Saturday, December 8, 2012

Practice 2012: Dan Cook on Generation of Value

Dan Cook, founder of Spry Fox and creator Triple Town and Panda Poet, gave a talk about efficiently generating value for players with the least amount of development time.

Presentation
  • Dan is concerned with how games can efficiently add value to the world. The two key terms here are "efficiently" and "value."
  • Looking at the cost curve of game development, you can see that games used to take two developers and $40,000 to make a game, but now it takes about 100 developers and $40 to $80 million. Expensive projects have a tendency to be conservative.
  • Dan lists technology in order from least expensive to most expensive -- Flash, paid mobile, Steam, free-to-play mobile, console digital, and console retail.
  • Dan uses a "development years : quality hours" ratio as a metric for determining value of his games.
  • He asks players to give his game a fun score on a scale from 1 to 5. He considers 4 and above as positive, and under 4 to be negative.
  • Designers have huge responsibility to create quality play. Dan was determined to keep the development years fixed for his games, but improve quality hours.
  • With Bunny, Dan tried to increase value through media. Bunny is a game that's the harvesting portion of a real-time strategy game. The problem with media is that players burn through the evocative stimuli like crazy. Developers end up on a content treadmill, continually making more content so that players will stay engaged with the game. As developers make content, players consume content, and it ultimately results in a 1 dev year : 2 quality hours ratio.
  • Bunny was just mind candy for players and it ended up consuming a total of 1227 years.
  • With Steambirds, Dan tried to increase value through capital goods. People like "stuff" so create lots of stuff for people to unlock and collect. There is economic value in these goods, as they help players earn more currency so that they can get even more goods. In Steambirds, players earn currency to collect planes which help them earn more currency. The problem with capital goods is that the game ends up with diminishing returns. Each additional plane added less value and the play space of fun was muddied by a lot of goods that were not fun. This results in a 1 dev year : 10 quality hours ratio.
  • Dan then decided to "screw it" and use math to reduce the number of capital goods that you needed to add. In Triple Town, players can obtain amazing high level structures, but the road to get there is an exponential curve. As a result, level 30 items take longer than the remainder of the sun's life to earn. While this results in a 1 dev year : 20 quality hours for casual players, it actually negatively creates a 1 dev year : -5 quality hours for players who like to grind.
  • This is similar to the story of the origin of Chess. The king offered to buy Chess from the creator. The creator agreed if the king would pay 1 grain of rice for the first square of the board, 2 for the second, 4 for the third, 8, 16, 32, and so on exponentially. The king, not well-versed in math, agreed to the deal. Afterwards, he realized that it resulted in millions of grains of rice, more than the kingdom even had. Enraged, the king executed the creator.
  • With Triple Town, Dan wanted to add value through skills, a human capital. Triple Town is a single-player tactics and organizational game. There is a whale player named Stephan who has a 62 million high score. The high skill barrier makes about 90% of players drop off, but for the remaining 10% like Stephan, there's a 1 dev year : 100 quality hours value.
  • However, some players write "improving Triple Town skills really doesn't count." They end up accomplishing nothing since the skills don't translate to the real world.
  • Dan tried to add more value through status. He implemented a leaderboard to Triple Town. This fundamentally deeply ties to the real world since people want to show off their skills to their friends. This creates a 1 dev year : 150 quality hours ratio for top players, but negatively creates a 1 dev year : -50 quality hours for people with low scores. It intimidates casual players and causes immense psychological stress. Leaderboards help some people, but are killing the rest. Studies have proven that normally smart people act dumber in the presence of "high status" people.
  • Finally, Dan stumbled on a value engine when he created value through trade. Capitalism creates value. Trading increases total value of all items since excess items can be traded for wanted items. This is also an transformative act. If a player with hot dogs traded a player with buns, they can both form a more complete item. This is a non-zero-sum act and a mutually beneficial act.
  • In Realm of a Mad God, the developers simply added the feature to drop items on the ground. Players themselves created a trading subculture. There was an explosion of emergent cultures from trading norms and languages to unofficial currencies and trading houses. There were also advocacy groups that would lobby for or against changes that the developers were about to make. There were trust networks that were formed to determine which players were fair to trade with.
  • The emergence of these subcultures blew Dan's mind. What if we're creating these games wrong? We shouldn't think of them as these packages that we build and release, but as engines and its players as the energy cells that powers the emergence of systems. Realm of a Mad God grew out of a hobby and became a lifestyle, resulting in a 1 dev year : 500 quality hours value.
  • Minecraft is a horrible piece of media. It has ugly graphics and a bad plot if any at all, but it's an amazing value generator. The ability to create cultural artifacts, build things out of your own imagination, play with friends, and learn while playing with parents make it a value engine.
  • Community generated narrative, or fan fiction, is also a great value engine. It turns out that Bunny, while not a very successful game, had characters that resonated with its players. The players culminated in forums and created a bunch of fan fiction.
  • We should think about more engines like creative tribes, sports, governments, and religions. Instead of thinking about game development as Creator --> Content --> Consumer, we should think of them as Creator --> Engine --> Community --> Culture.
  • Contents of an engine include building positive relationships, mutual benefit, persistence, specialization, working better with a group, shareable artifacts, and most importantly, non-zero-sum relationships.
  • Engines are still games and still need scaffolding. Tweaking culture feels like being a "dictator gardener." You are in control of an autonomous amorphous entity.
  • Millions will build communities, culture, and art around our games. We are creating legal systems for these worlds.
  • Small studios should not waste time making media or content, but just rule sets. This is an efficient use of both the designers' and the players' time.
Question and Answer
  • Mods have existed for ages and developers releasing tools like Hammer and the Team Fortress 2 hat system will further instigate them.
  • Realm of the Mad God outsourced most of the art to players since pixel art is easy to make. It resulted in a huge net positive since developers get free art and players feel ownership about their work.
  • Dan is not concerned about creating short one-off experiences like Dear Esther. However, he thinks the conversation about Dear Esther, in many ways, has more value than the artifact itself.
  • Positive psychology has gotten away from the study of "happiness." Hazing is a horrible experience for anyone, but it also has an extremely positive effect on group identity.
  • "Fun" is not the only positive value. There's also value in the "potential activity" even if you don't end up doing it. For example, someone can watch a commercial of an outdoorsy person with a hummer and end up buying the hummer without ever going to the woods. But the ability to go to the woods if he decides to is valuable.
  • The "big bang revenue model" is to spend a lot of time developing a game, release it, get two weeks of revenue, then see a huge drop off. Developers should look for a constant stream of revenue.

No comments: