Sheldon Pacotti on Cell: emergence's design, cellular automata, player education

Cell HD: emergence, the strategic shooter by New Life Interactive that immerses players in a microscopic world of cellular automata, impressed me with its unique ideas and presentation. Though perhaps not the most accessible game out there--partly because of its adherence to medical terminology and a "hints vs. hand-holding" approach--Cell provides a very rewarding journey for the playful and intellectually curious action gamer. I contacted New Life Interactive's founder, Sheldon Pacotti, to pick his brain about a number of Cell-related topics, ranging from the creation of the game's proprietary Automata Engine to the construction of the game's narrative. We also touched on subjects like the game's mixed post-release reception, and discussed the emergent nature of "cellular automata" and its differences from traditional artificial intelligence. focalbox Pacotti, who folks might already recognize as the principal writer for the first two Deus Ex titles at Ion Storm, has a wealth of experience making games. He worked as a systems designer, content designer, and lead designer on a number of projects after leaving Ion Storm, and also teaches game writing at the University of Texas. In summary, he's a really smart and talented guy. "My long-range target has always been an RTS made of 'cells' in which players build their own defenses and weapons, pretty much inventing the gameplay themselves," he said about Cell HD: emergence's inspiration. "I wanted to leverage cellular automata's potential for user-generated content." I asked him if in the case of Cell: emergence, whether the Automata engine preceded the gameplay concept, or if it was constructed to support the game's idea. "I started with the engine, not really knowing if the ultimate game would be a slow turn-based affair, or more dynamic," he said. "Then Asperger's Syndrome took over and I tooled and retooled the engine until I had close to a million cells running at arcade speeds." After two years of development, in which Pacotti was also managing a day job, family life, and a move to a new house, it was time to test things out. "It was time to stand up a 'quick' action arcade game to fine-tune the engine in preparation for something more complex. That also took close to two years," Pacotti said. "So I'm now thinking very carefully about what my next 'quick' application of the engine will be." Being impressed with the Automata Engine, I asked if he had any plans to use it for any future projects. "I'd love to make several more games with the technology and release an SDK at some point," he said. "Besides cellular automata, the engine provides a true 3D canvas, with voxel-based sprites being composited on top of each other with ink effects. (That's how the dynamic light works.) I think the blend of voxel-based sprites, all of which can also contain dynamic automata behavior, and the global cellular automata simulation, provides a vast space for experimentation." Opening up the engine to allow for user-generated content is a possibility Pacotti seems excited about. "The first thing I'd like to do is find a way to open the engine up to players somehow, because I'm still of the belief that user-generated content will lead to gameplay breakthroughs more quickly than me coding bleary eyed into the early hours of the morning," he said." Cell: emergence is a shooter at its core, and due to the dynamic nature of the engine, players need to be conscious of taking the best approach based on the circumstances of the moment, rather than simply plugging away at a set strategy until they succeed. I asked him to describe the types of decisions the AI makes on-the-fly. Calling it "artificial intelligence" is a bit of a misnomer, Pacotti said. "There has been a fair amount of talk about the AI, but I don't think of it as 'AI,' exactly," he said. "All of the behaviors of the entities in Cell proceed from random numbers--tens of thousands of them every frame, actually. For me, this is the magic of 'cellular automata,' and why I wanted to make a game based on it," he said. Pacotti went on to explain the difference between traditional AI, and the emergent way in which his game makes "decisions" based on a number of core rules and behaviors. "Very simple cell-cell rules can generate fluid, complex interactions, striking visuals, and even the sense, in many players, that there is some intelligence behind them. Though I've been told that the field of computer science formally defines cellular automata a 'AI,' I think of it rather as a deep simulation from which higher-order behaviors simply 'emerge.'"

Can you save the world girl?

Segueing from the Automata Engine, Pacotti discussed the game's narrative and its implementation. Both came about after the engine and prototype gameplay were up and running, but the inspiration came quickly. "I was seeing stories in Cell from the very beginning, when all the engine did was render battles of red and yellow boxes. But I also felt like I was gazing into a complete universe, especially once the simulation approached its current size," he said. I could see intricate structures and wreckage in the hazy distance and couldn’t help wondering what world they were part of. "I wanted the game to provide this sense of transport for players," he said. "I wanted players to believe that they were in a real place, with its own epic struggle between good and evil." From of these notions, and in an effort to provide reference for the mulit-colored voxel-based action in the game, Cell: emergence's cut-scenes were borne. "I wanted them to be lean, arcade style cut scenes--more suggestive than explicit--and I succeeded in places," Pacotti said. "My favorite scene is the three-panel interlude where the sick girl stirs momentarily and then falls back to sleep. No words; just a reminder that you are inside a body and that a life hangs in the balance." In a self-critical admission, Pacotti told me that if any of the scenes "fail," he feels it's because he included a bit too much narrative in certain places. "I couldn't resist creating a backstory involving terrorism and treason, and a three-act plot that shows the little girl, Eva, grow into a toughened superhuman adult (of which the 'emergence' title is only the first act)," he said. Even considering cases in which some critics complained that Cell's narrative was either being "too much" or "too little," Pacotti seems to have achieved what he set out to do with the game's story. "I'm gratified that even the reviewers who are most dismissive of the story take it as a fact that they just spent eight hours fighting disease inside the human body, which was the goal of having a story in the first place," he said. As in many good shooters, the action in Cell HD: emergence is fast and can get quite chaotic, with new threats sprouting up with regular frequency. However, the abstract (and beautiful) voxel-based visuals and infusion of real-world medical terminology add another layer of complexity that might be daunting to some players. That said, those attributes also really speak very well to Pacotti's inspiration and intent. "The joy of cellular automata, for me, is observing behaviors emerge from the simple rules," he said. "I wanted a game where players were looking deeply into the sim, studying it, the way I had my family study--with a microscope--a drop of the scummy red water that collects on top of the cover of our boat. (Really cool stuff living in there...) So, I wanted gameplay that involved 'solving' the mysteries of Nature and then conquering them." Pacotti extrapolated: "I wanted players to feel like they were living the life of a nanite who evolves into an independent being. I wanted them to figure out the simulation and devise their own tactics, in tandem with the scientists’ discovery of what their nanotech is becoming." Pacotti said that he attempted to help ease players into the experience, but conceded that if he had it to do over again, he'd make the tutorials a bit more rigid to help mitigate player confusion. "Help text pains me as a writer, because it breaks the fiction, so I'd rather say 'antibody' than 'translucent white voxel.' "Even though I added more and more hints and help text based on feedback before the launch, if I were to do it over again I would lock things down almost completely so that players know exactly what to do and can't proceed until they’ve mastered a concept," Pacotti said. Simultaneously educating players about new gameplay, terminology, and goals is a lot easier said than done, it seems. "The design ideal I was shooting for was the Valve technique of putting the player in a constrained situation where he can’t proceed without mastering an ability – part tutorial, part game challenge," he said. "The early missions of Cell are each skewed toward a single activity like that: shoot the Growth, defend the Membrane, spread the Antibodies, draw a pathway. I think where this approach breaks down is that it's possible, though not easy, to complete these missions without fully mastering a given technique." About six missions into the game, it's assumed that players know how to use their weapons and tools, but if a player hasn't effectively learned one of these core lessons, Pacotti admits that "there's a lot of on-the-job training which has to happen." He also sees some issues with how the game's HUD is introduced, which has led him to create a post-release tutorial video. He's testing the response to it, and says he'd like to incorporate some of its content into a future version of the game. In retrospect, he explained that had the HUD been explained more effectively, it would have done a lot to clarify the game's goals and mitigate player confusion. "The winning conditions of a level (high Membrane, low Growth; etc.) are represented by yellow circles on gauges at the corners of the screen, and losing conditions are in red. A visual effect calls attention to gauges that are either in trouble or nearing a winning condition. Unfortunately, tutorial subtitles explaining the meaning of red and yellow have not proven to be enough to teach players how the gauges work. Players are usually too wrapped up in a level to stop and watch how the HUD is reacting, which means that the most exact representation of how to win a level, which doesn’t depend at all on fictional goal descriptions, has remained a mystery to many players," he said. BOOM video 11970 Cell HD: emergence is a rewarding experience, and that's due in no small part to the fact that it doesn't hold your hand much. Trial-and-error is often involved, but makes finding a successful approach to a given level feel more meaningful. That being the case, reception to the game has been quite mixed. Pacotti recognizes this. "Given that the goal of this game was to put out a dead simple arcade game, mostly about shooting stuff, I have been surprised that it has come to be regarded as complex, and even foreboding. There are plenty of positive reviews and some very enthusiastic fans, but these are all people who devoted significant time to learning the game's mechanics and to beating some pretty tough levels." I asked him if he had any advice on how folks should best approach the game, or what he'd say to players still on the fence about picking it up. "To prospective players, I suggest tiptoeing into the game, first reading about it and watching the videos, and then, if the game seems to make sense and still looks fun, give it a try, expecting to pay attention to the tutorial text and to learn through failure," Pacotti said. "Think of it as Act 3 of DOOM, where you show up with nothing but your fist and have to make several suicidal dashes into the first level before you can make a plan of attack... and then you'll be in the right frame of mind." Cell HD: emergence retails for $8.99 is available for purchase from digital retailers GameFly, GameStop, Desura, GamersGate, and Greenman Gaming. A slightly less robust non-HD version of the game is also available via Xbox Live Indie Games for $5.