Thursday of last week, Ben Medler, from the Office of the Chief Creative Officer at Electronic Arts, came to talk to us about game analytics. If there is an expert in the field, it’s Ben. Not only was it the subject of his thesis, it really is a true passion for him. He was supposed to talk for forty minutes, he talked for over an hour and a half, and he could have gone on for much longer. Ben’s understanding of analytics and metrics differs from most people in that he thinks that, even though most people tend to think of data as dry and boring, they are wrong. Data is fun, we just don’t know how to keep ourselves from hiding the fun, thus the name of his thesis and his talk at GDC: Play With Data: The Many Faces of Online Game Data.
Ben originally wanted to give us a preview of his GDC talk, but after having read all of our questions, he preferred to present us with a few preliminaries first. His GDC talk is aimed at people who are already familiar with analytics. Based on our questions, he thought it would be a good idea to give us a bit of context about what he meant by “game data.”
He started by defining analytics as the activity of “finding trends and outliers in data sets.” Our questions showed that by “analytics” we were thinking of telemetry, the way some games send information back to servers about some of the players’ actions. But analytics are much, much broader than that. The data used in analytics can come from more or less anywhere. Ben’s objective in the matter is to try and expose to fun in analytics. Already, data is sometimes made “sexy”; it’s presented is visually beautiful ways. But Ben thinks we can go much further: he thinks data can be playful, and that creativity and imagination should involved in it.
Ben gave many examples of player-made systems where data is gathered to help current players play better, like “World of Starcraft”, Terraria map viewers or a page that tracks the length of the girl in Noby Noby Boy. He covered the issues involved in that practice, such as technical issues, cultural issues and legal issues. He also showed some analytics reports systems made by companies, for example, Battlefield’s battle logs. For the Noby Noby Boy example, Ben showed how the shape of the curve was indicative of how the game designer changed the rules of the game among the course of the game’s life.
In response to our question, “does any company present data well?”, he, again gave many detailed examples, including CCP’s API for EVE Online, how Valve shares data for DOTA 2, and examples from Bungie, Ubi Soft and Electronic Arts.
He then moved on to our question: what motivates players to be interested in data. He talked about intrinsic and extrinsic motivation. But he prefers Steven Reiss’s theory of the 16 basic motivations, as more accurate and helpful in his work.
Ben then went on to show examples of very unconventional gameplay data, for example, Evan Roth’s ink finger swipes of all 300 levels of Angry Birds, or photos of he and his friends playing Johann Sebastian Joust, where the trail of lights gives indications of how the actual gameplay went.
Ben then got more technical. He presented us with eight categories of play analytics, such as content databases, like user-made wikis, user-generated content, like in Little Big Planet‘s user-created levels, maps, leaderboards, tournament results, and actual raw statistics.
He talked about what he was working on right now: he’s working on a team that’s trying to make a system where games could be played fully be streaming and by using a mobile device as the controller. For example, all of the processing could be done by a remote computer, the local display device would constantly stream down the video output, and the input could come from a touchscreen cell phone or tablet. He’s trying to adapt existing games to that system: the challenges are not the same for all games. Some games, like Peggle, adapt fairly easily to that system. Others, like Dead Space 2 or Mirror’s Edge are much harder, because they require a controller with many buttons. Some games prove to be naturally well-suited for that system, like Tiger Wood’s PGA Tour.
He answered our questions about the relationship between analytics and game design: analytics, he said, are not useful to make games, but they are an extremely important tool for tweaking them. The fact of the matter is, he says, people lie (especially on forums). Analytics allow us to link what they say to what they do. It allows designers to target problems like the way people usually stick to a single mindset and playstyle and diminish the likelihood of a game having a dominant strategy. It helps designers compensate for the was social perception can overpower data.
Data as testimony
Ben finished with a story about his time working as an intern on Star Wars: The Old Republic. The game keeps a lot of data about what players do, but doesn’t present it to them. And this is key to Ben’s major project: games create a narrative and the data we can collect from them is a meaningful narrative about who we are through what we did. MMOs, with the way they constantly keep us in touch with other real-life people, don’t only create narratives and testimonies about ourselves, but about our relationships with others.
As we were very much overtime, by that point, there were only very few questions. Carl asked what the data in Ben’s Star Wars account was to him, if it was a scrapbooks of his journey. Star asked if there were often were contradictions between what people remember doing it games and what analytics show us. Ben said there was all the time. And Star also asked Ben if he was afraid of Google, like he (Star) was. Ben said he actually was afraid of Google as well, but games do not have the same ethical concerns as Google’s data collection because games are outside of real lif.