 Cheers. Oh, cool. Yeah, thanks for coming, everyone. So before we start, I've got a lead with a bit of a warning. I'm going to be talking about game design and game mechanics and a competitive online professional tournament gaming. But I'm not a professional online tournament gamer or a professional game designer or anything like that. I write software. I have done some game design and some gaming tournaments and stuff. But again, it's always like a kind of obsessive amateur level, right? Like kind of everyone else at EMF. And so I try my best here to like gather as much like genuine scientific research as I can. But a fair chunk of this is like kind of my reason conjecture as someone who's played a lot of games. All right, so first, I've kind of got a point out that all trolling is necessarily bad. So by definition, all trolling does wind someone up. But it's hopefully kind of, it can be constructive or funny or at least memorable. And so you get the picture. I kind of broadly dropped these into two categories, the first of which I really don't have a problem with. So this is an example of something in the first category. So this was a Boston World Warcraft. He used to hang out in the world and big, big organized groups of like 40 people would try to team up and kill him. But one group of guys, instead of killing him, they figured out that by exploiting this really complicated sequence of attack and run away and fame, death, and attack and die and do some fancy ability, then they could force this big demon dude to chase them more or less indefinitely. So yeah, gamers being gamers and being trolls, they walked him halfway across the world and dumped him in the other factions' capital city, where he just murdered literally everyone. All those corpses are player corpses. They're not sprites or anything like that. And if you search online, you'll find loads and loads of examples of these. They're great. And I think it's all online. You'd find teams of people who'd walk around and they'd find someone who'd gone to answer the door or make a cup of tea in their AFK. And while these people were AFK, they'd quickly build eight chairs, one for each pile around the person, because you can't move through chairs. And these people were then just stuck there until they got a mod to bump them a few taxes to the right or like switch server or something. And so these are obviously kind of trolling or griefing, but I think they're also pretty good. They're clever. They're memorable. They're funny. And they force developers to think more carefully about the games they're making, because then they know that if something can be exploited, then it obviously will. But then there's this dude. And this dude I really do have a problem with. And having been growing up around and having been exposed to the internet for a long time, I think probably everyone here has seen examples of this guy. But in case you haven't, in a kind of horrifying blue Peter fashion, I've got a bunch for you. These are real chat logs taken from League of Legends. It's like online tribunal system, which is like a community court. And I've got to put a warning, because the first one of these, at least, is really, really not kid-friendly. It's totally rude. So if you're easily offended or you're young, just shut your eyes for the next couple of slides. OK, cool. So yeah, so as I was saying, there's people who are just straight up offensive. We've got the kind of bigoted, homophobic, sexist, racist douche. And then the one I particularly like, hey, they're kind of really, really awful human beings. And we've got the guys who are like, oh, uninstall the game and then kill yourself. OK, so to us kind of people who are used to this, and once they're out of the context of the game, I find those kind of funny. There's a whole website devoted to these things. And the problem is there's people, a lot of people, who haven't grown up around the internet, haven't grown up playing games, don't have the ridiculously thick skin that a lot of gamers have. And those people are genuinely put off by this. Just one instance of this, let alone a 45 minute game of people just targeting them specifically and just not letting up. And that's not cool. Even the people who do have a thick skin, I don't know anyone who doesn't get tilted by someone just flaming them. So at the very, very least all you're doing is just putting off someone in your team and then making them perform worse. So a fairly common refrain other than, oh, grow a thick skin is, well, why don't you mute them? There's muting features in the game. But when you're playing in a competitive level, you really can't mute someone. Because if some dude's creeping up behind you and he's going to shoot you, and you muted the guy who saw him and is telling you where he is, then you're going to get shot on the back of the head. You're probably going to lose a round. The dude who you muted because he was flaming you is going to flame you even more. So yeah, that's a bit rubbish, really. I just want to walk towards a culture and a culture in gaming where everyone who wants to play and wants to compete is made to feel welcomed. But yeah, there still really isn't the case. People are quite homophobic, and they're really, really, really sexist and misogynistic. I don't know any women who play online computer games to use their microphone ever unless they're just playing with me or my friends. That's not cool. People should be able to compete and play at the same level as everyone else. And so I think this is also particularly important because online games, particularly, are a relatively new thing. And they're kind of these semi-anonymous spaces where you can play and compete and talk and hang out with your friends. But this never has to be tied back to your actual real identity. And I think for some people, particularly people who, perhaps, don't engage with traditional forms of competition like sport, maybe because they're not able-bodied or they're really socially anxious, then that's really, really helpful for them. And the more we can do to enable that, the better, really. OK, so that was a motivation for if you're an empathetic human being. But what if you're a cold-hearted game studio driven by money? What if you're Ubisoft? So there's a bunch of guys at Group Lens, which is this really cool research institute at the University of Minnesota. They scored loads of League of Legends players with respect to how toxic they are. And then they found that people who were new to League of Legends who were playing it for the first few times are like they hadn't got to level 30, whatever that means, I don't play League of Legends. If they played with people who were toxic, then they were far, far more likely to either stop playing the game entirely or to take an extended break. And so that's quite a result, right? That's saying that people who don't play games or at least don't play League of Legends are significantly part of playing more League of Legends if people are douches. And yeah, this effect kind of tails off as people get more, as people play more, sorry. And so the point of this kind of is that if your community is not garbage, then you're gonna have more players. And if you have more players, you're gonna have a stronger player base. If you have a stronger player base, you're gonna continue to sell your game. So you are financially invested in building a good community. Okay, so that was the kind of old model of video gaming, but the free-to-play thing is really big now. And so people apply the same arguments because free-to-play retention is even more important than it is when you're not free-to-play, right? Because when you're not free-to-play, you've made a whole load of money by selling the game in the first place. But if you are free-to-play, then you've got to keep people for longer because the longer you keep someone, the more hats you can sell them. So whichever way you put it, you are invested in building a good community for your game. And so the League of Legends guys, as I said, like they're in the hat business. And so they devoted a whole team to doing actual science, yeah, because it's really big money to them. If they can keep their players playing longer, then they're gonna make more money. So one of the first things that this team did was they assigned, so they've got loads of game data, right? There's millions of League of Legends played, games played every day. So they assign a toxicity index to every player by manually classifying thousands of chat logs. And as people who play games, and all these guys played loads of games, you might expect that there's gonna be some really, well, some fairly small minority of really, really toxic people, and they're consistently toxic. And then everyone else who plays was either kind of neutral or positive, right? But mostly neutral. Most people think, I'm nice. But actually it turns out that's totally not true. There is a small community or a small minority of toxic people, but actually it's a really, really small minority. And so this minority is also more aggressively toxic as well as more frequently toxic. But it turns out the vast majority of toxic behavior in toxic games turns from normal people who just had a bad day. They've had an argument with their girlfriend or boyfriend or they've been shouted at by their boss at work and they get home and they're really frustrated and then they start playing the game. And then the mechanics of the game interacting in such a way that it winds them up even more and then something happens, like they die or their teammate dies and then they go totally nuts. And to the other people in the game who don't know this person, they just think, oh, that's that normal like half a percent of really toxic people, there's nothing I can do about it. But to the perspective of the guy who's raging, he's not normally like this. This is just a bad day. And so this kind of brings us to this weird kind of place where like we know that the vast majority of toxic behavior is just normal individuals, but we also know that all gaming communities are not equal, right? Some are more toxic, some are less toxic, like some are really nice. And so we kind of like want to investigate why that happens. And here I'd be interested to see how much people agree with us, but I've ranked from left to right, like how miserable these games are to play with regards to like their communities. If anyone's played Heroes of New Earth, yeah, you know. I'm sorry for everything, that game like puts upon people. So yeah, so like why? This is kind of interesting, right? Why does this happen? Well, the first thing people think is like, oh, maybe there's selection bias. So like people who are horrible people gravitate towards horrible games with horrible people because they cannot just hang out and be horrible together. But that sounds accurate and I think it probably is accurate to some extent, but it's certainly not the whole picture because like loads and loads and loads of people play like Dota and loads and loads and loads of people play Rocket League and Rocket League has a fairly good community and Dota has a really toxic community, but a lot of the people who play Dota also play Rocket League. So like how does it work? How does one game like make them more angry than the other game? And so it kind of turns out that what actually happens is when you have frustrated players who have had something go wrong at your work or whatever and then you put them in an inherently frustrating game, then it just goes really bad. And if you can make your game less inherently frustrating then everything's gonna be awesome, hopefully. So I wondered why this was and I wondered why Dota was a lot worse than say Rocket League. So I wanna do like a sort of mechanical analysis of Dota and why this game is more frustrating and like Dota likes a kind of a whole range of games. Heroes of New York, which was the most horrible game I put on there is basically Dota, but like we've added tools to wind people up if you don't like them, which yeah, it's insane. And so you find, well, I think at least, this results from a kind of like trifecta of really grim mechanics. So first off, you've got like a really, really huge snowball and like I'll explain what these mean in more detail in a minute. And you've got a really, really huge snowball effect. You've got really like unintuitive mechanics that don't make sense unless you know a lot about Dota. And then you've got like a huge, really massive, insane knowledge burden before you can even start playing at any reasonable level. So this is a snowball. Like almost all games have a form of snowball mechanics. And what a snowball mechanic means is that like when you're ahead, it's easier to get more ahead. It's the opposite of a catch-up mechanic. And like different games do this in different ways. Like in Counter Strike, there's the economy system. In Dota, you get money when you kill people and you lose gold when you die. But the implementation of Dota's snowball mechanics specifically means that if one guy is really sucking, then it's better for you as a team to play 4v5 than it is to keep playing with him. Because by his sucking, he will be like actively feeding gold to the other team and making them loads better. Whereas in Counter Strike, if one guy is really sucking, like he's probably not gonna be actively griefing you. Like he's not gonna be shooting you. He's just not gonna be doing any work. So it's gonna be better to play 5v5 on the off chance that he kills someone. Whereas in Dota, it's really not. Like if a guy is really bad, the other team will just get way, way better. And so this means that flaming someone until they leave is your best chance of winning. So yeah, that's snowball. Now, I don't know how many people know what this picture is. Those of you who do will probably think, oh, that makes sense. Those of you that don't will be like what? So originally, Dota was a Warcraft 3 mod running on the Warcraft 3 engine. And like Warcraft 3 is an RTS that really wasn't designed for like competitive 10 player tower defense with hundreds of different characters. So lots of like really kind of weird side effects of the engine become baked in as mechanics because you can't get rid of them if you don't have control over the engine, which is what the Dota guys had. But over time, people become attached to these mechanics and they think they're good mechanics because they've grown up with them and they're used to them and they make them win over people who don't know about them. So with that in mind, this mechanic is called denying, right? And this is when you kill one of your own guys. And in Dota, you wanna kill your own guys as much as you possibly can. If you don't kill your own guys and the other team does kill their own guys, then because of the snowball mechanic, you're gonna lose horribly. What? You're gonna get really, really behind and of course like the tutorial doesn't tell you this because it's insane. Yes, so that's denying. Like I was gonna go into the specifics of why it happens but probably no one cares. So we've got the knowledge burden next, right? So this is a screenshot of the Dota 2 hero selection screen. There's 111 different heroes which each have at least four abilities. So that's roughly 444 different abilities in the game. They're multi-dimensional. So an ability might be like, oh it does some damage and like applies as slow for three seconds. If you want to play Dota in any way competitively, like not even professional level, just like low level tournaments or even like mid to high level matchmaking, it's just kind of gonna be assumed that you know every single one of these. What all of them do? If you don't know what all of the abilities do, then like you're gonna get flamed because you're gonna get killed by something that you don't understand and the snowball mechanic means that you're gonna get really, really far behind. It's ridiculous. So just in case that wasn't enough, there's also 148 different items. They all do different things. Some of them do different things on different characters or different things with different abilities on different characters which means that there's like actually just thousands like tens of thousands of insane edge cases like interactions between different heroes and different items. So sooner or later, you're gonna get killed by out of like nowhere by an interaction that makes literally no sense. Like you're gonna die, you're gonna have no idea why and it's gonna be because oh this hero had this item and I didn't realize that with this item it did this thing. So you're gonna be like this and then the dude on your team who's now really far behind because of a snowball mechanic, gonna be like this. Okay, so I just like kind of torn into Dota for the last 10 minutes but actually I kind of like Dota some of the time and like you might be thinking okay, so this game sounds really bad. It sounds insane. There's so much stuff that we need to know and the people are always really angry but actually like people still play this game despite all of the ridiculous stuff I've just mentioned because at the same time like the stuff it's really, really miserable and makes no sense to lose to. It feels awesome when you muster it. Like if you know this insanely convoluted data point that means that like your ability will actually damage this guy whereas normally it wouldn't and then you kill him and you win the game because of that like you're gonna feel like a god. So if you graph people's like frustration and enjoyment of Dota games over time and then you find that it's really, really spiky. Like you get these peaks where people love Dota because they're winning and they're winning because they know all of these insane awesome things and they're playing really well. But when they're losing, they're losing because they're losing to really insane things that they don't know about and it sucks and they've just got to like learn more and play more and it's really frustrating. And so this really spiky graph creates this kind of attitude where people in Dota really, really, really hate losing. Like losing is kind of a meme unto itself of how like unfun it is. And when people really, really hate losing they're gonna get more toxic about it because when they think they're losing and they think they're tied into like a 40 minute game but they're inevitably going to lose like no one likes that. Okay, so we're all hypothetically game designers, right? What can we do about it? Well, think about your design decisions really, really carefully. So like do you allow for fitting? Well, if you allow for fitting in a game like Dota then it means it's potentially gonna reduce its toxicity because the people who are getting really mad about being locked into like a 40 plus minute game, if everyone on the team agrees that this game is lost we can just forfeit, then they're gonna be less mad because they know they can always just bail out of it if they have to. But at the same time, like that's gonna potentially reduce the competitive like skill threshold of your game because if people are just like ducking out of all of the games that they think they're losing then they're never gonna like work out how to come back effectively. So in all of these you've got trade offs, like do you allow players to kick people on their other team? Well, this can stop griefing if someone's like team killing or like actively harming your team or it's just being really, really toxic like maybe you do just wanna get rid of them and that's great. But far more often than that case is the case where like someone gives up first blood or someone dies in a stupid way and some toxic guy on the team is immediately like kick him because he made one mistake and now we're really, really far behind. So yeah, again, it's a trade off. Do you allow cross team chat? Like can one team talk to the other team? It can reduce toxicity because other like people like to flame each other when they're winning and when the other team is sucking. But this like increased toxicity if you do disable cross team chat to get rid of it then your community is gonna take a hit because if you've got a good community where people like like to play games and like to talk to each other and stuff then they're not gonna have that positive interaction that they would otherwise. So this is kind of if you've already got a bad community then you might wanna do this to make it less bad but if you've got a good community then hopefully you don't need it. Yeah, and finally like what about voice chat? So most games do have voice chat enabled like there's a notable couple that don't. Again, interestingly like enabling voice chat can reduce toxicity because when you hear someone talking over the mic then you're like oh that's a natural person now I have some empathy they're not just like a faceless avatar behind a mystery keyboard somewhere. But also like it's harder to moderate verbal abuse reports if that's the thing that you do and also it does open up like more grieving opportunities like in kind of like the dude that just plays music down his microphone for the whole game. Again, like there's no easy answer to this and you just have to think about them carefully and how they fit into your picture like of your game as a whole. Okay, so I'm going back to League of Legends and again this is data from the dudes that did that study is a really, really awesome study like if you're interested in this go and watch the video. This is the only one for which I actually have any data or evidence but so these guys tried a bunch of really interesting things to see if they could reduce or improve their community. And so the first one that they did was with they set cross-team chat so being able to talk to the other team or not as a like an opt-in. And that like drastically reduced toxicity across a whole load of categories because now like the only people that wanted to like were interested in talking to each other were the people who'd explicitly said yeah let's talk to the other team. There wasn't like an overall decrease or I think it's like 2% or something I can't see from here but there wasn't a drastic overall decrease in total chat it was just the chat that was there was a lot better. And then the other thing they did was they added like sort of gameplay hints to the loading screen and things like that. And again like these improved the quality of the game significantly. Okay, again now these next two are kind of my conjecture about why I think these are good things. And so in Rocket League, those of you who don't know Rocket League is like a really kind of frantic game that's played with a controller. And in it the devs implemented this quick chat system and loads of games who have done quick chat systems like but they're all or they're mostly informational. And Sinox did things a bit differently so they made over half of them either apologies or compliments. And that does two really awesome things. So the first is that like because people are generally quite lazy and because the game mechanics aren't inherently as frustrating they're less frustrated by the game. And so it's much easier for them when someone messes up to be like, oh no problem on the quick chat. And then it is to like pull out the keyboard and start flaming someone. And the second thing it does that's really awesome is that it lets people like sarcastically troll each other in the way that like you probably would when you're playing with one of your friends. Like so like if you totally miss then you're gonna be like, oh nice shot. And this only becomes really obnoxious like when people start spamming it because before people start spamming it it's just kind of assumed, oh like they're just having a laugh. But yeah, so it's only obnoxious when they're spam and there's spam blockers in Rocket League. So if you say quick chat more than like four times in I don't know, 10 seconds or something then you're not allowed to say anything again for like the next minute. And like by and large the Rocket League community is really good. And I think this is a significant part of the reason why. And it seems small, but like if you play the game and you experience this, you'll be like oh this is actually really, this is really friendly. Everyone's just like chatting to each other through this system. And so next I wanna talk about Overwatch which it came out really recently and as part of their build up to releasing the game the devs made a whole bunch of like really interesting videos talking about like all the work they put into reducing toxicity and improving like the game feel for different people. And so the main thing that they did because this is a class based game is that like even though like Dodo and League of Legends and Team Fortress say are all class based games they're really lazy with their scoring systems. So like you hit tab and you see the scores of everyone in the game and the scores that you see are just like oh this guy's killed six people and they've died 12 times. Like okay if you're playing a role where your role is to kill loads of people then yeah that means you're probably not doing too well. But if you're playing a support role like you're a healer or I don't know like a guy who gives shields to people or something then like that stat is kind of meaningless but it gives people an excuse to flame you because they look at the stats and they're like oh we're losing. This guy has died two times as more times as he has killed someone. Let's flame him for the next 10 minutes because he sucks. So what Overwatch did was they worked out metrics that are actually helpful and like actually represent how well you're playing your role and then they highlighted them. So during the game if you're playing really well according to your role then it'll be like oh you're on fire you're doing really well and like everyone on your team and the other team can see that. And that has no actual like mechanical impact on the game it doesn't make you faster it doesn't make you stronger it doesn't do anything. It just shows to everyone else that you're playing really well. And again like it's kind of positive feedback here for you you're like yeah this is awesome. And the other thing they did was post game they display these four like little scorecards of four people who they think played particularly well or played their roles particularly well and you can commend them and like if you commend people then you're more likely to play more games with them in the future and the more commends you get like the nicer you are rated as a person. And again like these are really small things but the Overwatch community is also really really good so like they must be doing something right. It's difficult to say like how much of an interaction it is between the game mechanics and how much it is like kind of UI stuff like this or UX stuff like this. Obviously it's an interplay between the two but it's like I'm just highlighting them anyway. Okay so that was kind of really whistle stop tour. I wanted to talk about like well building a good community is important both financially and for the community is large. I did want to talk about games. Here is a new but purposefully foster a really grim community but yeah I didn't really have time. And so if you are interested in talking about that then like I'm going to be outside after the next talk because I want to watch the next talk. So just like find me then and we can talk about it and it'll be awesome. Thanks these are my references. If you're interested in them I can give them to you. The slides will be on to where or something. Cool. I don't know if we've got time for questions. Oh I've got some. Awesome. Go for it. Is there anyone? Oh yeah we've got questions here we go. I haven't played many games but a couple of games I've been involved in like Glitch. The reason why the community was so strong was because the developers were an active part of the forum. So relationships got formed and it was interesting how that pretty quickly stamped on any behaviour but became a collaborative endeavour. Yeah. Okay so I think the point there was that there's a game I'm afraid I didn't catch the name of it that has a really really strong community resulting from the developers like playing a big hand in like kind of nurturing that community and like hanging out on Reddit and the forums and things like that. And yeah like that's I think that's a really important thing as well like the game I mentioned earlier and Rocket League but even though that's a massively popular game the developers for that are known to hang out on like the Rocket League Reddit and they talk to people and they take community suggestions and yeah that's like that's a really positive thing both for like fostering the or like supporting the studio as well as kind of pushing back back to the player base yeah. Thanks. Is there a possibility that in games it's not just the gameplay but also the kind of feel of the game that is kind of affecting the community as well because like in things like overwatching and stuff it's all very stylised and like you see loads of like pictures on the internet even if you're not looking for overwatch you'll come across people playing it and drawing about it and talking about it and things like Splatoon for the Nintendo Wii U did quite well and seemed to have a good community because it was encouraging people to draw things so do you think that's another route to improve this? Yeah so the question was like do I think that games that have I guess interesting or like kind of more stylised aesthetics and like the look and feel of the game rather than the mechanical gameplay does that in and of itself like also foster a good community and why? And yeah I think that's like that certainly could well be true like Call of Duty and Halo Xbox Kids are a meme for a reason right? So yeah certainly I think like if you can take people out of like this grim gritty reality that they live in then yeah that probably like does make them feel more inclined to be nice than each other because the real world sucks and yeah everything should be bright and colourful that would be sweet Yeah thank you Sorry I can't quite hear you Sorry does this work? I grew up on subscription based games and found that the abuse in those was a lot less and it was almost as if players had a vested interest to behave and certainly I noticed when Dota came in in the free-to-play model that there was almost this you know there wasn't any concern from a player because you could go in cause absolute mayhem and if you got banned you could just go and create another count do you find that there is a more sort of toxicity for free-to-play games versus subscription based? Yeah so the question was do we find that there's more toxicity in free-to-play games because there's no barrier obviously no monetary barrier to entry than there is in like subscription games and yeah I hadn't actually thought of that that's probably a fairly interesting point I think that like yes is kind of my short answer but because you see this is reflected in the real world as well like outside of just games so like forums that anyone can hang out and talk on that you know you don't have to know people to get in or you don't have to pay money to get in or whatever they just get more toxic like YouTube comments yeah anywhere like where there's no kind of strong sense of community I think is likely to be more toxic and yeah I think adding a barrier to entry does mean that you're more likely to have a strong sense of community because by definition there will probably be less people involved in it whether or not the act of like actually handing money over to someone gets you more invested in making something nice for you and the people around you I couldn't say I think it's probably true but yeah I don't know thanks I'm afraid we don't have time for any more questions right now but the next session is also on computer games so if you're into computer games hang around for a conference session so yeah last thing to say thank you very much for talking to us thank you