Game Design Patterns for Building Friendships

Hi. Do I have any friends here at all? One or two. As long as there's just one or two, that's all you'll keep my heart alive throughout this very very very boring talk. Um I I tried like normally I like I think like maybe I'll put some jokes in. Like wouldn't it be cool to have jokes? And this time there's no drugs jokes, there's just data and information. I'm just going to try to like I'm pour whatever's in this head into your heads and hopefully things don't dribble out too much. Um they'll probably dribble out of my head. So um couple things turn off your devices that make noise fart machines just turn those off right now. Also there's a evaluations um there's data that is collected on you and shared with the global internet. So

no, fill out you fill out your fill out your questionnaires and stuff. Um let let me know how I could be a better human being. Um All right. So I'm Daniel Cook. I've been a game designer for a bunch of years now. They keep going up. I'm trying not to update that number anymore. It's kind of like it's the it's the 39 of 20 is the 39 of of game developer years. Um I've been writing on lostgarden.com whole bunch of a game design the writings. I like to think about tools that we can use as game designers to stop making so many how to put it stupid mistakes. Um like we tend to make the same mistakes over and over again and wouldn't it be cool to actually have like mental tools that we can approach a problem and solve it in a smart way versus like flailing around and then going and like blowing up a few million

dollars here and there. Um I co-founded a company called Spry Fox a small game studio. We've been going for about eight years now. We've made games like Triple Town. Initially I got interested like I kind of did the thing that game designers do which is like you know, it's all about the game mechanics and we want to do small elegant tight games that have deep you know evergreen gameplay. Um and then so I did that sort of thing for a while. Um but then I got into MMOs and I realized that people people are actually kind of important. Like people are are cool. People are interesting and you can build games with people that make people's lives better. And so this was a cooperative shoot them up online game and and and people couldn't really grief each other. So there there was there was a there was something nice about a game where it's not about betrayal or negative emotions. It's about trying to

build community and that kind of got that that idea seated in my head. Um we're working on another one. So we keep we've we've been making more and more online games. We made a um a sort of a multiplayer Animal Crossing called Bear Topia. We make a lot of cute games as well. Like our games are played by men women all ages. We're really interested in like the whole spectrum of humanity out there and how can you make games for them? And one of our one of our sort of slogans is making happiness. Um Like as opposed to making money which is also nice, but that's you know, like this is sort of a this can games contribute to humanity? Um So I don't build games because I love games. Like that's not that's not what I do. I mean games games are cool. Don't get me wrong. But games

games are a tool. As a designer I think of games as a tool that can help us bring happiness to the world. A tool that can make people's lives better. Um so I'm always looking for things that can improve happiness. So it's not enough to just design a game. Oh, we made a game. That's great. It's like it's like driving in a parking lot the first time you get your driver's license. It's a super cool thing, but it's even better when you're like going on a date night and you're driving to the movie with your date and you're excited and the the car empowers you to have this wonderful life experience. That's what games should be. So looking at like okay, well what improves people's lives did a bunch of research looked around looked at things for years and this is like one of the big findings that keeps coming up again and again and again in social psychology.

A small number of deep friendships leads to greater happiness in life. People have fulfillment in their lives. They start thriving when they have lots of small number of very deep friendships. Um and then there's a flip side to that which is large numbers of shallow status based friendships are actually really bad especially for teenagers. If you are a teenage girl and you are basically your entire social network is people sharing images of each other how cute you look you may be on the road to depression. It's not a healthy thing to have that. Um And so um given that sort of finding I asked myself well well what what can online games do? Like what can games do to increase this sort of like meaningful relationships? Um And we all have a fantasy. I think we all sat around the table and played games with with our family or friends

and we have this like sort of like oh you know, we're we're hanging out in this face-to-face interactions. Um and and like if someone does something we all laugh cuz it's a joke. It wasn't meant to be mean and everything's nice and wonderful. Um but the reality is is instead we're mostly acting with interacting with strangers. Um inter mediated by this very very thin pipe of the internet where all sorts of nuances lost all sorts of miscommunication happens and then we put on top of this we put on top of that systems and algorithms that are immensely dehumanizing that kind of strip our ability to act as authentic human beings away from the interaction from the reciprocation process. And and then and bad things happen. Um so you know, you you get match based systems which are this very common thing where we have relationships, but they're totally throw away. We interact with a

person and we're like oh we're never going to see them again. Of course I'm going to swear at them. Cuz it's like why not? Um friendship systems that you know, again it turns people into a number in some sort of viral marketing campaign. Again dehumanizing and we just do these things because they they make numbers go up. They make metrics go up and that seems like a nice idea and there's KPIs that are really important to somebody and we made them go up. So our salary might go up. Um But we do also things that are just naive. We're like chat. Chat see people seem to like chat. So let's just give people chat. Um and so you have these things like like Twitter where you're basically taking people and firing at it them at each other at super high velocity and some of those people turn into friends and you're like see Twitter is justified. And then there's all these sort of mangled heaps of bodies that are like lying in the gutter someplace because they didn't form friendship and they were damaged by this sort of naive

implementation of social systems. So can we do better? Um I think we know games build friendships. We know this. Like there are many people here who have probably made friendships in games. Who here has made a friendship in a game that's lasted? Look at that. Look at that. Games build friendships. So we know that is a fact. So how can we do that better? Um So I went looking again cuz I I I like sort of tying together different groups different different ideas from different fields and I'm like I wonder if anyone's thought about friendship formation before. And I found in the 1950s there were two folks. This is Schachter which is an amazing name. I wish my name was Schachter. Um and Festinger which is

also an amazing name. If I could go by Schachter Festinger and like have it in a big like name tag I would do that. That would be pretty cool. So in the 1950s if you ever heard of cognitive dissonance that is Schachter. So these are big names that like have done fundamental research in cognitive psychology social psychology. Um And like I was doing a search on one of their terms and it's still like there's like it since 2010 there's been like 4,000 research papers on this topic. So they came up with a theory that is actually still super validated. There's immense quantities of research. It's been studied you name it any field any sort of organization their ideas have been applied and validated. Maybe we should do that for games. Um so this is the proven model. Um, this is what it is. It's really pretty simple, right? It's like, how do

friendships form? And it says, well, you need these you need these four things for friendships to form. And if your game doesn't isn't designed to cause these things to happen, friendships are going to form at a dramatically lower rate. I mean, you may have some people that escape into chat channels and still form friendships through there, but in general, if you don't have these things in your game, your game's not going to be good at forming friendships. Um, so the rest of this talk is the super boring part. I've used up all my jokes. Um, we're going to go through each one of these and we're going to say like, what is this thing? And then we're going to say, here's some examples in games, some practical boring game examples, mechanics, systems, that sort of thing, that are doing it and doing it well, that we can just say, why don't we do that and do more of it?

Okay, proximity. This is your um, cool word of the day, propinquity. You can you can practice that in the bathroom mirror, propinquity, um, which is is essentially social distance. How far apart are people from one another? And propinquity is all about it is there's a couple reasons why you want the the rules of this proximity thing. You want repeat serendipitous interactions. So let's assume there's a bunch of people in a room and they are walking along and they're ignoring each other, but then they happen to face one another and see one another and they bump into each other and they say hello. That's a serendipitous interaction. They weren't planning on it. It was just sort of the shape of the room, the number of the people in the room, and they say saw each other and said hi. So you need serendipity in your in your system. Um, and you need to have repeat interactions. If those people bump into each other and they never see each other

again, then they're never going to become friends. It's just not going to happen. It's logistically impossible for them to become friends. Um, but if they see each other again and they actually have multiple repeat interactions, then slowly over time maybe they'll become friends. And you can sort of combine those ideas together and say, well, well, there's there's this thing called density, where the number of people in a room probably matters or the number of people in an environment probably matters. Like if you're living in the desert and the only things around you are cactuses, cacti, cacti. Um, then you will probably not make any friends. I think of what was it? Was it Joe from Snoopy? He lived in a desert and he was always alone and sad. Like, you can make games like that if you really want, but nobody will become a friend. Um, you can also do the opposite. You can make it too noisy. You can have people jam-packed in. We actually had

this problem in um, in Realm of the Mad God, where we would have the hub area and there were so many people in the hub area that it just became like people shut down. They're like, it's dehumanizing. I want to like ignore everybody and not pay attention. Um, so there's a certain density you need. Now, what I love about this, this is the sort of thing that just gets me incredibly excited. Um, we were talking about social psychology and now we're talking about logistics. And as game designers, we are excellent at logistics. That's numbers and people and pacing and timing and placement and all those sort of things. So if friendship in that there for this factor is just a logistics problem, we have all sorts of tools to solve that. So, this is an example of logistics thinking. Um, so a lot of people ask say, well, why not just play with my existing friends? I should just be able

to play with my existing friends. I like my existing friends. I don't want to form new friends cuz that's kind of scary. Um, why not just play with existing friends? So, let's look at the the logistics. Um, there's a thing in online games called the the concurrency ratio, where there's certain number of monthly active users that you have, there's a certain number of concurrent users, and there's, you know, a certain number who are online at any one time. So let's say in the best case, the best case, 10% of your audience is online at any one time. That's wonderful. Um, all right, let's let's crunch the numbers. Um, so assume let's assume a concurrency ratio of 25 to one and you have 10 friends. This is a secret, not very many people know this, but most people do not actually have 10 good friends. Um, most people have maybe five to seven. Uh, in online games, it's really really common to have see friends list with like one or two people in your friends list. So 10 is being generous.

Um, you come online and there's only a one in three chance that someone is actually going to your friend is going to be there. So most of the time you come online, you're not going to know anyone. And this is in a generous, good situation that doesn't actually occur for the vast amount of humanity in real life. So, we can take some takeaways from that sort of thinking. We can say, okay, well, it turns out that um, friendships, old friendships rarely translate to new contexts, particularly when you've just launched an online game. In part because the numbers don't work out. Like, the friends won't be there. It's just mathematically not going to happen. Um, so then the design problem starts to be, well, I would love to have real friends in there, but they're not they're not around yet. So I need to create an environment where people can form new friendships from strangers. And that's a completely different design

problem than hanging around the table at a board game and playing with with uh, people you know. Cuz it turns out that the psychology of interacting with strangers is totally different than the psychology of act interacting with friends. Okay. So, theory's awesome. I like theory, but uh, let's let's do some tools. Does anyone here own a hammer? Aren't hammers awesome? I think they're great. Um, oops. I screwed up there. Okay. So here are your basic tools. Um, you've got um, things like persistent identity. If you don't have identity, um, uh, if it shifts from game to game because you're randomly uh, assigning usernames, um, people are not going to have repeat encounters, you can't have friendships. Um, events. If you do an event, um, you

can increase the density in time, so therefore more people will show up, so you escape the desert, you know, problem and you can actually have people there if you have a low concurrency environment. Uh, daily incentives encourage people to come back, so again, you're trying to get people into your game so you can play. Um, and then offline communities, like, okay, what if you had an asynchronous community that existed side by side with your synchronous community such that people could connect up that way. So there's this parallel path that you can do. So these are all super good basic tools that work well. Um, and then you've got anti-patterns with that as well. Um, anonymous players. Well, if people are completely anonymous, they can't make friendships because they'll they don't know who each other are and they can't have repeat interactions. Um, the the desert, huge empty areas. Um, this is when you're getting into competitive games, you're like, oh, I know, I'm going to have lots of gameplay modes. Now, space is not simply spatial.

It's not uh, it's not a problem of geography. Like you can have abstract spaces well. And when you take your population and you split them across multiple modes, what you're doing is you're basically reducing density. You're you're you're making it so that people, those little molecules of people, don't bounce into each other and interact nearly as often. The same thing goes with skill. When you have an MMO and you have like the newbies and the experts and there's this spread of people across either a power gradient or a skill gradient, what you're doing is you're actually you're you are you are changing the density equation in your in your game of people who can interact with one another. Um, and it's really easy to do that. You're like, oh, no, this isn't a social system. This is this is merely the leveling system. It has nothing to do with social. But by changing these systems, they actually impact the experience and hurt and it can utterly destroy your ability to friend to create friends.

Um, this is this is this is sort of the whipping boy of um, of making friends um, matches. Um, the this comes from board games and sports. It's super popular. It's super super problematic. Um, so matches are used you start a match, you everyone has to be in there at the same time, you play a game, and then you end the match and it ends at the same time and you move on to the next match, hopefully maybe. Um, uh, so uh, this was a recent recent game that failed. They simply didn't have enough um, concurrent population to support their matchmaking structure. So people would have to wait too long, which kills the retention. Um, the retention drops, which makes the problem even worse and you get a death spiral. So, you know, me millions and millions and millions of dollars of gameplay screwed up by logistics. Um, and that's a that's an architectural issue, right? That's a structural issue to their game that requires that sort of initial starting conditions for players

to have fun. Um it does work. Matches do work. Matches are super effective if you have a mega hit. If you have a mega hit, you have a huge population, matchmaking is easy, wait time is easy, it all works out. So, if you have a mega hit, keep doing this. I'm not going to argue with you. Um but um it also sort of locks people away in these rooms and they don't interact as much. This this it's it's kind of if you're anyone who'd has any sort of like soul for risk risk mitigation and you have a chance that you might not make a mega hit, do away with this pattern. Consider something else. Um certainly we can do better. So, so this is one that I really like. This is sort of comes from uh from MMOs. Uh this is a instead of matches you do rooms. Um and uh this is a solution that

has uh density management. Um you can do repeat encounters and very very low wait times so you get less churn, which keeps your concurrency up. Uh really simple sort of like diagram how I think of rooms. And rooms could be a huge open world, they could be like uh you know, an intimate abstract um session, you know, there's there's tons of ways to do rooms, but the basic idea is you have people playing a game, there's a certain number of slots for people to play the game in the room, and then ideally you have these two other sort of things which is this idea that you can join in progress. You can pop in at any point. You want to start playing the game, let's start playing. Uh and you can leave in progress. You can leave at any point. So, it's very it's very friendly. It's like uh you know, it's it's respectful of of people's time, it's respectful of their schedules. Um it's it's a nice nice pattern. Um uh this is one um slither slither.io. So, there's uh dot io games are are massively popular right now.

Uh everyone here should be building one in case you're not. Um and there's no matchmaking wait. Like in order to play the game even including loading the game and entering the URL, pressing enter, it took me 6 seconds to get into a game and I was instantly playing with a whole bunch of other people. Um so, very effective. Um in a bigger context, you look at something like Guild Wars, Guild Wars 2, um and they have an interesting pattern which is an extension of room. So, you can take this very simple concept of rooms. Does anyone here from ArenaNet? Maybe one or two people. I I I see one. He's wearing the shirt. Um uh and um you you can take this simple room concept and you can extend it and you can start building things like a Lego block, a social Legos, you're building these architectures out of. Um so, this one uh they have a problem where what if you pop into a room

and there's no one else there? Well, the typical thing that you do is you leave. You're like, this is boring, a big empty room, I'm going to leave. So, instead they do they they build out this lovely map and they have little checkpoints on it and little quests that you can do. So, those little hearts are like little activities that you can do if you want. And they say, why don't you just do these little activities? And so, that's what people do. And but what that does is that keeps them in the room until other people start to accumulate in there. And once the other people are in there, they now have sufficient concurrency to do a much more engaging, much more social interaction. So, that's when they hit you with a public event and now oh, now we're all playing together when we were just hanging out in the room together. Um I think doctors' waiting rooms should totally do this. This would be awesome. Like you're sitting there in the waiting room, there's some magazines to keep you going, you know, there's like a sticky note and you're like, I wonder what the sticky note. Then you put on the sticky note, you say, secret loot, you know, or

something on the sticky note. People are looking through and they're like, oh, what's that? And then all of a sudden, boom, a dragon spawns. That would be that would be I think that'd make them better. Um Now, see, there's your simple pattern, you know, waiting room, go into event. Um and and again, going back to the theory, it it's um it's uh incre you have you've got you're managing serendipity, you're managing density, you've got this idea of this ideal density that you're aiming for, and so you can you can create that in your game. We did this on uh we made a a cooperative game Realm of the Mad God and we did the same thing here. Um this was a slightly different pattern. Um this was how do we get to that ideal density? Is there any mechanism we can do sort of like um where imagine again this idea of these people bouncing around like air molecules in a space, um is there any way we can sort of increase the density or shrink the density? And if you think of a balloon,

that's exactly what happens in a balloon, right? You can kind of increase the balloon and shrink the balloon. Can we do that sort of managing of density with um game servers? And so, here we have we have these little instance servers that are running and you can jump into one and you can look up there and you can say, oh, wait, there's 14 of 80 people in that room, so there's 14 slots filled in that room. Uh and then we've got another one with another number of people in that room. And then there's a hub area here where everyone's gathering and saying whether they'll join those things. Um now, this is really cool because now we can just go and create the number of instances that we want um to make the ideal density. And so, you kind of a hub area and then people freely flow back and forth between the hubs and the instances and then you just add as many instances as you need to create the amount of space you need for the number of concurrent users. Math. Right here.

Um really simple, right? Just simple simple load balancing. Um then, you know, what happens when the population shrinks? Um so, we have this idea of garbage collection. Like, okay, we need to shrink the thing, but if we just go and like shrink it, maybe people will get upset. So, in Realm of the Mad God, we we made a game play reason for this. The each instance has a big baddie, a big bad evil person. Um and uh the the god of that that area. And if you kill him, the whole thing gets destroyed. The whole the whole world gets destroyed. But that gives us a nice narrative excuse to just delete that instance. And now we send everyone back to the hub and now we can load balance the sort of like how many how many instances we need. Um Guild Wars has a slightly different. They were kinder in some ways than we were. We just we just have people in a permadeath game fight a giant monster.

They decided, hmm, maybe we should ask them and give them something nice. Um which is nice. So, here they bribe people to leave the server. They're like, hey, uh we're going to like it's kind of boring in here. Do you want to go someplace else? We will give you stuff to go someplace else. Um and if that didn't work, they kick you out. So. The carrot and the stick. Um so, as I was thinking about this stuff, I'm like, okay, that's all good. What what what the density systems, what they are all doing is they're basically increasing sort of uh serendipity. Um but what I'm really interested in cuz I want those those those meaningful relationships, um how can we um how can we increase that repeat interaction? So, we worked on this uh MMO which nobody has played, I guarantee. It was on a a a platform that doesn't um doesn't uh travel as well as some of these other ones.

Um and so, this was a this was a multiplayer Animal Crossing-like game uh and you had little bears and they were running around. Uh and uh they were all part of a the same cohort. What we would do is we would stick 40 players together in a village and we say, you are going to live in each other's sweat and bile for the next 3 months, year, doesn't matter. You're just going to live together in this place and you have to get to know each other. And this and you've seen this in schools. Um many schools do this where you have a class, a cohort, and you like stay with the same cohort, and you either become lifelong enemies or lifelong friends. Um and and it's this this really wonderful intense repeat interaction and and increasing pattern. Um and uh we had this whole system for like managing this, you know, like what if villages fail and how to migrate and what's the narrative behind it. And so, when you build these things, you think

of the flows. Think of the flows of people and think of how to how to maximize. How do you maximize um the opportunity for people to stay together? Yeah. You can do this in first-person shooters. Um so, first-person shooters simple thing. Like when you're going from one match to the next match, why don't you keep the same players together, especially if they're on the same team? It's the same idea. Same exact idea. Um so, here you know, you you you you you start to get tools, right? You start to get a list of tools for sort of managing your population of players across servers. And now, when you build your game, you're thinking about this stuff. You're building all these things in. Like um uh, BearTopia had all these systems going on simultaneously in the game, and it was a fundamental part of how we designed the game. It wasn't something that we thought of naively later. It was like,

"Oh, wait, we're building this game and we want to encourage friendship. Let's make sure that every piece of the game, the the idea of villages, the the the idea that you're staying together, the idea that you're going to leave." Like, every once in a while people wander off on a boat adventure to go someplace else, you know, to explore the old country, and you know, they come back, and then the land is changed afterwards. And so, we're building these ideas, these concepts, these logistical concepts narratively into the gameplay. Okay, similarity. This is a short and horrible one. So, first first rule of similarity is, um, people make friends with people like them. If you have a group, um, and you randomly mix up a bunch of different people and they have some sort of attributes to them, people will look around the room and they will instinctually make friends with people

like them. And this has been studied in all sorts of situations, and this is an incredibly powerful thing. So, like, "Oh, those orange those orange things, like, anyone else wearing orange? I'm instantly your friend." Just like that. Um, like, I want to become friends. Um, now, um, what what what's happening here is, um, we perceive that similar people are going to be like us, and then as we build friendships, it's going to be cheaper to negotiate with them and figure out how we could work together. It's like, "Oh, if you're like like me, we're already talking the same language, we already probably have the same ideals, the same norms." And we sort of we we sort of project these like benefits to similarity. Um, this is not always positive. Um, in many many many many cases, it's incredibly negative. Um, and it's also a wildly strong

filter. Like, some there'll be someone with a piece of cake sitting there saying, "Would you like some cake?" And it's obviously amazing cake, and there'll be someone next to them with a knife who wants to stab them, and who has but looks like them. And they'll go up to the person with a knife and they'll say, "Wow, we should become friends." Like, it's it's a it's a brutal brutal thing that has been repeated in thousands of studies. Um, so, um, this game, which I always mispronounce, Agar.io, is my attempt at it. Um, another another, um, one of these dot io games. So, they said, "Heck, like, people really like this nationalism thing." And so, they allowed people to put, um, stickers basically, uh, skins on their characters of national flags. And for a while there, what was happening is there was these massive nation versus nation battles going on in this game. And along

with it came all the baggage. All the baggage of nationalism. Like, it's like, you know, slurs just, uh, just uh, there were denial of service attacks, there's just horrible horrible things because you you took in this like tribal identity and you applied it within your game. Um, so, how the heck do you use this thing? This is like, it's really really strong and it exists in all of our communities, and if we ignore it, we basically basically import the worst elements of modern society into our games, and people create these poisonous communities based off bad ideals. Um, and so, and and the the the diversity of our population drops, our our populations become like, we we we we basically start doing gatekeeping where you're not like us, you're you don't belong here. Um, it's harder to acquire people players, it's just a miserable

thing. It's like one of these dark secrets at the heart of humanity that makes designing online games really difficult. Um, you know, you can you can put people in factions, you can design these factions, um, you can have people have sort of shared experiences, positive shared experiences. Um, you can have them do shared goals, um, and these things help a little. They sort of work. Um, but they're not that effective in general. The one thing I've seen goes back to the fundamental reason why similarity is so strong. And the fundamental reason why similarity is so strong is goes goes back to that that idea of shared social norms. We are deeply deeply pattern matching creatures. And we're not that smart most

of the time as individuals. What we do is we look around us and we say, "What are other people doing?" Particularly people like us. And then we copy those behaviors. We start to mimic those behaviors in this new unknown social context, and we start doing that. So, you can use that to set up positive fictional identities. You can create a tribe for your players that is actually super positive. So, in BearTopia, there was this, um, we're like, "This is a happy little village where people are nice to each other. Everyone's an immigrant, so there's vast diversity." That's That's actually a similarity that you're all immigrants. Um, and you're building things together and you're working towards these common goals, and everyone's just really nice. Um, and what we've noticed is, um, it uh it actually works really really well. Um, people don't people don't get angry at each other

because that's not the tribe that they're in. They're not in the get angry at each other tribe. They're not in the like, um, wear black and beat each other over the head with a stick tribe, right? That's not this game. And so, by creating these positive identities, these prosocial, uh, player roles, um, you end up sort of setting the norm for this is what this place is. Um, and people will share that identity, and they'll adopt people into that identity, and they'll be like, "Oh, you're like me, so therefore we should be friends because we're all in this very strong prosocial identity." Um, so, some some rules there. Do not reference real strong real world identities directly. Like, keep that out of it. If there's a strong real world identity, sort of separate that from your game. Um, the um, yeah, and and it's really effective. Um,

what so, we make another game called AlphaBear, which, um, is a game where you collect cute bears, and it's funny because you should read the support emails we get. The support emails are like, "I'm playing this with my child and I just wanted to let you know." Um, or "I think I found something that horribly deletes all my data, and I wanted to apologize." Um, um, and it's just the sort of game that attracts and presents a viewpoint on the world that is pleasant and nice. And you we can build those sort of games. We can make that sort of game. There's nothing about games that say it has to be about ultra violence and the fall of man as we turn into animals. Um, what if we make games for real human beings being the best that they can be. Um,

I like that. All right. Um, reciprocity. Anyone asleep yet? No? This guy, wake him up. So, this is reciprocity. All right, this is we're going to we're going to do a simple thing here. Um, he's fallen asleep. You're going to punch him in the shoulder gently. All right, and then he's going to look at you. And he's going to go, "What the heck?" Right, that was reciprocity, right there. Tit for tat. Right there, he could have punched you back, that would have been an extension of that. Um, but, um, you you make a friendly gesture or offer, um, and then in return, you know, the other person gives a positive response, and you've got a nice little reciprocation loop, right there. Uh, doesn't have to be material, it can be, you know, attention, recognition. Two people walking across the room, one person looks at the other person, the other person responds by meeting their gaze. That is reciprocity right there.

That is a reciprocation loop. Um, now, what's interesting about reciprocity is every time you do that interaction loop, you build this magical resource known as trust. You're building trust with each other. It may be just a little bit. Like, when you meet glances, you're not building very much trust. You're building a tiny tiny little amount. Like, "Oh, this person seems okay. They seem to be following basic social norms. Um, maybe maybe we can continue this later." Um, but trust trust is interesting because it goes back to trust is shared social norms. Um trust is I know I can predict what you're going to do in any sort of social situation. I I know what you're going to do and I know how I'm going to react and you know how I'm going to react and therefore if there's a complicated task or a risky task, I know

I can rely on you. That's sort of the economic sort of underpinnings of trust. And if you don't have trust, then there you can there's just a limit to the complexity of thing that you can do in life. Um GDC could not have been put on without trust. There needed to be trust that there would be people here that we would meet each other, that we would have good talks, and if if that's broken in any way, then the whole large organized culture just falls apart. So trust is wildly important and reciprocation is how you get it there. Now, trust is is hard. Trust is super hard. Um trust trust grows very slowly. Um it can take years for people to trust one another. I come from uh I come from Maine. Maine is a cold weather climate, which basically means that every time

winter comes, there's a 110% chance you are going to die. That's just fundamental to the culture. It's going to happen. Um and you have limited resources and every day is like survival of the most brutal kind. And it may not actually be like that cuz we have stoves and stuff, but that's what we're all thinking deep down. And so trust is like wait, anytime I like ask someone over to dinner or ask a favor of somebody or someone asks a favor of me, like I'm giving up scarce resources that could have been essential to my survival. Um and so trust grows very very slowly. We slowly say, "Okay, I suspect that person is not going to steal my awesome wood pile." Um and eventually over years, you might get to the point where you wave hello. Um

that's cold cold weather climates trust growing slowly, but then the horrible thing about trust is if someone like betrays that trust, it goes away super super fast. Like if I see that guy taking wood from my wood pile, he he it's over. I'm never going to talk to him again. I'm going to spread rumors to everyone else. I'm going to like, "This guy, he takes wood from the wood pile. That's my wood pile." Right? And it's like one moment and maybe he really needed that wood, whatever. It's like trust can drop quickly. Um trust also um allows us um to increase the stakes of each interaction. Like maybe we get to the point where we're like I'm letting him take stuff from my wood pile. Awesome. That's good. We're trusting each other. Um but it's expensive for me to like make dinner and have the fellow over for dinner. That's

like what if, you know, what if we don't get along, what if we're only good at the like saying hi level and we should probably should have kept it there. Um so trust like as you start interacting and as you build up that history of interactions, you start being willing to do more for the other person. You start being willing to take greater risks. Um but the flip side of this is deep friendships are incredibly expensive. So the the end result is you can only end up investing in a handful of of deep friendships. So we have a lot of shallow friendships and then a very very few number of deep friendships. Okay. So trust source of civilization, very important. You can tattoo that someplace. Trust is the source of civilization. I have many tattoos on the bottom of my feet. That way they're it's like private, you know.

Um okay. Now there's a downside to this. Um we are social creatures and we are wired to reciprocate. So there's this wonderful experiment where you go and you give person $1 and then you ask them to buy like a $20 thing from you and they're total strangers and they absolutely do not want this $20 thing, but because you gave them the $1, they're like, "Oh, I'm biologically wired to reciprocate." Um and then they go and boom, they'll go and they'll they'll complete the sale with you. Um so you can create these reciprocation systems, but they they end up being um uh they can they can actually be kind of uh manipulative if you're not doing it with good intent. Some tools. Okay. Anti-patterns. Uh anything that reduces trust. Um if you're building a trade system and people can scam one another, you're actually actively reducing the global trust in your community. Um if there's

um hacking that's going on, you're reducing trust and you're causing people to not want to become friends with one another. They start to become offensive. They start to become like manors and you don't want that. Um power differentials. Um like I think of the nation right now with the extreme income inequality going on and there starts to create this like system of mistrust where like people like you are thinking in a way you have access to resources that I don't. You're taking them from me and you start to get this this sense of distrust. Um now what's fascinating this is this is one of my absolutely favorite things about uh uh reciprocation tools. We have built an immense number of reciprocation tools. If you are a startup in this area and have taken VC money, you have engineered reciprocation loops.

And you just try to get them out there and you try to get people using them as much as humanly possible because damn they're cool and they have high engagement and people love them and they're socializing and interaction and all your numbers are going up and it's super exciting. Like chat, let's do chat. Let's let's have friends lists. Um let's go and have checklists where we're like trying to like compete on the number of things that we're selling and we're comparing that with each other. In games we can do guilds, we can do gifting. Like all these things we can dump into a game um or app if you're an app person. Um and we can just do that. And then the question is wait wait, but but like we talked about Twitter. Twitter is the the case study of throwing random crap into a thing and trying to get as many people to use it as possible and then watching the wreckage fall out. Um the human wreckage. Um and that's

maybe maybe maybe we could be smarter about it. I have a I have a I think we can. I I think we can. Um so it turns out that if we go back to that idea of like trust builds up slowly and there's low cost international interactions early on and then high cost interactions later on, maybe we can structure the way we interact inter uh introduce tools to people based off levels of trust. Um this is sorry, that's more than I like to put on the slide, but um this is different levels of friendship in a game. And you can think of the the the absolute lowest level of friendship as

someone playing by themself. Single player games not commonly known for their friendship formation capabilities. Um the next one is what kids do. If you ever watch kids play games next to one another, they'll they'll do this thing called parallel play where they're playing they'll they'll play next to each other, they'll kind of like glance at each other out of the corner of their eye to see what they're doing, but they won't actually interact. Um then you've got um ambient co-op. We're helping each other. It's like oh like in Steam Birds uh here I'll I'll do some examples here, but ambient co-op we help each other without meaning to. Soft co-op. Um we're intentionally benefiting others, but there's no requirement to do that. There's no requirement to actually help each other, but you can do it if you want. Uh and then hard co-op, the Nintendo puzzle where like someone has to step on the pressure plate but so before the door opens and if if you don't step on

the pressure plate, the door will never open, the other person won't be able to go through and that's the sort of locking key hard co-op. If you do not co-op, you will cooperate, you will fail. Um and then finally we get all the way up to friends where we're spending spending time together, we're trusting each other um and we're willing to collaborate and communicate and solve. We're able to actually interact at all these different levels. Um but we can do hard co- uh cooperation without backstabbing. Um yeah. And so that's friendship levels. Uh this is sort of like a very simple interaction on Discord. Why is Discord so popular? Um, they do a lot of smart things, believe it or not, in terms of friendship formation. So, one of the things they do is they actually pay your bill when you join a um a new server. They pay your social cost of joining the new server by making you seem funny and witty. Um, so they introduce they're like, "Oh,

this person's here to kick kick butt and chew bubble gum, you know?" Um, and they they make up a little witty thing. And so, they're doing this very simple thing to sort of like ease you in and introduce you to the community in a nice way. And then so and and I see people actually respond to these. Um, not always, but sometimes they'll respond and they'll reciprocate to this little thing. And so, at the very lowest end of the friendship ladder, you can do you can do these things to automate like the the the suggestion of how people should interact. Um, in in our in our MMO Steam Birds Alliance, um, we do this thing where um if people kill an enemy on the screen, everyone around them gets the XP. So, it's it's a a form of super ambient co-op, right? Like, we're happening to bene- benefit each other. We're happening to play play together just by

being on the same screen and shooting the sa- and and someone shoots an enemy. Now, that can move into uh soft co-op where we're both actively shooting the same enemy because then we're like we're thinking, "Oh, wait, I'm being totally selfish here." This is a totally selfish activity, but by shooting that thing while you're shooting the thing, we'll shoot things fa- we'll kill things faster and get loot and XP faster. So, it's a totally like very low cost interaction that results in us doing something together and completing reciprocation loops and having a shared experience. And and now suddenly we have shared goals um and all that is sort of just sort of naturally happening just by playing the game. Um, I I really love this quote um on uh Destiny 2. So, they have a thing where they have uh non-zero-sum uh guild rewards. So, if if someone goes and gets a

a reward in the game and they're part of the guild, then everybody gets a reward. Um, and uh the the designer on this, Emmy Chung, is saying like the Super Bowl ring philosophy. Um, when someone wins the Super Bowl ring, they they don't just give it to the people who were on the field and who played in that game. They give that ring to everybody. And it becomes this sort of like shared like, "Hey, everyone helped do this. I'm going to reward everyone along with us." And even if you What's nice about that is even if you were actually in the sort of hard co-op um environment, you were in the hard co-op match where you had to work really really hard together, this is something that benefits every level. And there's um one of the challenges with degrees of friendship is is you've got a mixed audience. You've got people who are casual friends, people who are deep

friends, people who don't know each other all that well, people who are like amazingly good friends and they're probably going to get married inside your game. Um, which happens. Who here has had a game where someone's married somebody else inside the game? Yeah. Yeah. So, that's the level you can get pretty intense here. Um and you need game mechanics that work for that entire spectrum. You're not necessarily just developing game mechanics for one group at a time. You're looking for game mechanics that are flexible enough to work for all different levels of friendship that enable the super super deep friendships as well as the um more casual ones. Um this is a pattern that I love, uh roles. Uh roles exist in a lot of things, but I just like talking about Puzzle Pirates cuz it's super cool. It's a game where you're a pirate and you play puzzles. I

mean I don't know how you can beat that. Um, that's what it looks like. Look at that, pirates playing puzzles. I just want to take a moment and appreciate that. Um, you fail if you don't work together. Everyone's got a different role. They start relying on one another. There's a differentiation of of of roles and capabilities. Um, there's as I was building this talk, another talk came out that talks all about trust. I highly recommend you go check it out. Um, this is the spectrum of trust. Raph Koster has a copy of this up on his website. Um, and this is his slide on roles. So, it's this idea that like low trust environment, you you don't have differentiated roles very much. Um, like I can do pretty much anything you can do. In high trust environment, you're allowed those those roles to spread out. Now, I need to rely on you. Again, this goes back to the soft and hard co-op idea.

Um Yeah. All right, disclosure. This is a pretty simple one. Um in order to make deep deep deep relationships, you've worked your way up. You're doing more expensive interactions with another. You have to um you have to become vulnerable to the other person to reach those those maximum levels of trust. Um, you need to have riskier uh trust building going on. Um, the problem here is by revealing things about yourself, it's actually like what if the person rejects you? Everyone here has a fear of rejection. Someone's going to look at them and say like, "Oh, if they tell me about this intimate detail of my life that all this work

building up this relationship is going to go away." Um, but it's the only way. It's the only way to get to that next level. And we need to make games that safely support vulnerability and disclosure. Because if we don't, bad things happen. Like um if we reveal things too soon, you get all sorts of stuff that happens here. Um so the way I think about it is there's opt-in disclosure at the right time and place and there's premature disclosure. And if your game causes premature disclosure, then you get all the negative impacts. You get horrible horrible things. Like this is an example of that. Um, this is a this is a woman who is playing Overwatch and um she can't get she can't use voice chat without getting marriage proposals. Because voice chat reveals that she is a

woman and it's a premature disclosure and it turns out there are a lot of sexists jerks out there that she didn't actually want to be friends with, but because she's forced to disclose this information, they're able to go and put all their their horrible nature to work making her feel unwelcome. And so, she wants to like here he's I mean, this is a very clear like critique of the system that says, "Hey, you forced me to disclose something I didn't want to disclose." Um and what happens is people turn off chat at that point. And now your opportunity in the future for a meaningful moment of vulnerability is lost. friends friends at an early point. Um, and that your community as a whole suffers because of the systemic things you put in place with your naive social

systems that caused people to defend themselves to turtle Um, so you know real names are horrible idea. Um, identity and persistent identity is a good idea. Real names are miserable idea because it forces pre- me premature um disclosure. Um, voice chat also a bad idea. Location is a bad idea. Even if someone's purchased a particular item is a bad idea. Like be very very careful about what people can share. Um, and if they want to share, give them tools to share, but make it an opt-in activity. So, this is where I kind of put things on like where you should start introducing various uh interactions. Um and you want those private spaces where people can talk to each other so that they can have those vulnerable moments. And anyone who says like, "Oh, no, we

can't do chat because chat is risky." Chat is actually essential. You need people to be able to talk to one another. Um, this is the cone of silence uh idea that we did in Realm of the Mad God. Um, you have secrets and people can't hear each other outside the cone of silence. There's ways of creating these very simple opt-in systems that the players don't even think about this, but you this is an this is a super simple thing you can do that like totally changes the um the ability for people to control their experiences. They can watch someone coming towards them and say like, "Oh, I'm going to stop talking about that right now." Um and of course um you need to give real friends, which will come up, a way to skip this. You need people to opt in to skipping this. Okay. So, those are the basic things, you know, and

that's how you design for friendship formation. This is sort of like where those things map. So, you've got these four sort of like tools, and then where they map onto sort of the level of trust. Um and like certain things like if you're dealing with like parallel and solitary play, you probably shouldn't be thinking about disclosure. In fact, that's a negative thing to to have in your game. But by by the time you're dealing with hardcore up and friends, that's an important thing. And so, you can kind of like think of your game as this sort of like funnel of like people becoming friends and how do you support it at every stage? If you've ever built a match-based game with a large number of people where the social identity is beating the crap out of one another, you may have noticed that getting people to be nice to one another is a hard problem.

It's This is just hindsight, but it may have been a good idea to actually build a positive social structure into your game from the very beginning. Because it's crazy hard to retrofit. Um I I will call out League of Legends here because they have done heroic work trying to fix the experience of a really bad architecture. Um heroic work. I mean, I think we should applaud them for the work that they've done. But at the same time, you don't have to live their life of making that mistake, right? So, build it in from the beginning. Um and this is a design process. Social is not something that is a separate pillar. It's not something that's an adjacent

system. It's not tacked on at the end. It is fundamental to your game, and you should build you should be thinking about social. You should be thinking of it honestly as the reason you're building your game. You should make games so that people make friends. Um some resources. You can look this up. I'm totally out of time. I apologize to all the people who really care about that. Um and that's it. I got no time for questions, but I'll