In what was accurately self-introduced as "a fairly heavy design talk" at Game Developers Conference, Ubisoft Montreal creative director Clint Hocking said he would attempt to "bridge the gap" between intentionality in game design, and what he calls improvisation -- in the process providing a fascinating look into how a team can turn an apparent game design failure into a new success. Kicking off, he referred back to previous talks, defining "intentionality" as "the ability of the player to devise his own meaningful goals," as phrased by fellow designer Doug Church. He showed a video from his game Splinter Cell: Chaos Theory in which a player entices a guard to kick down a door that the player his rigged with a mine, sending the guard plunging down an elevator shaft -- while the player watches via cameras he had laid. This required knowledge of at least nine game mechanics, Hocking pointed out, with understanding of some 36 interconnected relationships between those mechanics. He followed up with a further example from Deus Ex: Invisible War. Such actions require a roughly equal balance between composition of an intended result, and execution of the actions required to achieve it. "This [equal] balance between composition and execution is a very game-like system," he said. By contrast, a heavier balance of composition is more puzzle-like, whereas something like driving in a video game has a heavier balance of execution -- it requires little planning but a good deal of skill in real-time. "Games are not only a medium for the creative expression of their creators, they are also a medium for the creative expression of their players," Hocking pointed out, speaking on the difference between interactivity in general, and intentionality. "Intentionality is a subset of strategy that is concerned with expression beyond the limited constraints of winning the game." He explained that fire propagation didn't find its way into Far Cry 2 simply because "fire is cool," but rather as part of an attempt to build on the avenues for intentionality as explored in Chaos Theory. However, it didn't entirely succeed -- "this is how we failed," he admitted. Hocking described the Ubisoft Montreal team's expectations for how players should play Far Cry 2, relating an elaborate hypothetical scenario involving plans with deep levels of intentionality along the lines of the Splinter Cell example above. But, as he noted, "No design survives first contact with the production reality intact," and a number of those intended systems were either not able to be implemented properly -- or "poorly conceived" in the first place. He then described a more realistic situation -- a much more straightforward scenario with the player simply accepting a mission, heading straight to the objective, run into an unfortunate random scenario, becoming engaged with the enemy earlier than intended, and ending up in a close-quarters battle far less intentional than the ideal. "As engaging as the outputs were, we were very far from what we set out to build in the beginning," Hocking said. "The more we failed, the more we whittled away on the composition phase until it was almost gone," referring back to his earlier mention of ratios. But as that happened, "the play experience got better and better and better -- what the fuck?" As it turned out, that structure was much closer to a more traditional shooter experience, "it just wasn't the game we intended to build. ...But at the same time, we were embracing what the game itself wanted to be." Interestingly, this led to a game that was actually very difficult to demonstrate, because the level of intentionality ended up being quite low -- a few minutes before a PAX demo, he realized "I was going to be making it up as I went. I wasn't going to be demoing Far Cry 2, I was going to be improvising it," noting that it reminded him of his earlier days being in a band and performing a live show. Still, paradoxically, "improvisational play can be highly intentional," he said. What ended up occurring with Far Cry 2 was not that the execution phase lengthened to compensate for the composition phase shortening -- rather, both phases ended up being quite short, but alternating very quickly. "Randomness and unpredictability for where improvisation is born." System's like Far Cry 2's malaria and weapon jamming, which introduce randomness, ended up having much more influence over the final experience of the game than was expected with the initial design. What happened was then that they were the triggers that kick the player out of the execution phase back into the composition phase, leading to the rapid back-and-forth of those two phases. In contrast, the consequences for getting kicked out of the execution phase in Chaos Theory has a huge impact -- the game is so reliant on the player executing his careful plan, and the game is so slow-paced, that it makes more sense simply to reload a saved game. But in Far Cry 2, that disruption ends up being part of the game, and there is such a level of chaos to begin with that players did not end up feeling the need to reload every time something went wrong; rather, they would adapt to the new factors. Another game with a similar dynamic, Hocking said, was BioShock with its Big Daddy fights -- particularly early on in the game, when fighting a Big Daddy, the player generally is kicked out of his composition phase a few times before he has figured out an efficient way to take the enemy down. But a game can't "overwhelm" the players with improvisational demands -- BioShock has its vita-chambers, and Far Cry 2 its revitalizing buddy system, for example. Unlike nearly all other games, there is no actual repetition upon that failure, however -- you simply move to the nearest "safe" point and continue. Hocking painted a contrast between the terminology of "beating" a game and "completing" a game -- he thinks of games as something to be "completed." For example, in Splinter Cell, you are never explicitly taught the specific combination of skills allowing you to pull off the earlier example. Such actions are separate to what is actually necessary to "beat" the game -- it is an improvisational action that will only occur if the player invests himself into it for a purpose other than simply beating it. It is unnecessary for a game's core goals to be difficult simply for the sake of challenge, Hocking argued: "We as designers need to reject the notion that games ought to be punishing or abusing." Showing a striking photo of a Muhammad Ali boxing match, he added, "Look how fucking beautiful play can be." "When we strive to understand the thing such as a game...we eventually arrive at mastery, and when we master a thing, we destroy it," he said. "It can't happen any other way in the end. Mastery is not a prerequisite to improvisational play. The only prerequisite is confidence, and the only prerequisite in making the game is that we do not" discourage the player from improvisation by "humiliating" the player.." Concluded Hocking, "Let's invite them in, and let them play."