Katherine Isbister presented the results of a survey exploring the relationships between players and characters. She didn’t have a definitive answer to the eternal question of game character development: whether it’s better to create player-characters with strong, fully-developed personalities, or to provide the player with a blank slate onto which he or she can project his or her own emotions or values. Basically, some people like it one way, and some like it the other. More interesting were findings suggesting that in-game interactions are as important as story or dialogue in creating connections between players and characters; Aeris’s death in FFVII wasn’t upsetting simply because she was a likeable heroine, it was upsetting because the player had invested a lot of time in levelling her up and she was a powerful party member.

“Immersion” is one of those terms that’s mostly had all the meaning buzzwordized out of it, but Clint Hocking went a long way towards bringing it back into the realm of usefulness. He distinguishes between sensual immersion — the overwhelming feeling of place where the gap between player, avatar, and world closes to near zero — and formal immersion — the recursive, repeating engagements that speak to the pattern-matching machinery of the right brain. He also talked about ways to reconcile the two perspectives, and the difficulty of developing each type of immersion without breaking the other.

Nicole Lazzaro’s talk compared Halo to Facebook in terms of how well they build or reinforce social bonds by creating feedback loops between people. It’s an apples-to-oranges comparison, of course, but there were a lot of interesting observations about the broad but fleeting interactions on Facebook vs. The highly structured, repeating experience of Halo multiplayer. Both systems can result in a cascade of “snowballing meaning,” where in-jokes and shared experiences lead to a more tightly-knit social group.

One thing i can never get enough of at work is data on how the game plays, how my AI is performing, etc. Ramon Romero gave a presentation on some of the instrumentation that Microsoft has for collecting and presenting gameplay metrics, and I’m all psyched to start pestering my engineers for some of what they’ve got. Gathering data isn’t that hard — log the critical events and pour them into a database — but sifting through that data to find meaningful exceptions, and then presenting those exceptions in an understandable way is an art unto itself. And some of the death maps Romero showed from playtests of Halo 3 were pretty artistic, if Tuftean data visualization is up your alley. The big revelation was that qualitative interview data is just as valuable as hard behavioral measurements — noting which spots on a level caused players to express frustration is at least as interesting as tracking how long it takes to find a waypoint.

Also: Nintendo had Super Smash Bros Brawl playable on the expo floor. I am even more excited about the game now than I was a week ago, which is saying something.