far between.

Simulations, then, seem best suited to teaching concepts, tactics, general approaches to situations rather than specific actions. And the limited depth of our simulations today means that, even when we have real-world analogue interfaces, it's imperative that the people operating the simulation provide appropriate context and additional instruction to those who might mistake the partial simulation for an accurate recreation of a real situation...

WW: Simulations are great for understanding processes that are outside of our experience. You can play with time or scale. You can interact with molecules and planets. Many designers tend to map them into instinctive structures that we already have, either through analogy, or through gut-feel. And for a lot of things, of course, the human mind is better than any computer.

Clark: Can games change the behavior of players outside the game? How could you maximize the transferability of game-learned experiences to life?

JB: Because behavior is based on what we know and have done before, games become part of our overall experience. The degree of impact they have is much like the rest of life and is dependent on what the player brings to the game in terms of knowledge, skills, and attitudes; the context and circumstances in which the game is played; and what happens before and after the game. I’ve facilitated “off-line” simulations in which some participants exhibited extreme forms of emotion and carried feelings from a simulation into their relationships with others throughout the course of a two –week institute and again months, even years, later on in follow-up workshops.

I believe the transferability of game-learned experiences can be maximized by being clear about the purpose of the simulation before using it and by thinking of it as one tool in an overall learning experience. Setting an appropriate context with the players in advance is important; as is making sure that the players understand the rules and roles. In some simulations, guided practice may be needed before starting the actual game.

From my perspective, the most critical elements of a simulation come after the game itself. Debriefing what has happened – what a player experienced, felt during the simulation and is feeling afterwards, what strategies were tried and what happened, what other strategies might have been applied, what else the player needed to know or be able to do, analogies to real life situations, how the players’ own values and experience influenced their actions -- are all important items for discussion.

WS: I think the most anyone can say about the effect games may have on
player behavior is that some people will be affected in some ways at some times by some games. Not much there upon which one could or should base an educational program (or public policy!). I think the key to using games to influence behavior and/or learning is to put specifically conceived and implemented games in an appropriate context (a classroom, for example) and use them to teach very specific, targeted things. In general, I've seen very little evidence, anecdotally, that game-playing has any more influence on behavior than any other entertainment medium or social situation.

Clark: How accurate does a simulation have to be to be a valid teaching tool?

WW: In most interesting fields, like weather modeling, predictive simulations are very difficult or impossible. However the property of weather being unpredictable can be a property of a good descriptive simulation.

Let me give you an example. Say you put the ball on the tip of a cone, and let it go. A perfect predictive simulator would tell you exactly which side of the cone the ball would fall for the exact condition set up. A descriptive simulator, like SimCity, would probably use a random variable to decide down which side the ball would fall. While that simulation would fail at being predictive, it would teach both the range of possibilities (i.e. the ball never falls up), and also from a planning perspective, it teaches that you can’t rely on predicting the exact outcome, and how to deal with the randomness.

I have seen a lot of people get misled. I see a lot of simulations that are very good descriptive (like SimCity), but a lot of people use them predictive (like a weather model).

Clark: How do you research topics for a new product? What is the role of subject matter experts? A person from what perspective (i.e. academic, practitioner, consultant) has the most useful information and perspective for making a simulation?

JB: Our ideas come from many places. With George Lucas as our creative director, they often come directly from him. He has a great interest in simulations and their potential as a learning tool. We also research commonly taught topics in schools and think together about which ones can be done better using computer technology. It is important to us to connect our work to those things being taught every day in classrooms.

Subject matter experts play a key role in making and creating simulations. We like to put people from lots of different perspectives around the same table when we’re working on a high concept. We’ll bring together academic experts, practitioners, and others with interesting backgrounds to help us identify the most important themes and ideas for us to convey through the simulation. For example, when we were creating Star Wars® The Gungan Frontier™, our ecology simulation, we worked with teachers, population biologists, ecology and science center staff, zoo educators, etc. You need a mix of those whose work focuses on the big ideas and concepts, those who are working directly in applying that knowledge in everyday work, and those who are translating that into educational experiences for others. None of these alone is sufficient. I also love to add to the mix that person who brings an unusual connection to the topic to the table.

WS: You research topics for a game project pretty much as you would any other kind of media artifact. You figure out what you're trying to
achieve -- what you hope players experience as they play -- and start digging up books, magazines, movies, web pages and anything else that can help you get where you're trying to go. I hate to sound mysterious or mystical but there's no one answer to this question. If you're making a game set in real world locations, go find information about those locations. If you're making a fantasy game, you might want to read every fantasy book you can get your hands on so you understand the genre's conventions well enough to represent them well (or undermine them, depending on your intent!). In making the kinds of games I do, "experts" really don't enter the picture much. (It'd be tough to find an Illuminatus who'd admit to being one and orcs and elves don't talk to humans much.) But if you were making a game about the LAPD or the Navy SEALS, you better talk to some real LA policemen or some SEALS... But, really, where games are concerned, experts and real-world information sources are only so valuable. Making a successful "real world" game involves knowing where to deviate from reality in the interest of fun more often than it does knowing reality inside out so you can cleave to it religiously.

WW: That is my favorite part of a game. I usually do a game that I am interested in. I find someone who is either controversial, or between two fields. I will use content experts as canaries in the mine. I did SimEarth based on Lovelock’s Gai.

Clark: Gaming is starting to reach the masses. What lessons can e-learning learn as it tries to do the same?

JB: Stay close to the learner (your customer) – understand his or her needs, interests, and context. Be very clear about you’re trying to do and don’t try to do everything. Use the computer in ways that take advantage of its unique capabilities. Get good at doing each thing well before spreading out to other things. Just because things are possible to do, doesn’t mean you should do them – especially when it comes to the high-end stuff. It’s important to create the highest quality possible product with tech specs that match the installed base. And, to develop products at a cost that keeps the price point appropriate for the budget realities of schools or other customers and allows developers to sustain themselves financially.

WS: As developers try to reach larger and larger audiences (to offset
larger and larger development budgets!), we have to focus on a few critical points, I think: Any time we think something is too simple, we have to make it simpler. We absolutely must streamline our interfaces and make them so intuitive users forget they're even USING an interface. We have to make sure users know exactly WHAT they're supposed to do at all times and challenge them to figure out HOW. We might even want to consider leaving the fantasy ghetto behind and giving people subject matter they're already interested in -- in other words, make games that have built-in appeal to a larger audience.

Clark: What are the elements that make a simulation immersive (i.e. to make someone who is playing the game buy into the illusion)? What happens that breaks people out of "immersiveness?"

JB: There are many factors that affect this. Perhaps the most important is what I’ve heard George Lucas call the “immaculate reality.” Attention to the detail and cohesiveness of the simulated environment is crucial. One discordant factor breaks the illusion. Likewise the choices and actions the player faces must fit within that universe and its internal “rules.” One of the important tasks a player faces in a simulation is making sense of the world he or she has entered and figuring out its internal rules – both its consistencies and its intentionally designed inconsistencies. It is important not to break into a simulated experience with peripheral information. For example, you wouldn’t want to interrupt an ecological simulation with something flashing questions about your experience at you. Save the questions for later.

WS: At the most fundamental level, making a simulation immersive involves removing as many obstacles as possible between player and belief in the reality of the depicted world. Obstacles take many forms -- shifting camera position during play (e.g., third-person conversations in a first-person game); forcing players to switch from real-time gameplay to separate interface screens; dialogue presented in text, rather than spoken, form; objects that don't behave and/or can't be used like their real-world analogues... I'm not saying you want to recreate the real world but, certainly, you want to strive for internal consistency, at least, so players aren't reminded they're "just playing a game" any more than necessary. What you want to do is (and I'm about to reveal a boatload of prejudices here!), create a game that's built on a set of consistently applied rules that players can then exploit however they want. Communicate those rules to players in subtle ways. Feedback the results of player choices so they can make intelligent decisions moving forward based on earlier experience. In other words, rather than crafting single-solution puzzles, create rules that describe how objects interact with one another (e.g., water puts out fire, or a wooden box dropped from sufficient height breaks into pieces and causes damage based on its mass to anything it hits) and turn players loose – you want to simulate a world rather than emulate specific experiences. Okay, I'm officially failing to get my point across so I'm going to stop.

WW: The more creative the player can be, the more they like the simulation. This might be giving them a lot of latitude. People like to explore the outer boundaries. There is nothing more satisfying than solving a problem in a unique way. Another derivative: being able to describe yourself to the game, and the game builds around you. It also helps if a player can build a mental model of what it going on in the simulation. This has more to do with the interface. All it takes is one weak link in the chain to blow this. Have them think they understand it enough to start testing their theories. At this point they are reverse engineering your program. You want to give them an entry path. People will say, “Oh wow, my mental model was way off.” Most of my games use an obvious metaphor and a non-obvious metaphor. They think it is a train set, but they come to realize it is more like gardening. Things sprout up and you have to weed.

Clark: How do you reward or penalize a player within the context of a simulation?

JB: I’m a strong believer that logical consequences are the best and only reward or penalty for a player in a simulation.

WS: I'm not sure penalizing players is ever appropriate in a game. Well, that's an overstatement. Obviously, there are inevitable penalties associated with failure but, when you can just load a saved game and try again, how severe do you want those penalties to be? Basically, game development is about presenting players with genuine challenges and then providing sufficient rewards to keep them feeling good about themselves and eager to tackle the NEXT challenge. Reward schedules are critical. But penalties? Punishment? Sounds like entirely the wrong tack to take. But maybe games are different from more educationally motivated sims in this regard...

WW: With a game, you need some kind of reward structure. I try to have several goal path; I try not to force them down anuyl. In SimCity, you can go for happiest people, or biggest city. There has to be paced reaward structure. Some rewards – RTS, forced multipliers, some things that a – you can build a unit that allows you to do something – decisions that force them to think long term vs. short term gain. Giving them strategic decisiojns. Giving the peoiple maximum creativeity. The games are really just problems.

There is never one way. One way kills creativity. New ways of solving problems drives people in wanting to share experiences.

The photo albums in The Sims are important. We have to create new ways for users to share. Otherwise people used to feel as if they wasted the time they spent playing a game. So it is important to have an artifact that they walked away from the game. They can show people. Getting people to engage other people with what they learned is critical. If you can get people to talk about, it creates a snowball effect. You have to create glue. The community becomes the effective tool for learning.

Clark: Any ideas at how to measure the effectiveness of a simulation? Please?

JB: All of us who work with these powerful tools realize how complex and difficult it is to measure their effectiveness. Subjective tools such as pre/post discussions, interviews, writing activities can capture some of this. Rubrics have been developed to look at actual participation during the course of a simulation, but I tend to oppose that approach. Test performance on measurable items may be used, but it is impossible to separate what combination of things contributed to the learning measured and to account for any learning not covered by the test items. If you figure this one out, let me know!

WS: I haven't a clue. The measure of a game's success is in sales or, possibly, in critical acclaim. If players tell me they played Deus Ex for 12 hours straight without eating or going to the bathroom, the team succeeded; if they stop playing after five minutes, we've failed. Similarly, great reviews tell us one thing, bad reviews tell us another. From a slightly different standpoint, if players describe the way they solved problems in a game and, in doing so, describe situations the team didn't preplan and didn't anticipate, well, that's a big win. But in terms of measuring the effect of a simulation or determining what was learned by a player, in any specific way? Beats me. You'd have to talk to an educator or a psychologist about that. And even then I'm not sure I'd believe the answer. But maybe I'm just a cynic!

WW: You first have to develop intent. The most interesting things to use simulations for are the hardest to measure. Teaching creative problem solving is very difficult to measure. A simulation is more like on-the-job experience. It is a broader element. You have experienced a larger landscape of possibilities. How you measure that, I wouldn’t want to hazard a guess. It almost feels to me that the forces that demand tight metrics would not co-exist with simulation users. They may be incompatible ecosystems.

No comments:

Post a Comment