Game Economist Cast
What does the new wave of open economies mean for monetization? Will negative externalities overcome cosmetics economies in the long run? What exactly does a game economist do?
Game Economist Cast is a roundtable discussion of the latest developments in mobile, HD, and crypto games through a bunch of people figuring it out using the economic tool kit.
Game Economist Cast
E34: A Theory of Optimal Economic Balance
Is game balance bullshit?
The crew goes toe-to-toe debating Sirland’s Don't Use Math in Balancing Games. Chris emerges from his Roblox hibernation, Eric tells us Street Fighter is more accessible than platform fighters, and Phil goes bonanza for All in Hole.
Chris is dying, I figured it'd be better to put my corpse in front of the camera than not show up.
Speaker 2:His last deathbed wish was to be on one more podcast episode. Let's start with utility.
Speaker 3:I don't understand what it even means. Everybody has some kind of utils in their head that they're calibrating.
Speaker 1:There's hardly anything that hasn't been used for money.
Speaker 4:In fact, there may be a fundamental problem in modeling that.
Speaker 4:We, here we go. The boys are back in town. We were supposed to keep up our normal schedule. There was a mishap with a guest who were trying to get back on before the christmas season. Let's see if that happens. If not, early 2025, where the guests will be coming, they'll be rolling. They're still more in the pipeline, so it's just the normal crew today getting back into season.
Speaker 4:We have some amazing things to talk about. As always, I'm going to be talking about battle pass trends. I titled it. What the fuck are these three battle pass things? A little bit of clickbait, but there's three things in battle pass we should definitely talk about, revisit, so to speak. And then we also have a wonderful david sterling video don't use math in balancing games which is something that he released on youtube. We'll have a link to it in the episode description if anyone wants to see it or watch it. David sterling, I think, is a very interesting human being. We'll talk about him. I think he's certainly a youtube channel that's worth subscribing to. He's got the hot takes. Let's see if we can out hot take him, inferno take, burning take, warm take. Before we do, let's talk about chris, who's barely alive. Let's talk about what chris has been sipping herbal tea got a selection of good things on sale stranger just for an fyi.
Speaker 1:Even though this is the symbol, this is just tea in a bailey's cup okay, non-alcoholic. Right now I'm trying to nurse myself back to health. I've been a little bit sick but as a result, I've been playing mobile games and I've been reading a book. I've been playing a lot of Roblox games. So finally, on your phone yeah, on my phone, because the PC and console experiences are not great. So when you play Roblox on the console, this sub menu that's layered on it's almost a meta menu layered on top of the actual game that allows you to interact with the menu items. So maybe you want to go to your roblox menu or you want to exit the game. You have to press select in order to activate that menu and then you can.
Speaker 1:It's like very confusing and it's not communicated. I played on console for a little bit and I was like this is not great. And then I played on my phone and I was like this is not great. And then I played on my phone and I was like, oh, this is what Roblox was meant to, how Roblox was meant to be played. I've actually been enjoying it. It's a really interesting experience. It feels like I'm hanging out with kids that are younger. There's a culture that I don't understand there, right Like the names of some of the games.
Speaker 4:What's up kids?
Speaker 1:Well, I'm, well, I'm trying. Yeah, I have a cool avatar. They've got a lot of free gear and items and stuff. I anyway, uh, I played adopt me, uh, a little bit. I played the fashion game, which just like it's.
Speaker 1:I just don't understand it. There's just some games that I don't understand, even though there's some of the most popular games in the ecosystem um, my favorite are like the mini games, so they have a lot of tower defense games, so I found myself playing some of the tower defense games. Then I asked myself am I better off playing this tower defense game inside of Roblox or would I be better off with another app on my phone, the idea of having this central place where all my stuff is? I have one symbol on my phone, the Roblox symbol or app, instead of 30, 40, 50 different uh apps. I believe they do streaming as well, which is pretty impressive. Game video streaming or no, no, just I believe all the games are streamed. Otherwise there would be like a million terabytes of data that they'd have to download because you can jump into any experience on your phone. I'm not really sure how they do it.
Speaker 1:I know that they have a huge cloud gaming department, though, within the company I think it's all marvelous, so kind of yeah, it goes in the face of that whole like anti-cloud gaming uh narrative, which I think is probably true when it comes to xboxes and playstations uh, it's not as true when it comes to roblox. These experiences are seamless. They're cool like they're, they're it's. It's seamless to play these games. You're able to jump in to so many games even quicker than we talk about the Xbox store, which is pretty easy to go find a free game, download it, play it. Roblox is even faster, which kind of gets to, in my opinion, the not so great parts. It's really hard to sift through the millions and millions of games to find a game that I like and some games, when I get into them, they're just like monetized to the, to the some inappropriate word, they're completely destroyed. With monetization there's, you can't click through the monetization to just get to the game.
Speaker 1:I think that's mostly a critique of individual experiences. I think as a new user, I played Roblox in the past, but this was like years and years ago. So I created a new account. So it's cool to go and experience it as a new player again. I think the first time user experience can be a bit overwhelming. So those are my initial thoughts. I'm enjoying a few different experiences. I'm opening it up and playing around with it. Yeah, I've been doing Roblox. I've also been reading a Brandon Sanderson book. That has gotten interesting.
Speaker 4:So, on the discovery, this is a really this is a really interesting point. So there's this company called the high pipe, who I actually worked with and I think they're really brilliant human beings where they were making tick tock for games, in the sense that you could look at a game as though we're tick tock video and then you can swipe upward and you can see a new game which makes discovery like super fast and what. One of the things I've been trying to figure out is like how you it's a matching market like you want to pair people to the highest valued game that they have, and so like how can you do that the quickest? And so part of it is you can do revealed preferences where you just have to play games, but that's actually pretty time intensive. Do you think there's a way where you can basically induce this by doing like a survey saying okay, are you into racing games?
Speaker 4:Like netflix kind of, has you rate sample movies for sure, starter, spotify why don't they do that? Why don't they say pick a genre, tell me what? Like an artist's like title, does title?
Speaker 2:but a lot of things like even when you sign up for twitter, you have to pick some topics. Yep, yeah, blue sky threads do this.
Speaker 1:Yeah, that's actually. That was actually one of my main kind of uh recommendations, as I was mental recommendations as I'm looking through this app. I'm like us. Just a really short survey at the very beginning would help so much Because I, as a player, don't care about dressing my character up. I'm not really motivated by those types of games. I'm more interested in story games, fun tower defense types of games. Yeah, I totally agree, phil. I think it would go a super long way.
Speaker 4:What kind of questions would you ask that you think would reduce search costs or do this process the quickest? Like, how do you think there's a way we could do? Is there a way we could do like economic incentive? I don't know is. Is there a way we could be economists and approach this problem?
Speaker 1:there's a couple of issues with the, the incentive structure, and not issues with roblox, but issues with creating an incentive structure to promote games. Because the games that get, I assume, pushed by the algorithm, are these games that have lots of views and lots of high ratings and I think we've talked about this on the cast before rating inflation, where you go out and you say, hey, I'll give you 10 in-game credits if you give my game a 10 out of 10. And I'll find some games will be a three-star experience and I think that's a five-star experience, and vice versa, you'll see games that are five-star experience, that feel like a two-star experience. So that's the first part is how do you accurately rank these games in terms of the quality of experience? There are a hundred different dress-up games, there are tens of thousands of dress-up games, tens of thousands of racing games, tens of thousands of tower defense games. How do you determine which one is the best intro game? So that's on the, that's not even on the discovery side, that's, if you're creating your, your sorting algorithm. That's like setting the matrix up so you can actually serve this matrix to the customer. So I think that's probably one of the challenges is figuring out okay. A what are the different types of games we have here? And B what are the best games within each of those genres?
Speaker 1:I assume one of the most challenging things about working on Roblox is that there's a continuum. Everything is a continuum. Not everything fits into these specific categories. Even the tower defense games, you still have an NPC that's walking around the environment placing the guys. So it's always going to be this three-dimensional experience, even if it's effectively a 2D game. How do you rank everything in a matrix? And then how do you serve that up?
Speaker 1:They have two primary ways. They have search and then they have recommendations. So you can go and you can search. Here's the item, here are the things that I want to, that I'm interested in, and they have some really cool. They have some excellent language translation services as well, so you can like, you can Google, you can search in Portuguese and you're going to get all the games that are relevant to you, even if they're not originally in Portuguese. And then there's the algorithm.
Speaker 1:I definitely think I wish there was more kind of oh, I'm interested in shooter games, I'm interested in racing games and being able to create those categories in my For you page, not necessarily because it's easy enough to go to the search menu and say I want to go play a racing game with dragons, and you're going to find a game that has dragons in race cars. But I want that to be on my main menu and I just want, I want those types of games to be served to me and they do that sort of based on your history. I think like some customization. There would be amazing Intro survey what kind of games are you interested in? And then what kind of genres do you want to follow?
Speaker 2:Let me add a point to that, which is maybe the point of Roblox isn't actually to serve you the best games, but it's to keep you and your friends together and to that end, the mashing problem is not actually independent between the search, the two players, the players trying to find games. Maybe they just want to funnel. Everyone's playing adopt me. Maybe there's a better pet game, but all your friends are playing adopt me. If we encourage you to play adopt me, you'll talk to your friends more and the whole platform will be so.
Speaker 1:Yeah, maybe they're not even trying to serve the best games and that's that's why I say I feel out of place. It's because I don't have any friends on the platform and friends are like. That's half the fun of it. It's for, especially for the people who are super into it. It's like facebook for them, it's like instagram for them. They go there to hang out with their friends and there's a lot of features that allow you to jump to your friends experiences. I hop on and my buddy is playing adoptopt Me playing Fruit Blocks. Yeah, that was the one I was playing. I just jumped to that experience. I'm not even. I don't care, I'm not trying to find a racing game. I do think for this more traditional gamer, somebody like me, discovery is a little bit more important. Maybe it's not as important for these guys who know exactly what they're looking for.
Speaker 4:They just don't do a lot with friends. I have recommended for you.
Speaker 2:But to your point, eric shouldn't be like here's what my friends are playing or watching, because everyone's moving in social groups. Isn't that what chris is saying, that there actually are a bunch of friend features. You just have no friends, or?
Speaker 1:do you have friends on roblox? I don't have friends on roblox. There's a button that allows you to warp to your friends games or the experience that your friends are in. That's what meant. So if you know the 10 games that you like to play and the 10 games that your friends like to play, it's perfectly fine. For a first time user coming in who's not familiar with the platform. I think that a more curated experience would be better. Answering a survey, having categorization of the games where I can say I want to just play racing games. They do this, but it's very algorithmic. It's not as curated for the specific interest of the player as it could be. Treat me like I am literally a blank canvas, even though I'm not. I have these preferences as a player. So it really is this zero information environment, and I'm sure that as I keep playing, it will become better and better. I just think there's a way to jump start that no, I'm with you.
Speaker 4:With you, that makes a lot of sense but it's super interesting.
Speaker 1:I think roblox is probably the most interesting experience that I've. That I've had. It just makes you go. What is going on? I play a lot of games and this is the one that was like holy crap, crap, don't understand this, and maybe that curiosity is what is keeping me going back.
Speaker 4:And it's wild. When I was looking at the quarterly stats for the quarterly earnings, I did the math Half of US Canadian kids and American kids will play Roblox each day. Dude, it's Divided by under 30. Yeah, yeah, I can't even believe it.
Speaker 1:And I think that's probably their biggest challenge is like how do they get outside of that group? And I know that there's. According to public releases, their average age is increasing, so that's cool. How do they break into those other demographics?
Speaker 4:I think there's another problem here, which is that because the Roblox fee is flat, because it's 80%, yes, ladies and gentlemen, it's 80% of every dollar spent goes to Roblox. Not only that includes Apple's fee, just to be clear. So it's really 50% them 30% them and cloud services.
Speaker 4:So they do a shit ton, but the ultimate reality is when someone's taking 80% of a dollar, that means that you can only scale so much, and so these games can own, like the top game on roblox maybe does 80 million a year in take-home pay for the developers, and if that's the top game, remember it's power loss. So everything goes down from there and I think they've really limited the size of the games that can grow on the pop, because, because a game can only grow to the size of the platform, right, that's, it's TAM.
Speaker 2:I mean 50% of all North American kids is huge.
Speaker 4:I wonder if they'll go to the inverse marginal tax rate of Steam, where what happens is that remember, steam actually reduces your tax rate, your marginal tax bracket, as you earn more money. I think if you earn more than a million dollars on Steam, I think it's 25% instead of 30%. I forget the exact brackets, but I would argue that they should do something like that, because to me the threat is to your point, chris, like if I'm going on to roblox and I'm only going on to play adopt me, when I click that button I don't think about roblox, I think about adopt me. And so if that's the case, if that's where the brand loyalty is and remember there's also a lot of spinoffs, like adopt me has toys and toy stores, like they're starting to build a real sub brands where it's not about buying roblox, that's about buying that shit then why doesn't adopt me just spin off into its own mobile game? Fuck, fuck roblox. Make adopt me in unity and publish as a separate app.
Speaker 4:Well, because I think that's development is yeah, I guess that's fair.
Speaker 1:I do think it's really misleading to use the 80% figure because, like you said, you're talking about Apple's cut. You're talking about like you're talking about infrastructure as well, all of the development tools. It's basically that is 20% profit. That's their profit margin. That's like how much they're getting after the fact or, I guess, net income, because they haven't paid their payroll taxes yet or their payroll yet. But so I'm not sure that it's fair to say that there's a huge deal to be had.
Speaker 1:I really the point about you're only as big as the platform that you're on. What happens when you grow outside and this is a very important point Roblox is getting better and better every single day. But there is very specific type of game you can make on Roblox. Right, there's only so much fidelity that you can add. There are some really cool experiences with high fidelity. There's Phantom Forces is like a first person shooter, csgo type of game and that'll get better and better. But not going to see a AAA studio developing on Roblox anytime soon, and I actually think that's a good thing, because I think that games, I think game budgets have just gotten out of control. I like the tax bracket idea you had. I think that's really interesting.
Speaker 4:I think there has to be competitive pressure. That's why Steam did it right. Is they had so much competitive pressure from other platforms? I think Fortnite really needs to threaten Roblox for them to consider it.
Speaker 2:They were throwing money at developers with their whole engagement share program, throwing money at developers with their whole like engagement share program, which they rolled back because apparently it was losing them too much money.
Speaker 4:The problem is fortnight's not growing. Fortnight's actually been shrinking. It depends on what you set your index to, but fortnight is not 50 of us and canadian kids under the age of 13. It's much, much smaller than that. It's 300 mau 380 million mau I thought was the last total for Roblox which is more than the size of the United States and it's more than Xbox, PlayStation and Steam combined. Just for context, for listeners.
Speaker 1:Well, I was going to say with respect to the UEFN, that's their developer environment. I think Roblox has a huge first mover advantage, so they're the incumbent. It's going to be hard to house them. It's the same issue that they're having with the egs, that the game store steam is the first mover there, so they're trying to come in with this product. That's effectively the same thing, maybe better in a few ways, maybe not better in a few ways, and it's just really difficult to dislodge their biggest disadvantage that the games that they develop on the roblox platform are capped at the size of Roblox. So, while it does have this kind of ceiling effect that you're talking about, like you can only grow as big as the platform. That is simultaneously a problem and one of its biggest advantages.
Speaker 1:Why am I developing on Roblox and not UEFN? Why am I not just doing it myself as an indie studio? Because I want access to those hundreds of millions of players. I want access to every single kid or every single young adult in the United States. I want access to the tools. I want access to that platform, that distribution platform. That's something that almost nobody else can provide. Maybe you do have this like top player. Your biggest games are eventually going to move away, which I don't think that they have seen that. By the way, i't think there's any been any major players that have left the roblox platform. I could be wrong about that, but all the big ones that I know of are still around. They've been around for 10 years. So, yeah, I think that's probably a gift. It's both a gift and a curse, and they're just going to keep getting bigger, so it becomes less of a curse as time goes on.
Speaker 2:I've been playing Rivals of Aether too. It's basically Smash Bros. It's like Smash Bros for Smash Bros players. If you've ever heard of Project M, it was like a brawl mod back in the day where people would hack their Wii's and modify it and add all sorts of characters and new mechanics. I would say it's like the spiritual successor to that. But yeah, it's the first platform fighter that actually has drawn smash players like actually drawn them out, not just paid a bunch of marketing money to them and they all quit after a month. Um, because it's actually good. It's designed by, like, people who have deep experience in the genre. Um, and it feels great. I do think it's market size is limited to people who are ready smash enthusiasts. They're not really pulling in new players or anything. But yeah, game's fun.
Speaker 4:Don't have too much to say about it this looks very interesting, eric, this is on switch and it looks like steam it's on steam.
Speaker 2:Yeah, the previous one was on consoles too, but this one.
Speaker 4:Okay, so this is rivals of aether 2.
Speaker 2:Yeah, is it of mtx yeah, it's, uh, their shop actually. I don't know if I've seen this structure before. So it's box price, a quote, quote, unquote. It's like a pseudo live service game. They're not pushing like all the characters and stuff. Game content is released for free, but there's a soft currency gold and a hard currency bucks and here's the interesting thing. So all the items in the store you can buy both ways, but the stuff purchased with bucks is always available a la carte, and then they have a rotating store that offers those same items for coins at a fixed exchange rate. So it's basically the same price, but it's limited window availability Are hard currency, coins are soft yeah.
Speaker 4:So essentially, you can buy anything you want with soft currency, but only through the FOMO store. So essentially, if you retain long enough, you'll see everything priced in soft currency Exactly.
Speaker 2:Interesting.
Speaker 4:So you basically are short-cutting the rotating shop. If you spend hard currency, you're forgoing the period, exactly, yeah, so the longer the rotation windows, the longer the wait, technically, the more beneficial it would be to purchase hard currency. Yeah, and the rotation window is pretty limited. So if you're waiting for a specific, longer the wait, technically, the more beneficial it would be to purchase hard currency, yeah, and the like.
Speaker 2:The rotation window is pretty limited, so, like, at any given, if you're waiting for a specific item, it could be months jesus, did they make this available, this information public, like when a routine is going to happen, or is it random? I'm not sure.
Speaker 4:I haven't looked too much into it let's put it this way do you think it would be beneficial? Should they?
Speaker 2:they definitely should not, because then you create that like sales training like you're. I'm just gonna wait for this thing to show up why not?
Speaker 4:why not do one? Why not be explicit about that? It's basically like a timer. It's basically a timer like why can't you just tell me if a timer is big or small?
Speaker 2:because when the time is uncertain, I think people are more likely to want to shortcut it. That's like a risk conversion thing that that's fair Okay. And you still get most of, I think, the PR bump of what free players can buy.
Speaker 4:But it's not. I guess we can go Frank Knight here. I guess it is calculable, though, right? It's not uncertainty, it's not even risk, right, because it's not random if something shows up in the shop.
Speaker 1:It's based on a predetermined schedule, schedule unclear. Is it possible that it never shows up again, or is it?
Speaker 2:just like they literally randomly select. Yeah, I don't even know if it's. It might even be personalized and random, I have no idea. But the thing that really stood out to me is interesting, though, is a lot of games with this soft and hard currency. They'll have different items that have different ratios, or some items will be only available for one and others will only be available for the other, and there's probably upside to that. But there's like a lot of adjustment and calibration involved, and they were just like whatever it's a small indie team, there's whatever it's like every. There's a fixed exchange rate. Everything's available in both stores, just in one it's limited window and the other is all a card. They don't have to have some project manager's job. Who's is to like design, which bundles are set at which price, and shit like that how do fighting community guys feel about micro turn tech?
Speaker 1:because you think about yeah smash bros. It was always just like hey, you earn the characters over time as you play the game.
Speaker 2:No, there's dlc yeah, the new one. There's dlc. They didn't monetize it very aggressive. I think it's 15 bucks for five characters and there were two sets. Generally speaking, if the game is good, people don't care. But also fighting games, there's not as much power scaling, so there's a cap, like you buy a character and then that's it, and so as long as you buy the three strongest characters that you want to play, you're good.
Speaker 4:There's less character switchings. Yeah, Can you even make this live service game? Then it's only the Ubisoft one right Ubisoft Fight.
Speaker 2:I can't even think of what you're talking about. Assassin's Creed, it's a fighting game. Brawlhalla Ubisoft owns Brawlhalla. Yeah, that game is slowly dying. Multiverses was, I would say, very similar to Brawlhalla.
Speaker 4:Is this just impossible to do, To build a live service fighting game? No, I don't think so not at all.
Speaker 2:I think a lot of games are basically functionally there and if you consider COD to be a live service game through its, like, regular releases, there's a lot of these fighting game franchises that every year, every other year there's a new game. People pay the box price. There's also DLC in the game. Yeah, it's not super lucrative, the market's not super big, but I don't think it's a far cry to say that they haven't. They're not doing lightsabers.
Speaker 4:It's just that it's just not as lucrative, so it doesn't make headlines, so you do think it's the case that they can just stick to the strategy of charging for new characters and cosmetics, and that's enough. Because what you just suggested, though, is, once I find my mains, I don't really want to buy new characters, and if it's not a lot of character switching, I'm basically tap out once I purchased the three best cosmetics from my mains.
Speaker 2:Yeah, so what they tend to do is the new DLC characters often are strong or, at the very least, very interesting and unique. Cosmetics is an endless. There's always some new fashion that looks cooler, so you can always keep hitting that button. I have a question. It's just a matter of scale because of, like the fixed content, development costs and variable revenue.
Speaker 1:Is it wrong to rebalance the game in order to sell characters?
Speaker 2:I think most Western players would say yes. I think most live service games justify it by saying they're shaking up the box to keep the gameplay and the meta interesting and the designers don't want to talk about the implicit effects on revenue. I think the implicit effects on revenue aren't that dramatic, but I think if you can get people to play multiple characters, main multiple characters, they'll spend more, and so if you can design the game whether it's you're specifically trying to nerf and buff people to shift them if you set this expectation this is a game where you should play multiple characters you'll have a wider monetization base. Yeah, rival Data 2 play multiple characters, you'll have a wider monetization base. Yeah, rival dealer 2 check it out, I'm gonna show this game. This is the game I've. I haven't fell in love with the game for until for a while, until this.
Speaker 1:So about to say. I'm still waiting for the day that eric describes the fighting game that makes me want to go out and play fighting games you just honestly, you just need a couple of friends who are into it. That's the main driver I need a switch is what I need.
Speaker 4:So all these games are on nintendo the platform fighters are so much more accessible. Like I will give that to eric, it's like the controls are wait.
Speaker 2:I wrote a whole blog post about how they're actually not accessible and uh multiverses made a huge investment mistake because they believed them to be accessible when they're not like when I think of like a game like street fighter.
Speaker 4:The thing that's always prevented me from getting into it is that there's really deep combo lists and pretty, I would argue I don't want to say convoluted, that's not the right term but it's like pretty deep mechanics that require an understanding of. Oh, like this bar charges up every time like I get a parry or something of that nature, so what? But when I play like super smash brothers, there's really only one or two layers of button combos like move dash. Dash with stick plus A, like analog stick plus button is basically half the game. That feels a lot more accessible to me.
Speaker 2:I'll say I think it feels more accessible because it's presented more accessibly and probably the people you're playing with are not as competitive. But if you go into the mechanics, platform fighters are insanely complicated. Like in Street Fighter, to move forward, you walk forward or walk back and maybe you can double tap the direction to dash. In Smash you've got walk, run, turn around, wave, dash. You've got shield state and outside of shield, in shield state, there's a restricted set of options you have, and don't even forget about the ledge play knockdowns, get up attacks. It's insane. The amount of movement, fidelity and state-based actions that you have is incredibly complex. And there's all these videos of fighting game people trying to play Rivals of Aether and just getting super confused and complaining about the poor onboarding. It made me reflect that actually there's this common narrative that platform fighters are accessible, but I think that's just. Smash has broad appeal because it's got Mario and Pikachu.
Speaker 2:But if you dig into the actual competitive mechanics and like the complexity of the game. There's a ton under the hood and if you've ever played against someone who's way better than you, you just get washed and it feels awful right. You're just like this guy's, just moving faster than me and doing stuff that is impossible for me.
Speaker 4:I don't even know what's happening. Oh, I've definitely had that frustration. I've had that frustration in fighting games too, where I feel like I get absolutely wrecked. But isn't that just speak to the cliche of easy to pick up easy to pick up and learn, hard to master?
Speaker 2:yeah, but what makes smash easy to pick up? Because if you play street fighter against another noob who's only hitting single buttons and ignoring, I would say not having to understand combo lists.
Speaker 4:I've never had to understand a combo list in Super Smash Bros, and that is the main barrier to me.
Speaker 2:You don't need that to play Street Fighter either.
Speaker 4:I'll tell you.
Speaker 2:You cannot play Smash competitively if you don't understand combos. But you can still play casually, right and the same goes for Street Fighter.
Speaker 4:Are the combos in Super Smash Bros explicit? Because in Street Fighter, like I need to do x, y plus, like I have to do a sequence of moves that can be like six. It's like notes and it's like a song. It's a song that has all these notes I gotta hit, whereas in Super Smash Brothers it feels they're separate notes that are strung together and that just ends up being a combo. But that's more organic than the inorganic ones in Street Fighter, where it's like literally to even activate this move, I'm going to hit 30 buttons in 15 seconds in a particular sequence, whereas this I can do whatever. It just happens to be the case that in Super Smash Bros these things fit together.
Speaker 2:That is fair. The combo trees are more free form in platform fighters. But even in Street Fighter you can just walk up to someone and just hit light punch and that's a combo, right? I don't know. I think there's a lot of and, frankly, you can play Street Fighter fine if you just space and do single hits, you and a friend can just go at it. You don't actually have to learn the combos to have a good time as long as your opponent is on the same level as you. And I think that latter point is the key is that there's more casual Smash players, so it's easier to find someone at your level. It's not that the game is intrinsically more easier to pick up or anything.
Speaker 4:It's about flow state right, like getting into, like these fighters. It's about once I understand the combos, I can then do all these things which give me this. It's that flow state of optimal difficulty, optimal success. Is that fair? Isn't it harder to get flow state in a Street Fighter than the Super Smash Brothers?
Speaker 1:I find Street Fighter easier to play than Smash Bros Interesting.
Speaker 2:Here's another point. In Smash Bros, you can kill yourself. In fact, top players kill themselves regularly, right? What other game can you think of where it's like a normal thing for an expert player to kill themselves?
Speaker 4:That's a really good point. You can kill yourself in Dark Souls too. I would say that the fail state is also super interesting, Like the fail state in Super Smash Bros is that you become more elastic, which is very unique. I'm very surprised very few games have done this. I worked on one other game called Rocket Arena rest in peace that used this, which was a third person shooter. Interestingly enough, where it used this kind of like elasticity, more damage you get, the more likely you are to get booted off the map quicker. Weight your weight goes down, but it's interesting that all the other fighters will use health and you just die when your health goes to zero yeah, I think it's great because it changes the game state a lot.
Speaker 2:But again, here's another layer of complexity where the combos that work at zero percent are different than the ones that work at 50 percent. Right, and you have to know that if you want to be able to convert. Probably.
Speaker 4:But and anyway right in all two d fighters, like there isn't variability in health between. Yeah it's just a bar that goes to zero but what will change is the damage per particular hit between a character like a punch might do. Okay, that's a very interesting thing that shows in the whole constant, because in smash your elasticity changes or your weight changes, I man, you make me actually want to play this. Eric, I hate you.
Speaker 2:I would say right now the new player on Boeing is pretty bad, but I know they're trying to put in a tutorial in a single player campaign.
Speaker 1:This has actually made me want to go play Dragon Ball Z on Nintendo or Game Boy Advanced.
Speaker 4:If anyone wants to play with me. I really need someone to play Sparkling with me. I paid $80 and I'm just sitting in my Steam library. I haven't even opened it $80?. Yeah, I bought the Ultra Premium Mega Edition.
Speaker 1:You bought the DLC upsell. You could buy a Game Boy Advance and Dragon Ball Z, goku's, whatever for probably $40. You can get it for free Game Boy Advance.
Speaker 2:Just download an emulator. Yeah.
Speaker 4:Oh, eric, I could see you being a contributor to Dolphin back in the day. Did everyone run smash on dolphin, or did? I guess everyone ran it locally?
Speaker 2:yeah, melee is all on dolphin.
Speaker 4:The, the online play, is a modded version of dolphin on my side I've been playing this game called all in whole, which I'm fucking obsessed with. If you guys have heard about this just came out globally. It is similar to a game that was on steam originally and then has been translated to hypercasual. You could almost call it reverse Katamari Damacy, if you remember Katamari Damacy, where you would roll around and you would collect objects and your ball would get bigger and bigger and then you'd keep collecting more things. This is similar in the sense that you have this dark hole and it sucks in objects and the more objects that get sucked in, the larger the hole gets and and you keep sucking more and more objects in. And it started as a hyper casual game and it was so enormously successful I was actually working at Homa as a consultant at the time they came out with the original version called hole and filler their version.
Speaker 4:It was just marvelous seeing how many what we call attempts per success there were. So it take about. On average, I think it was around 11 attempts per success to beat a level, which is mind blowing. 11 times you would fail before you win. Which per success to beat a level which is mind blowing 11 times you would fail before you win, which is a brutal difficulty curve, and it was set up that way because what would happen is, when you lose, you would get gold along the way and you can use gold to basically vertically progress yourself. So you would start with either a larger hole or maybe the time would be extended.
Speaker 4:So the idea is that you would grind through each level, but of course it's a miserable experience because you're failing 11 times. So what they've done in this version it's been two years since the last one, by the way is that they've added the the match three template to it, which is something I've done in another project and earlier cast. We're talking about like why match three is the best game economy of all time and I still believe that because it can be layered in as a template to so many different game modes, and this one is no different. They basically take that core gameplay and, instead of having that grind element I talked about, what happens is that you just need to collect x objects in a particular world and then you get to the next world.
Speaker 2:Yes, that's three template. What do you mean?
Speaker 4:So the batch three template I call it the saga template is that first of all there's linear levels level one, level two, level three, level four and there's a fail state at the end of level and I can usually spend either extra moves or I can extend the time, or if I'm playing another game that's like match 3d I can extend the backpack because there you're collecting objects and you only have so many objects and when you get three objects that are the same on your backpack, then they combine together and so if you have too many objects without a match, that's another fail state that can. So you have this idea of a fail state and then you spend to extend it. But then what they do is they add a lot of pressure to that fail state. So what you want to do is you want to win in your first attempt consecutively, because there's a lot of benefits to that. So they layer on things like streaks, where if I don't win on my first attempt, then my streak is broken and really broken, and really I want my streak to continue because I would have to play, I would have to build up the streak again by by winning consecutively, and then they'll add even more pressure to that, where they'll have an event that's called episode race that king came up with, where I'm competing with other people in a pod and the first person to have five wins in their first attempt wins, wins the episode race and wins currency. So they they've created this really incredible template that I think you can graft onto a lot of different core gameplay. So in Coffee Golf we did the same thing, where you had a certain par and as long as you came in under par, then your streak would continue. And then we had a bunch of events that were based on you getting a win in your first attempt and, of course, if you were above par or you ran out of shots excuse me you could buy more shots Right. So that's been the match three template that's been grafted on to all in whole as well, and it's worked really well. They got all the events, they got the streaks, they got so many things, and not only that, they innovate on the fail state.
Speaker 4:So one of the things that's really challenging is that when you have match three, your fail state is based on moves.
Speaker 4:You have X number of moves to reach some sort of designer driven goal, whereas match 3D and all in whole have times and a goal, so basically I have to suck in, let's say, 30 apples within two minutes and 30 seconds, or else I lose, and so that's generally. It's a tough thing to do, because most players hate time pressure. That's a big change from match is not having this time pressure, and so one of their big innovations that I fell in love with is they added bombs on map, so if I accidentally suck in the bomb, I die right there and then and that's how they drive a lot of the fail states is adding more bombs to levels, which I think is pretty brilliant and so, when the fail state triggers, it offers you the ability to spend to what revive and continue playing exactly because I want to win on my first attempt because I want to win the episode race or I want to keep my streak going what's the innovation here compared?
Speaker 1:so I I think I remember you recommending the whole game and I played it and I was like this is fine, it's good. What's the innovation here? Why would you play like all in whole instead of whole in one or whatever?
Speaker 4:because it's got the saga template. It's so much more enjoyable. You don't have to grind for 11 attempts per single level it feels like there's much more momentum there's much more to compete for.
Speaker 1:There's, yeah, there's far more levels.
Speaker 4:There's far more levels. You only have really. You only need one or two attempts to beat a level, rather than 11. They don't have this vertical progression element. You don't have 11 fails before success. It feels much more fluid and much more meaningful to win, or to win on your first attempt, which adds tension especially with the bomb innovation.
Speaker 1:That's interesting. So was the losing a real driver of retention? Or I guess was there a negative relationship Because I found myself just wanting to suck things. I was still getting progression, though out-of-round progression.
Speaker 4:Yes yes because you would get coins. So they wanted you to fail a bunch, because each time you failed you would collect coins, you'd upgrade yourself and then it'd be more likely you could win.
Speaker 2:That's still pretty brutal, because you're still failing, you're still getting coins, but still quote you lose like you're not going to the next level yeah, my my experience with it was I realized it was all about you just can't do it until you buy enough upgrades and I was like, oh, the level is just try to level 10 times. And I don't want to do this anymore.
Speaker 4:Yes, yes, and that's brutal, and this is a much smoother experience.
Speaker 2:There was no feeling of if I play right, I can win.
Speaker 1:Yeah, that's true, that's a good point.
Speaker 4:There's no amount of mastery it feels.
Speaker 1:Speaking of whole games, roblox. I can't see. It doesn't matter, just imagine a bunch of holes.
Speaker 4:There's a bunch of fill-the-hole games. Does it fill the hole in your heart for these games?
Speaker 1:Be a hole Hole simulator. Become a black hole.
Speaker 2:Yeah, you get the picture. The more things fall in the hole, the bigger the hole gets, which is the opposite of how holes work. I guess that is how black holes work.
Speaker 1:I haven't really thought about that. It is satisfying to have a giant hole that everything just falls into it.
Speaker 4:I agree. I'm surprised we haven't seen someone take a shot at free-to-play Katamari Damacy. Basically, take the same concept, but just do Katamari Damacy Instead of it going into the hole and getting bigger. You could do everything the same.
Speaker 2:Just make it katamari style. Yeah, just the ball gets bigger. I would license that ip and try to pull this off. It might be visually less clear because the ball obstructs your view of what's going on.
Speaker 4:Yeah, and you've got to zoom, you start zooming out too. There's a lot of there's a lot of technical problems that make the sucking thing better than the growing I just want to shout out donut county as the, the indie game that actually produced this concept.
Speaker 2:Yes, that was it.
Speaker 4:Thank you, thank you, eric. Thank you, eric. And I would say this is another thing hyper casual studios are doing now, ever since Survivorio stole Vampire Survivors, like I know, a team at Voodoo has this where they literally just go on Steam and they play Steam games and they try to make hyper casual versions of it. Remember Backpack Battles? Voodoo has made a hyper casual version called bag fight. You can play and I hope someone does this for bellatro like where the fuck is my free to play? Bellatro? It's social casino man. Like why isn't zynga poker doing this shit? Zynga, where are you? Like pivot? Pivot to this. I want a free to play version. Is I want to spend money in bellatro. That's the only thing I don't like about bellatro is I want more pvp and I want to more money.
Speaker 1:Did Bellatro just come out on mobile?
Speaker 2:I assume it's the same game One step closer.
Speaker 1:Yeah, it's the same. It's also probably still cost money.
Speaker 2:I think it's super cool when games are able to take meta progression elements and import them between games. Like you mentioned, the match three template, the schema where oh the few, but like when you have a fail state, the option to buy out of it lots of levels along a track, the streaks and all that it's cool to see how modular some of these elements are.
Speaker 4:Yeah, no, I agree, I think this stuff is really cool. I feel like in mobile the paradigms are like much sharper just because the capital costs to start a game are much lower and so basically you reach consensus really quickly because N basically just implodes. It's a much more brutal process because you don't have many barriers to entry, which I think just makes the evolution process so much quicker, so much more brutal, like sink or swim baby.
Speaker 1:We could probably just put a different name, and we're almost always all playing the same stuff Phil's playing a free-to-play game, Eric's playing a fighting game and I'm playing some random shit, tumeric, tumic. Eric's playing a fighting game and I'm playing some random shit, some turmeric, tumic, turmeric. That'll be tunic, will I feel?
Speaker 4:like chris has the widest range. Yeah, I would agree with that, to be honest, that's because I think he does honestly that's because I like to try what people recommend to me.
Speaker 1:I'm very much I. My problem is I know that I have a very specific type of game that I like and I force myself to experiment and participate in that multi-armed bandit. So I don't just end up in the local minimum. But I still love souls like and I think I'll never love another game more than a souls like. But you know, it's worth exploring other games.
Speaker 4:I wrote a piece this week called what the fuck are these battle pass things, and this is based.
Speaker 4:This is based on a piece I wrote, actually back in 2021, which was actually a series that ended up being called the Economics or Battle Pass, or Broken, let's Fix it. It's by far the most successful thing I've ever written on the internet, just in terms of view count, and all I try to do is basically lay down what I call the ADMC model, or the Average Daily Mondestation Cap when we think about Battle Pass, and the function of that is very simple. It basically it argues that we can assess the total spending cap of a free-to-play Battle Pass by only a couple components, one of which is the fixed costs of the pass, which I model here as FC. If you're watching on YouTube and it should be we'll now have video on Spotify as well. I figured out how the fuck to do that. That'll be up in Spotify, but it's FC, which is the fixed cost, and then it's the number of tiers times the tier price, which is captured by this Y part of the equation, and then we divide that by the number of days the pass is available for, and that gives us the average daily monetization cap or the maximum amount I could spend each day if I wanted to fulfill the maximum amount of spend in the past. So just to give you like a simple example, fortnite might charge $10 for the past. They might have 100 tiers. They might have $1.50 per tier. So I could do $1.50 times 100, I get $150, I'd add that to $10, it's $160 I could spend maximally and I divide that by the number of days the pass is available.
Speaker 4:So let's say it's available for three months, which used to be the model. Then we can figure out what is, on average, how much I can spend, and so this provides a little bit of a benchmarking tool to figure out what the monetization depth of a pass is. And so people have been like fucking with this model quite a bit, and so one of the things you can do to fuck with this model is you can reduce D. So if you reduce D or the number of days the pass is available for, if you make a pass that's monthly or bimonthly rather than, let's say, bimonthly or every quarter, what will happen is that you will get a higher level of ADMC doing this really well monthly battle passes. So they got this down to only one month. There have been reductions in battle pass sizes across every game. Cold duty does it in only two months now.
Speaker 4:But to be able to do this, you also need to use really low cost marginal content, and so one of the ways that marvel snap does this is that they have fractionalized and storable content like store, like currencies, which are low cost control c, control v, or you have a title, marvel snap which I think is a really innovative cosmetic where, okay, your world's funniest boss, it's just a string of text. So there's a lot of ways that you can make this really low cost, and so other games have been basically playing with d. That's one of the. That's what the fuck are these three battle pass things? That's one of them. Apex legends is another great example that tried to do this, where they split up their seasons into two, so they would have two passes in a three month season, so that's six weeks per pass. That makes D a lot lower, which increases your average daily monetization cap. So that's one thing you can do. You can fuck with D.
Speaker 4:The other thing that's been happening is people have been adding additional lanes, so now, rather than having just one premium lane, we're going to have a premium plus lane, so we're going to upsell you on that. So black cell pass from call of duty is a really great example of that legend of mushroom has passes for fucking everything and they have multiple lanes for everything. So rather than having one battle pass, they have three battle passes and each of those passes have an additional lane. That's even more premium. So they're're trying to fuck with the FC number. So how can I make the fixed cost higher rather than $10? Can we get that to 20 on average? Can we get that to 25 on average? So I would say that's the other what the fuck battle pass thing is trying to fuck with the number of lanes.
Speaker 4:Other thing that's happening is that there's this weird thing I can't figure out, which is subscriptions. I haven't been figured. I haven't figured out why that so few games have used subscriptions for their battle pass mechanic, but it's becoming more common. It feels like you would just want this to be a reoccurring purchase, of course, as we know like the power of defaults really matters. So I've been really confused why we haven't seen more battle pass subscriptions.
Speaker 4:People have sometimes casually called them subscriptions and they're not. I just want to be clear on that. You are wrong if you are saying that A subscription is a recurring charge, and so Fortnite Crew is a great example of this, where the Battle Pass is automatically renewed and they give you some additional benefits, which is really nice, for, I think, about $14 a month. It's a steep discount. And the other one who's doing this is TCG Pocket, pokemon Pocket. If you guys have been playing that. They actually give you a trial, which is really interesting. I've never seen someone do a trial, but they give you a two-week free trial, which I think is a great way to get people signed up for the subscription, and they also give you additional pack openings per day. But the main benefit is your recurring battle pass.
Speaker 2:Do you have to pay for the subscription to get the trial, or do you have to put in your credit card for the recurring payment?
Speaker 4:you do, but it's usually always an apple pay. You usually always have your credit card already, because I think even to set up apple I think even to set up an apple account you need to have a credit card attached to it that's pretty similar.
Speaker 2:What, like spotify, like netflix, hbo do, I think?
Speaker 4:yep, and it's very weird, we haven't seen more of this in games, right is there?
Speaker 2:yeah, it's a really good point, because battle pass are basically a seasonal regular. It's basically a monthly subscription, but you don't you have to re-up it manually. Is there like a payment legal issues or I don't know I don't know.
Speaker 4:I've just been confused by this. I thought it might be something technical with setting up subscriptions in the back end, but here's the other thing. Remember, on apple and google, if you have a subscription that lasts more than a year for a particular user, your rev share also bumps down to 15%. I'm just I'm really confused we don't see more of this. I thought we'd at least start to see it on web shops, which have become more popular that they would offer it on the web shop, not even on the app store.
Speaker 4:I think the only thing I've really come up with on why this hasn't been wildly successful is that really just a Beccarian view that essentially, like players are assessing their time horizons and when they see a subscription, they immediately are integrating the shadow price that they're not going to renew the subscription, and so the implicit price of the subscription is actually much higher than, let's say, the $10.
Speaker 4:Like that free trial is actually $5 in their head, because they're calculating the expected value that they're actually going to cancel it in time, and so that is actually has it. That makes them hesitant to do any sort of subscription, and a lot of free-to-play games time horizons are so uncertain that this is really just a risk aversion of me fucking up, and that being so, essentially like all this risk is essentially increasing the shadow price, and so people are not. This is why it hasn't been adopted in games is the best I could come up with. That resonates with me the way I actually came up with an experiment. I would argue for this, which is that you should implement the recurring charge if the beckering view is right. But I would argue, is that you automatically you specify that cancel. You'll automatically cancel it if players don't come in and claim the battle pass.
Speaker 1:I think that reduces the sludge problem and I also think it reduces it reduces the shadow price well, you're saying people are less attracted to a subscription because they anticipate that they won't cancel in time for a single time payment for a battle pass where it's just month to month. In this, or at least, a potential solution would be to add in a reoccurring cost that automatically cancels itself, which, regulatorily, you could imagine a policy that lawmakers would make and they'd say hey, if your thing is not being used automatically, you're not rendering any service, so you should cancel, correct?
Speaker 4:And I think it's really easy to have the action. It's if you claim the battle pass or you earn the first tier, so to speak.
Speaker 1:Oh yeah. So on this idea of experimentation, but going back up to your model with the fixed costs and the days, has anybody of you done any research on which of those is what's the marginal? Sorry, what's the elasticity on those? If you were to increase the fixed cost, how much does that increase your KPIs versus decreasing the days?
Speaker 4:Is there an optimal.
Speaker 1:It's an optimization problem.
Speaker 4:I agree. I would say I think the tiers have been a disaster. That's the one thing I've gotten really wrong in my battle pass analysis is I really thought we could do better at selling tiers. I really thought that was going to be a big revenue driver for us is that you would basically buy the gap between the tiers. So if you were in tier eight, if you ended up at tier 80, at the end of the season you would you'd buy up to tier 100. And I thought we could optimize against this. We could use like the saga playbook of we'll have some levels that are hard, require a lot of experience points. People want to buy those. I really thought we could do all this shit and that's basically been a disaster. No one's done that. No one's making money a lot of money from tiers, and no one's think.
Speaker 4:That chapter is largely closed, although I haven't completely given up on it. What has the high return? Fc or d? I actually think I don't know. From a spreadsheet perspective, chris, on which one has the high return, I would say most people are fucking with d. D has been more popular than fc. There is fc like the bat, like the black cell pass, but that's one of the few examples. By and large, it's d that's being moved I think ultimately hard to make a.
Speaker 4:It's hard to. It's hard to change d, I would argue. I would argue from a production standpoint. Actually, I think d requires more resources because, like launching a pass is also a marketing beat, like fortnite, for instance, a pass is always a marketing beat. So there's just there's a lot of dependencies on this legend of mushroom and a lot of how mobile has done battle. Passes is more flexible because they're less specific. They're literally just like giving you generic resources. This one game called mobile legends adventure just gives you, like you can just buy any pass you want.
Speaker 1:Each pass basically just continually gives you a single resource and so you, literally, you have 30 days to earn it my intuition was that d would be a higher, would have kind of a higher elasticity. Because, let's say, I'm used to seeing a ten dollar price tag for a battle task, or twelve, twelve dollars, and I say, oh yeah, the fortnight battle pass is twelve dollars and it's for a month, um, or for three months. And then I start to see a four-week battle pass and then a two-week battle pass and they're still the same price. I don't know, it just seems like you'd be a lot less in or they'd be a lot less flexible in terms of their purchasing. They're used to seeing ten dollars, whether it's for. It's like when I go to aldi versus costco. The aldi price is much cheaper, which I'm happy about, but at the end of the day, I'm getting the same per ounce price whether I'm at costco or aldi. The costco price is much higher, but that cheaper price is so much more attractive, I guess, my point being-.
Speaker 2:Wait, you're not getting a better unit price at Costco?
Speaker 1:No, okay, some products I'm getting a better unit price, but honestly not really. Now you're also that's assuming quality is identical, which it's obviously not, but when I think about cheese or something like this, they're pretty much the same price price. I feel like people are more sensitive to the skew price, the actual price on the item, and less sensitive to details behind the scenes. For example, it would be easier for me to sell a 10 skew that lasts for two weeks than a 20 skew that lasts for four weeks.
Speaker 1:I guess is my point like that 20, while it's the same deal as the other one, it's much harder for the player to choke down because it's this really, while it's the same deal as the other one, it's much harder for the player to choke down because it's this really high price. It's just more expensive than they're comfortable purchasing. Just like when I go to Aldi and I say, oh, this is great, this is only $4. Yeah, it's cheaper than if you went to Costco, but you're also getting a quarter of an ounce. It's like this tiny little amount of cheese versus a freaking pound or two of cheese.
Speaker 4:I think if we were to set up that experiment, chris, where we hold this. So basically, we could set up an experiment where we screw with FC or D but we keep a DMC constant between both variants, would help tell us. I think that D would end up winning. That battle pass has this other problem where if you're acquired at, let's say, the middle of a month and a three month battle pass, that means your time to be able to earn the last tier is significantly shorter. You need to play at a much higher rate in the remaining days to be able to earn the reward. You have this very awkward problem. So if I buy the pack, if I am acquired into a game one day before the pass ends, that means it's a really shitty deal for me to buy the pass, and that problem is reduced as long as D is reduced and that FC does not change that, and so I think D would win on that plus. One aspect is that it makes far more sense based on the variability and acquisition date.
Speaker 2:Some games used to offer a similar thing where, let's say, it's an XP boost, you pay $5, you get XP boost for for the next two weeks, which solves that timing issue.
Speaker 4:but this battle pass structure seems way more popular so apex legends also try to solve this where you would have weekly missions and the missions would only unlock each week. And so if you came into the past, let's say in week eight, which is the last week of a two-month season, then all the missions would be available to you and you can earn them concurrently. So it became like a backdoor battle pass XP boost because the mission stacked, but it's still. It still is a problem. It's still basically resting XP right in some sense that we're trying to develop. Like no one's done a really great job at solving this problem. I think D tries to take care of it. That battle the apex legends I think apex actually moved away from this, but that's another thing that I've seen that people haven't solved is like this acquisition problem, like how do we build wrestling xp in?
Speaker 1:so it's not a shitty experience if I'm acquired, let's say, the week before a pass expires so the shorter intervals is better because it gives the player more opportunity to get in at the beginning of the pass versus at the very end of the pass, correct?
Speaker 2:So it's interesting to see the evolution of the role Battle Passes play in the monetization portfolio. I think we've been talking about how there's always efforts to increase the spend depth of Battle Pass. But if you look at when they first came out, I think a lot of people it seemed Battle Passes were just an extra monetization system. It was meant to be a very good deal, but you were only limited on how much you could buy and it was meant to be a broad but shallow monetization thing. Like everyone, if you're going to play, you might as well buy this thing, but it's a great deal. If you want to buy more, you just use more stuff at the worst price point, like a first time discount price discrimination thing. Nowadays it seems like a lot of games, battle passes are their primary monetization method and, as a result, they're trying all these things to add depth to this new channel. And yeah, I don't know. It's interesting how that's replaced.
Speaker 4:It's shifted from like the broad appeal, shallow thing to oh no, this is our main thing, this is the main way we make money, and I would say that the one last thing that I think has been interesting in battle passes, and this is when the one that's most confusing to me is this skill tree based reward selection. I think, eric, you saw this in hell divers before fighting fortnight. I think, for fortnight actually might have been before hell divers. I'm not quite sure about this yeah, I wouldn't know.
Speaker 2:But I have duty, sorry, go ahead I haven't played Fortnite in ages, but I saw this first in the Eldivers.
Speaker 4:So the idea here is that when you're earning progression, you're earning battle pass currency, or we call it battle pass points. Sometimes it's represented in stars. And then when you buy the pass, you're presented with a page of rewards. Let's say it's 10 rewards and each of them are priced in that battle pass point currency. And so you have some autonomy to choose the order in which you want to get things with on the page. But the other thing is that you can go to the next page without necessarily buying all the objects on the page. So in some sense it almost becomes a shortcut to get to the last item you want. Of course you can go back and buy all of them if you want to, but it's become so. The way they also gate this is that you need to make X number of purchases or you need to spend X number of stars on a page before you go to the next one. So functionally this this functions as a shortcut where you can get to the last tier item much quicker than the first tier item. And normally I'd think this is a problem because again it reduces tier spend, essentially like the effective number of tiers goes down. So if you have a really compelling item that's in tier 100, usually what you can do is you can balance the battle pass where they might end up at tier 80. And so they need to buy the rest of the tiers to be able to complete the pass.
Speaker 4:But this basically lets you shortcut it, and from what I understand with people that I've talked with, there wasn't any deep economic reason for this change other than the fact that they just wanted to revive Battle Pass.
Speaker 4:They want to make Battle Pass interesting, and by giving players autonomy they thought the Battle Pass just became more interesting. So no deep economic reason here. It reduces a little bit of the tier problem. But I argue the reason tiers never took off this tier balancing or trying to make money on tiers is that ultimately, the number one way people make money in Battle Pass is that people buy the next Battle Pass, and ultimately one of the strongest predictors of whether or not people buy the next Battle Pass is if they had a good experience in the last one, which is usually defined by completing it or not having an incredibly difficult time to do it. So that's why I think tiers have gone out of the way and I think this is just another way that you could argue, this is LTV positive as long as it increases the retention rate in Battle Pass purchases enough.
Speaker 2:It's more engaging.
Speaker 4:All right, David Serlin, he hates math.
Speaker 1:We saved the shit-talking for the shortest part of the segment.
Speaker 2:I'll say the title.
Speaker 4:To introduce the video a little bit. David Serlin don't use math in balancing games. It is a little bit of a click-baity title. When you go into the video he's a little bit more reasonable. He's arguing that using math in a PVP game just doesn't seem to make a lot of sense or it ends up being subtractive from a lot of different experiences. Again, he's very reasoned here. He goes through examples of other game designers who have come out and said this. He tries to point out he's not arguing for authority. Come out and said this. He tries to point out he's not arguing for authority. What is really his argument? It just it creates more noise will be a good way to frame his argument that we don't.
Speaker 2:Yeah, there's a lot of pitfalls and traps to using math, and when he says math here he means both like heavily mathy parameter adjustment characters move speed and attack sizes and damage values. Right, if you stick too many parameters in there, it makes it hard to adjust.
Speaker 3:In balancing games, especially multiplayer competitive games, math and data-driven methods are way less useful than you'd think. I'll make up an example for you. Let's say we were developing a fighting game like maybe Street Fighter for the very first time. Like there isn't Street Fighter and we're creating it and we create a character, chun-li. Let's give her longer throw range than other characters. That's good, that's useful.
Speaker 3:Okay, what else can she have? How about she has a pretty fast walk speed? Okay, walk speed has an interaction with throw range. The faster your walk speed, the longer your throw range kind of feels like it is. So if you already have a pretty long throw range and you can walk quickly, that's a really good synergy. So maybe we should take how good her throw range is and multiply it by some kind of synergy factor with how fast her walk speed is. But what would that synergy factor be? Would it be 5%, 10%, double? What units are we using? What are we even multiplying? Anyway, it's all just made up and it gets worse and worse as we go on and also just a lot of data traps.
Speaker 2:If you take data at face value, like win rates or whatever, it can lead you down the wrong route.
Speaker 4:And he spends a lot of time talking about the fact that it's much quicker to talk to players. So he's taking aim at what he calls math here and he's also taking aim at a lot of analytics, particularly in the early phase. And then he's also arguing, he argues a lot of. He just says experimentation is pretty much infeasible because there's the skill element which is hard to control, for One person could be more skilled one person, the other person could be less skilled.
Speaker 3:And then we're going to use all that data. Right, that's what's going to guide us on which things to change. I think that's going to go a lot more poorly than you might realize at first glance. So let's walk through an example. Let's say that an expert play tester comes to us and says hey, I have a concern. I think that this one character is really way too good. And I know that it's a little tricky because at first glance the character doesn't seem too good. But let me just show you there's this thing that you can do with this character. It's really difficult to do, but players are going to eventually master it and when you're really doing that thing and leaning on it, the character is way too good, like 10 times too good. So that's the concern that this player is bringing to us.
Speaker 3:Okay, ready to use our data-driven approach? So first we've got to kind of kick that player to the curb because I mean what? One person saying one thing? I mean we need a lot of data. So we say, hey, you got to play a bunch of games and show us this.
Speaker 3:So he comes back. He says I played against three different people and I was pretty much able to do what I was worried about. So yeah, it's a big problem. And then we say, ok, hold on there. Three games Three, I mean what kind a sample size of three? That's way too small to do data driven stuff, so you're gonna need to play way more than that. But also a different problem is this player is coming to us. He's really good at our game and he's actually better at it than those other players. So even if he played like one of those people five times, ten times, twenty times, beat him a lot. Beat him maybe 15 of the 20 or 20 of the 20. What does that actually mean? Now we have to care about the skill levels of these players.
Speaker 2:And to add some context. So David Serlin is like a fighting game old head. He used to be like a top Street Fighter player. He's made several indie games and a lot of them are heavily inspired by Street Fighter. So, frankly, a lot of them are taking fighting games and trying to make a reduced form game of them, turn it into a card game, turn it into a board game, turn it into the simplest variation of Street Fighter you can, which I find very interesting from a thought experiment standpoint. But the games themselves I feel like he famously made Chess 2, where he tried to add asymmetric teams to chess, which is a fun idea if you like chess. It has pretty tried to add asymmetric teams to chess, which is a fun idea If you like chess. It's just a thing.
Speaker 2:It has pretty poor reviews, though. He's used to working on small games with not large player bases and he comes from the in-person fighting game world, right. So I would say he has a lot of hangups from that. I really appreciate how much he talks about his ideas, but frankly I think some of them are bad. Takes, for example he said that any game with bans is intrinsically not a competitive game and he used this to argue that MOBAs like League of Legends and Dota are not good competitive games because you can ban the champions your opponents will play, which is what have you never heard of? Hate drafting, but anyway. So yeah, interesting guy, interesting topic.
Speaker 2:I think the crux of his argument is there's a lot of pitfalls with using data and math in balancing, which I think are true. But I think what would be more interesting because he doesn't really go into, he explains a few of them but he doesn't address how to solve them. So I thought it might be worth talking about a lot of the common pitfalls and I'll go through some models. I'm going to start with the very simplest one, in which the data works at face value, and then I'll add sort of additional factors which complicate things. So imagine you have a multiplayer competitive game, a lot of different characters and you got players of different skill levels. So if in the simple world, if all characters are played uniformly across skill levels, that is, annie and bob or whatever, or these characters, the high skill players and low skill pairs are both playing them and everybody is playing against each other in a giant matchmaking pool, then win rate is actually a very valid way to look at balance, right, if everyone's playing all characters equally and whatnot, then yes, some characters win more, those characters are stronger, right. However, that is not what actually happens and, in particular, there is a matchmaking system, matchmaking rating, mmr or ELO or whatever.
Speaker 2:The games try to take high skill players and match them against high skill players and low skill players to match against low skill players. And if you imagine, every player just has one character they play, they hard main that one character and they play enough games to converge. Their true skill level, their true matchmaking rating, their win rates actually all converge to 50%, which makes it totally useless. Every character has a 50% win rate. You can't really judge power that way. What you can judge is the frequency at which characters appear. So this is another useful metric in balance is how often different characters or archetypes are appearing at the higher skill levels. Even if everyone plays till their win rate converges to 50%, you'll still see more frequent games for the high skill characters, sorry, the high power characters. So I think these are the two base cases where you should be looking at win rate and you should be looking at play rate. And the more people switch characters, the more win rate is valid. The less people switch characters, the more I think play rate is valid.
Speaker 2:However, there's also a whole bunch of other factors. So let's consider learning curves right. Some characters are designed to be easier to learn for a new player and some are designed to be harder. So in a shooting game, for example, a shotgun is very easy for a new player. A sniper rifle is harder to use. And even if you're just looking at the top level and you're saying, oh, shotgun sucks, so they're not overpowered, you don't actually realize that at the low level, shotguns are terrorizing these new players who can't aim and might actually be a huge design problem at the low level. And you're ignoring that because you're saying my balance statistic is usage at high level.
Speaker 2:Again, you have to account for the newbie-friendly, different strengths at different skill levels. There's the player experience factor as well, where the more you use a particular character, the better you get. Every character has a learning curve, different characters have different learning curves and the player bases for different characters are at different points. We saw this in League of Legends where, where these weird esoteric characters would have very hardcore player bases and not a lot of players early on the learning curve, and so a character might have a high win rate just because most of their players are experts, whereas a broad appeal popular character might have a bunch of newbies playing them and it screws with the results.
Speaker 2:There's asymmetric matchups right when, like, rock beats, paper beats scissors, and so if rock is overpowered, then paper will have a high win rate, even if paper is not overpowered. If you nerfed rock, maybe paper would actually be fine. There's team synergy. I could go on and on, but the point is there's a whole bunch of different factors that confound all this, and when David Serlin says you can't, data can lie to you and you can't trust it too much, these are. There's a whole bunch of reasons why you can't. It's hard to interpret these things at face value, and so.
Speaker 4:So I think if, if we just start with a I don't want to say a cheap deconstruction, but if we take his argument at more face value, if you can't use math in games or like it isn't, let's say it isn't beneficial to use math when balancing for pvp games. Okay, so dps damage per second very common metric that we calculate when we do balancing very informative, tells us something that's math, doesn't that just destroy his whole argument right there. And then you're done for the damage.
Speaker 2:Divided, damage per second is a formula, it's calc, I think his point and he talks about this is if you over parameterize it or, sorry, if you reduce the dimensionality too much. Right, like a gun does damage over time and maybe have a certain number of bullets. If you just look at DPS, that might not tell you the full story. Right, it might not be telling you about threat ranges or the burst damage potential or a whole bunch of other things.
Speaker 4:So that's argument for being a good analyst.
Speaker 2:That's not an argument for not using math in games, though, which is why I think his whole argument basically just falls apart. Yes, his whole thing is like your analysts suck and they're gonna read everything at face value.
Speaker 4:Well, it's. It's even worse that don't use it, because I can't read this right. Don't use it. One of the things he talks about is is just talking to players, because of all these pitfalls of of using analytics, but he falls to the same trap, which is like the whole point of using analytics. But he falls to the same trap, which is like the whole point of using analytics, is that we know that there's a bias in how people recollect their experiences. We know there's a bias in how they express themselves. That's the whole point of data. Is it's supposed to be a different type of observation? It's revealed preferences. It's based on human behavior, not what people say, but what people actually do.
Speaker 1:I'm just surprised that experienced game designer falls into a trap like this, like you can't be serious man. I do think, like probably this is one of those situations where he's saying hey, here's an alternative way to think about this. Here's why you might want to be concerned about using data analytics to, to, to guide your game design. So I'm not sure he's saying you should never use data analytics and you should never use math when you're designing. So I don't know if that's a fair criticism, because it's also like the best way to try to write something is just be like oh, this is my suggestion, I think, in maybe a little bit. You stay in the middle and you can never be wrong. I am actually going to come out in his favor here.
Speaker 1:There are some cases, especially for a game, where it's not exactly clear what are the metrics. If we're trying to do a matchmaking algorithm, that's relatively well understood. So the metrics are, we mostly understand what metrics we want to keep track of, like DPS, kd ratio, but there are some instances where it's not obvious what the analytic, what the data point you should be looking at is and what level of aggregation you should be looking at. So I actually I run into this a lot in my job, where everything will look fine and. But I know that I'm getting complaints from the community about some sort of issue and it is really tough to sift through.
Speaker 1:Okay, what's what's a valid criticism versus what is a like just people bitching, because people like to bitch. So I definitely feel for him here. There have definitely been cases where I've figured I've discovered a problem and I'm able to create the analytics that identify that problem, but only if I know what the problem is. It's almost like this can't measure the problem because you don't know what the problem is. So I agree, in general, metrics do a really good job, but I certainly think you can fall into a trap in relying too heavily on them.
Speaker 2:At Riot we used to call that data-driven versus data-informed. Data-driven is no, the metrics are the truth, and if the metrics are good, then it's good, whereas data-informed is like the metrics say, it's good but we're also hearing all this anecdotal feedback, so let's combine it to form our worldview.
Speaker 4:But that's bullshit. But that's bullshit because because that dude, that's because that doesn't say anything. That says nothing, it doesn't speak, it doesn't say anything, it doesn't have a face, it doesn't have a voice all right, fine, the interpretation of the data say something, though they literally say exactly like a hypothesis exactly, but I exactly, because you have to have a theory of measurement too. So I I know it can sound like semantics, but I think it's really important, because I think when you say data informed, or even what the data says, it's escaping responsibility.
Speaker 1:It's escaping responsibility from arguing that you have a theory and this is the empirical evidence for the theory but why wouldn't you take the easiest approach possible and just ask the players? We just we had this discussion in our game, econ discord like yes, you could technically find it in the data, but you could also just find it by asking somebody running a survey so, first off, a whole set of cognitive biases and what's the term?
Speaker 2:Hidden incentives, right, players might advocate that their character's bad to try to get them buffed, or add another character strong to get that character nerfed, to win their own self-interest. You see this all the time in games. With frequent balance updates, collection bias, you might have a totally different player base. Your most vocal players are not representative of your whole player base. I think the high-sk skill player example is very common In League of Legends. Everyone's like oh, garen's the noob character. Garen sucks, garen is stomping at the low skill levels and nobody is complaining about that, because if you complain about that, you look like a dummy and all your people on Reddit will make fun of you for being a dummy. Yeah, and also like just social information bubbles right, we've just social information bubbles, right. We've talked about information cascades, right. Like certain memes or ideas can proliferate even if they're not true.
Speaker 4:And yeah, imagine like asking people okay, what rank are you? What rank? Imagine if, instead of like, actually achieving a rank in a game, you literally just got to decide what your rank is every day. I think I'm platinum, I think I'm diamond, I think part of it is also the game is wrong.
Speaker 2:The game says I'm gold. The game is wrong, I'm diamond. Fix your game.
Speaker 4:Yeah, okay I get it oh man, I got a headshot. How that headshot me? I've got so many times 10 year olds in call of duty. What the fuck man? That was totally a headshot, I don't know.
Speaker 1:Authoritative server says something different I certainly do not consider myself a qualitative scientist, but I do think that there's something to be said for like sociological research, for political science research, for these more qualitative fields you can put together. You're not going to go on to Reddit and be like, oh, obviously this is broken because Reddit's complaining about it. So I think there's a lot of nuance here and I think it would be unfair to A unequivocally throw out what he's saying and B unequivocally claim that analytics are the only solutions.
Speaker 2:So that's what I'm saying. I'm saying he's both Agreed.
Speaker 4:Agreed, it's a tool. I don't think we should rule out any tools from our tool bag. If he came out and said we should only do analytics, I'd throw that out the door too. I think that's the wrong take. Ultimately, we need empiricism, we need qual, we need quant, we need theory, and all these things come together.
Speaker 2:I think that's what we're best at I think what amin said about survey data was really insightful. If the players incentives are aligned with yours, the qualitative feedback is very good.
Speaker 4:If they're misaligned, like we're talking about ranking or nerfs, you get garbage and that is again you need to correlate the surveys with like skill level right, become even more insightful. It's like how to skill, how to high skilled players respond to the survey versus low skilled players. How do I even? I think one of the things I really wanted to get to that I feel has been so poorly defined in games is like what is it? What is the theory of balance? What does it mean to have a balanced game?
Speaker 4:And so I've heard like basically two definitions. One has been equal win rates, which I think people have pretty much thrown out the door as not being viable because of the skill problem. There could be self-selection. It could be like if you sell a character, then only people who are paid end up buying the character, and maybe people who have money tend to have a higher win rates. There's a bunch of noise that makes win rates problematic, and so then the next definition I've seen is they want to make each character in a PvP or each weapon, whatever it may be, each one be viable. So it isn't necessarily that each weapon has the same win rate, but each weapon is viable, and that is much harder to define. But I've yet to see a really crystal clear definition that game designers have rallied around. This is what it means to be balanced. They basically just bullshit each other and just ah, when it feels right, what does?
Speaker 2:that even mean Balance should not be a goal in and of itself. In fact, a lot of the most fun games are not balanced, and it's fine. A lot of games that have stood the test of time, like Counter-Strike, like Smash Bros Melee, are horribly unbalanced. You look at the gun usage distribution in Counter-Strike it's like AKs and M4s all the way down. If you look at Super Smash Bros Melee the top tiers it's like super duper popular and it's like overwhelmingly stronger than the weak characters. Yet these games live on and have thriving player bases and are super compelling. And I think the fact of the matter is, what matters is if the game is fun. And instead of win rates or play rates or whatever, what you should be looking at is are the characters that are being used a lot? Are the play styles that are being employed a lot? Are they fun to play against? If Counter-Strike was all shotguns, let's say shotguns are super duper strong, it would be no, it's like peekaboo hide around corner and shotgun people.
Speaker 2:The fact that rifles are the best gun means that the gameplay is centered around this kind of mid-long range duels, right and fighting. And you have your snipers for the super long range. You have your shotguners for the super long range. You have your shotguns for the super close range and they do. There's strategic reasons to use them, but the base mode is the mode they wanted to create, which is mid to long range distance fighting. Again, if you look at Smash Bros, melee Fox is a very fast, aggressive character and they want the game to be fast and aggressive. He's in your face and he's doing all sorts of stuff. That's the type of gameplay you want to create. You don't want to make a game where the best character is a projectile, camping, runaway and shoot missiles, right. What are you doing at that point? You're not even playing a fighting game. You're playing a shoot like a long distance shooting games. So the game should be tuned to create the dynamics of gameplay you want, right?
Speaker 2:Yeah, balance is an interesting lens to look at that, but that's not the goal.
Speaker 4:What would it be a perfectly balanced game for you, Eric. What would that mean?
Speaker 2:I don't care about the balance, is it fun?
Speaker 4:So I guess what you're saying, though a game is perfectly balanced if it maximizes retention and fun engagement, so to speak.
Speaker 2:I would say to your point the second balanced definition you gave, which is everything has a function and a niche. That's more about we didn't waste any design space. I didn't make a tool that nobody uses. Every tool you make and the content you put in the game, you want it to be consumed at some point, but I don't think that's actually important. For the game to be fun, you can have something nobody uses and the game would be no different.
Speaker 4:So in the Super Smash Brothers example you gave, where, let's say, there's really only three or four viable characters, you would say that game is. Maybe balanced is in the right word, but we could call it balanced because ultimately those three or four characters end up creating, they end up creating engaging gameplay and ultimately that's what matters. And so the game is balanced because it has engaging gameplay and so you would only do a balance update if you felt like it wasn't creating engaging gameplay so isn't part of the, because there's this other dimension to balance, which is that balance is almost like a puzzle.
Speaker 4:When you have new content is they're always solving for the win maximizing solution. You're solving for the meta, which is why these metas evolve as people do balance updates. It's content that you're consuming when a balance update happens. How do I think about that?
Speaker 2:I think that's it. It's content it makes. If shaking the box makes the game more fun and engaging, great, so it's almost even if it's almost it makes.
Speaker 4:If shaking the box makes the game more fun and engaging, great, so it's almost even if it's almost about purposely creating imbalance, because that's content and creating waves of imbalance yeah, as long as the puzzle has to be interesting, though right.
Speaker 2:If the puzzle turns out like oh, option a is the right option every time, it's not a good puzzle and players will critique it as this game is imbalanced, but but really it's. The puzzle you created was bad because there's one solution that works every time.
Speaker 4:So should designers be direct in the puzzle solving, Like they when I'm thinking about designing Super Smash? If I was a designer on Super Smash Brothers, should I design in such a way that this meta is going to merge with this set of characters and then when I do another balance update, I'm purposely trying to have this set of characters become viable again? Or should it just be the spontaneous order of players to discover it and I'm reacting?
Speaker 2:I think both can work. I know in Snap we often intentionally, when there's a new season, a new seasonal card which we know a lot of people will play, we know this card is going to be strong. We should make sure there's enough support cards for it. We should also make sure there's enough counter strategies for it. So there it snap, I know, is very intentionally designed that way. There's also a game, like I said, that a lot of fighting games in particular are often much more. Here's the base tools.
Speaker 4:You guys allow the emergent properties to so if I were to, if I were to wrap the, if I were to like summarize eric's theory of balance, is that the only proper set of like balance values or system values? We could think about it like a spreadsheet of like attack and health for each character and all the characteristics they have. We could think of that as bundles and we could create different bundles which change different values. You know, at some point there's some bundle which maximizes engagement and ultimately we want to pick that bundle, and then we also know that there's diminishing returns to that bundle, so we want to create another bundle at some sort of cadence that players have to solve. That's ultimately, what's about. It's about the y variable being engagement, not balance. I don't know. Am I off base?
Speaker 2:okay, no, that's right, and balance is a fucking great theory, eric.
Speaker 4:I love it. I love this theory. Dude, I'm signing up. This is the econ theory of balance.
Speaker 2:I'm glad this is a sub stack man win rate and stuff is like a useful indicator. Right, it's a useful diagnostic metric. It is not the end goal so I don't know can we think about? Is there any more?
Speaker 4:is. Is there any concrete quantitative way to measure exactly what you're talking about? How do I know if this is the the optimal balance that's creating something fun and engaging? Is it intuitive?
Speaker 2:it's hard because you can't have like clean experiments, right?
Speaker 4:you can't but we can still have a model or a theory right. Is there any way we could know if we're on the right track or the wrong track, Like, how do we know if something's imbalanced or balanced?
Speaker 1:If the players are playing it, then it's good, right? I feel like you're making this way more complicated than it needs to be. If you have high engagement, high retention and a decent player base, especially within your category, I think that's the Y variable.
Speaker 4:But I want a level change. I want Y plus one right. I want to pick the system design value that gives me even more engagement.
Speaker 1:Well then, that's an easy experiment. You just serve up a different system to your B group and see if that increases engagement.
Speaker 1:It's more about engagement than monetization. I think this is actually also relevant to non-competitive games like economy games. I actually think that there should be imbalance in economies, right, like you might think that it's balanced, if everybody gets to use every item or everybody gets to have one of everything, um, that actually makes for a really shitty economy, though, because if everybody can do, if everybody gets to use every item or everybody gets to have one of everything, that actually makes for a really shitty economy, though, because if everybody can do everything, then there's no trade, and you want trade, and you want prices to be higher in one region and lower in another, and the only way to get that is through imbalance. So I almost I have this philosophy. We talk about it in my company a lot.
Speaker 1:These asymmetries glitches, they actually cause a lot of interesting stuff. So that's why, like when we'll have one faction in the game we'll say, they'll say, oh, our side of the map sucks, like we don't have access to this resource. Yeah, you do have a disadvantage in that particular metric, but you have an advantage in this other metric. You guys might have not, you might not have good access to copper, but you do have really good access to iron. So it's this balancing act where you want them to enjoy it, and if people are complaining, that's no good. That's the objective function that I care about. If everybody hates it, even though I'm happy as the designer, then it's not what we set out to do, and that goes back to what I was talking about with respect to player feedback. Yeah, I think I like this idea of imbalance is a good thing. Imbalances, it can be okay if everything's perfectly balanced. What does that mean and is that necessarily the objective? So it extends, I think, also to economy games as well.
Speaker 4:All, right, let's wrap it up. Boys game. Economist cast episode 34 in the can. Happy, happy Thanksgiving, guys, happy.
Speaker 1:Thanksgiving Happy.
Speaker 4:Turkey Day Happy Turkey Day. Enjoy Watch some football. Kids, Peace out. We should teach this to our children. Economics is major. Everyone has to major in economics. Number one for personal survival.
Speaker 1:Economics is major.
