At #GDC18 @danctheduck is talking about how so many naively-implemented social systems (like say, twitter dot com) are deeply dehumanizing, and that we could do so much better at creating spaces to foster friendship 
He’s presenting a framework, based on psych and sociology research, on how games could do better at helping players make friends.
First, proximity: people have to be able to bump into each other in unplanned, serendipitous ways, repeatedly. Density makes a huge difference to allow you to bump into the same people again and slowly move towards friendship
Fortunately for online game design, this is a logistics problem and we can solve those—but not in the naive way that leads to “let’s just have players play with their friends by default” which will lead to bad results for most people! Online games must design for new friendships.
Basic tools to allow this: persistent identity for recognition in repeat encounters, events to boost density, daily incentives and side-communities like boards to boost repeat encounters.
Anonymity, huge worlds, proliferation of game modes and player categories for “feature richness” work against friendship. So do classic matchmaking systems: need high concurrent population, separate players after the match, etc. Only work for megahits!
Better design: room-based play that allows people to join and leave games in progress. slither.io and other .io games as an example. “Everyone should be making one, if you’re not.”
Extended versions: active waiting rooms as in Guild Wars II, rooms that have single-player activities for you to do if there’s nobody else there (yet!), big events (boss fights) in public space.
Elastic density is possible with instances that pop in and out depending on how many are needed, this is in many ways just simple load balancing. Garbage collection of too many instances : make sure there’s a reason for an instance to end and disappear.
For repeat interactions: preserve cohorts, in this case groups of 40 players become a village of bears that must life together, for good or ill. (Has anyone told @tonitonirocca about Beartopia?)
List of tools to support serendipitous and repeated interaction to help the first level of friendship formation, proximity.
Second stage: Similarity. Like it or not, humans tend to look for “people like us” when trying to find friends in a group, assuming that the cost of forming a bond will be lower. Of course, in many case this can be incredibly destructive too.
The anti-pattern @danctheduck points out is the agar.io game Nationalism, all built around national identity, and therefore all the most poisonous, slur-strewn, gatekeeping baggage that comes along with that.
Well-known alternatives: put players in factions (Horde/Alliance), give them shared experiences to bond them. But his favorite pattern is creating new social norms that allow for positive fictional identities.
Beartopia is a nice place! A village of bears that are all immigrants. It’s not a place where people are mean, the bears help each other. So with that set up, players adopt new players into the positive social norm. Utopias over dystopias, heyyyy
“Games don’t have to be about the fall of humanity and people turning on each other as we devolve into animals. What if we made game worlds about humans being the best we can be?”
Third stage, Reciprocity. A gesture offered, and responded to. Even something as simple as eye contact made and responded to can start to build an important resource: Trust. It’s wildly important.
(This is also the operational definition of “trust” that I used in designing Consentacle, btw —Naomi)
Trust grows very slowly and can crash very quickly, in a betrayal. (Dan says it grows even more slowly in cold-weather climates where even a simple interaction costs a lot of heat!)
Trust building allows for higher stakes, but we can each only afford a small number of deep friendships where the trust interactions are very expensive to reciprocate.
Anti-patterns for trust are exploitative of humans’ wired inclination towards reciprocity, scams, power problems, capitalist schemes.
Naive tools for reciprocity are often just dumped willy-nilly into a system as features. Twitter, again, as an example. Give people lots of ways to interact! Then spend years shoveling the wreckage.
Rather than just adding Every Feature at once, what if players could graduate at their on pace up a series of levels with increasing trust interactions.
Dan says Discord is doing a lot of this sort of thing well, even at the very start, helping to ease newcomers arrival with a little humor, a social offering
Another example: in Steambirds, everyone gets XP from a nearby kill. All helping each other, without even having to decide to do it. Destiny 2’s guild rewards where all members benefit.
Part of what’s important about sharing across a whole guild is that these social systems should work for a wide variety of people, not just those at highest level of trust or heavy players.
Varied player roles help in trust too. Dan references the work on trust in games by @raphkoster & Aaron Cammarata, which is some very thorough and important research on this subject. I recommend! raphkoster.com/2018/03/16/the…
Finally, the highest levels of friendship allow for intimate disclosure of private information that risks rejection or problems. So Dan recommends making these features gradual and opt-in as well.
Voice chat is a very difficult interaction for many players to deal with at low levels of trust, with strangers! Because as many women have found, you get the worst and most inappropriate comments immediately.
The big takeaway from this talk is to think about social aspects of play as a graduated process that complements certain gameplay, not just a set of features that all players can or should use with each other right away!
