<cite>Chris:</cite>
<time>0:00</time>
<p>I was at Costco like two days ago and I got a dollar 50 hotdog, so don't, mess with me.</p>
<cite>Phil:</cite>
<time>0:04</time>
<p>on. Just to clarify, Chris, if you're an economist at Costco, are you raising the price of that hotdog? Are you fixing that? Are you pegging that to</p>
<cite>Chris:</cite>
<time>0:10</time>
<p>I saw like a breakdown of their income on some, I don't even remember where it was. And they have like visually razor thin margins compared to a tech company or something like that. But it's just consistent. It's like Coca-Cola. It's, Brand that just keeps delivering. I think that's their philosophy. Their philosophy is customer satisfaction. That is the top priority without a doubt. I know that if I go to Costco, no matter what I'm getting, except for maybe with some of their clothes, which I'm a huge advocate for Costco at any time, I'm decked out in some Costco Drip just saying</p>
<cite>Phil:</cite>
<time>1:00</time>
<p>you like a Kirkland tattoo or something. I</p>
<cite>Chris:</cite>
<time>1:03</time>
<p>I, I haven't quite gotten to the Kirkland brand like sweater. But anyway, you know exactly the high quality you're gonna get. So I think it's just one of those things like how much are they gonna possibly make up out of hotdogs? I think that you could do that analysis. I think you could say, okay, what is the counterfactual of raising the price of a hot This is ridiculous. Raising the price of the hotdog by a dollar to, to be in line with inflation or$2, or I think like with inflation it should be like four 50 right now or something like that.</p>
<cite>Phil:</cite>
<time>1:30</time>
<p>You heard it here first. Here folks. Chris is taking the Costco hotdog to four 50</p>
<cite>Chris:</cite>
<time>1:37</time>
<p>In line with inflation. No. You do that and you do that calculus and see what the loss in, like customer loyalty is what, like it's a small effect. It is a small effect, but when you're Costco, half a percent, a 10th of a percent matters. Like it matters as Costco. Their margin they operate on these like very tight, consistent, like they're an immovable object. Everything that they do, their customers are extremely sticky. That's their whole thing. More sticky than even Amazon's subscription model. Because it's food. It's this like necessary thing. Also, I think like you don't have the option to you buy a year at a time at Costco, I don't think you can just drop out, but I, so no, I wouldn't raise the price of a hotdog, I don't think.</p>
<cite>Phil:</cite>
<time>2:19</time>
<p>I, I think there are ways you could do this. Because it is a cost center, right? The hot dog. The hot dog is a cost center. I wonder if you could start to blend it into a meal, like you can only get the hot dog through a meal.</p>
<cite>Chris:</cite>
<time>2:30</time>
<p>No, the hotdog comes with a meal or it comes with a drink. At least I think it, yeah, it comes with a drink. So it's like already a meal.</p>
<cite>Phil:</cite>
<time>2:37</time>
<p>a buck 50 and you get a hotdog and a</p>
<cite>Chris:</cite>
<time>2:39</time>
<p>Yes. And a big drink. Like an American size drink. I always leave the drink behind because I'm not like a soda drinker.</p>
<cite>Eric:</cite>
<time>2:45</time>
<p>Is it's the brand of Costco is you're getting a great deal on this giant hotdog, Like even if they're losing money on it and yes, it gets people in the store, but on top of that it makes you think, oh yeah, all the prices here are super low,</p>
<cite>Phil:</cite>
<time>2:55</time>
<p>This is actually signaling, this is come into our store. We have cheap hotdogs. You're gonna consume that immediately. Come for the hotdog. Stay for the bulk</p>
<cite>Eric:</cite>
<time>3:04</time>
<p>everything else is cheap too,</p>
<cite>Chris:</cite>
<time>3:06</time>
<p>I mean, I seldom just by a hot dog. I'll buy like a smoothie or something, or my wife will get that and I'll get a hot dog. Or she'll get a hot dog and I'll get,</p>
<cite>Phil:</cite>
<time>3:14</time>
<p>It's a marketing tool almost. And you, to your point Chris, it's nothing on the p and l, hotdogs are nothing</p>
<cite>Chris:</cite>
<time>3:20</time>
<p>yeah,</p>
<cite>Phil:</cite>
<time>3:20</time>
<p>game Economist, cast episode four. We are here with our usual hosts. I am Phil Black, out of Game Economist Consulting. I am joined by my two other wonderful, unusual co-hosts. We will have a guest soon, I promise. I promise. Not that there's anything wrong with Chris. how are you?</p>
<cite>Chris:</cite>
<time>3:38</time>
<p>doing well. My wife and I are in the middle of buying a house, so very exciting times. Indeed.</p>
<cite>Phil:</cite>
<time>3:43</time>
<p>And we have Eric joining us as well. Eric, how are you?</p>
<cite>Eric:</cite>
<time>3:46</time>
<p>Hey, doing all right. I'm in San Diego with the parents for Chinese New Year. Taking the little kids out to one of those parades.</p>
<cite>Phil:</cite>
<time>3:51</time>
<p>Just as a reminder, we're gonna be doing a mail episode in the future. Continue to send your mail to the cast. We gotta think about five or six messages last week. It's
[email protected]. Continue to send along. Send us memes, send us troll questions. We'll read it all.
[email protected]. Today we are gonna talk about two articles, only two articles. Today, I will be trying to milk for as much as possible my article on Marvel Snap. Why is it struggling to scale? It's the vertical progression. Stupid. I did talk about this on another podcast that shall not be named far inferior to our podcast, obviously, but</p>
<cite>Chris:</cite>
<time>4:30</time>
<p>of it.</p>
<cite>Phil:</cite>
<time>4:31</time>
<p>never heard and Eric, what are you gonna be talking about today?</p>
<cite>Eric:</cite>
<time>4:34</time>
<p>I'm gonna talk about matchmaking specifically in role-based team games. And you know what they can learn from mechanism design because a lot of matchmakers try to home brew their own algorithm and then they fall into classic mechanism design traps.</p>
<cite>Phil:</cite>
<time>4:52</time>
<p>Chris, what have you been playing days?</p>
<cite>Chris:</cite>
<time>4:55</time>
<p>I've been playing a lot of mobile games because I realized that was something that was not in my rep. I started playing this super hyper casual puzzle game that was recommended in our group chat. It's basically Sudoku, but a little bit easier and a little less annoying. Nanogram I use the nanogram.com app. There are ads. It's a little bit annoying. I've also been playing Marvel Snap on Mobile. when I initially started playing Marvel Snap, it was on it was on PC like I was playing it on my pc, like an ultra wide monitor, like Marvel Snap. It's this little strip in the middle of the TV or the screen. So I started playing it on mobile and now I get Now I'm like, oh, shit. They made like the perfect mobile card game, not the perfect card game in general, the perfect mobile card game. So I'm super excited about that. And then I've also been playing something in between, not so competitive called Poly Topia which was recommended to me after I posted about finally playing Marvel Snap on Mobile</p>
<cite>Phil:</cite>
<time>5:47</time>
<p>And Eric, how about yourself? What are you up to these days? What are you playing?</p>
<cite>Eric:</cite>
<time>5:50</time>
<p>I've been going back to Super Smash Bros Melee. That's I heard someone use the term home game and that, that sounds right to me. It's like the game I always come back to. and yeah, I</p>
<cite>Phil:</cite>
<time>5:59</time>
<p>the game</p>
<cite>Eric:</cite>
<time>6:00</time>
<p>melee the one for the Game Cube released in 2001.</p>
<cite>Chris:</cite>
<time>6:03</time>
<p>Do you play it on Game Cube or do you use an</p>
<cite>Eric:</cite>
<time>6:05</time>
<p>Nowadays they use emulators. There's there's this whole fan mod community, which they created net play for it, and they actually have like better net code than any of like Nintendo's official games. It's crazy. It's a deep dive rabbit hole if you wanna go into that sometime. But yeah, I love that game.</p>
<cite>Chris:</cite>
<time>6:21</time>
<p>So I know for like old games, like on the super on the Super Nintendo, if you want a perfect if you want no latency, perfect connection, you have to use like an old school monitor the big glass big giant cube that most of us you know, listening to this have seen at least once in our life. I can't remember if that issue persists for Nintendo Game Cube, if you wanted to play old school Game Cube.</p>
<cite>Eric:</cite>
<time>6:40</time>
<p>For the longest time that was true. Nowadays, computer monitors are getting better. and the crazy thing is, so the Game Cube has a built in three frames, so that's 0.5, 0.05 seconds of delay built in. but in the emulator, they somehow make the game run faster and cut down that delay time, so to make the game run faster than it could on original hardware. And then they use that delay time to make the internet net code better. Yeah, anyway it's, a crazy rabbit hole. But yeah, the people care a lot about what monitors they play on.</p>
<cite>Phil:</cite>
<time>7:09</time>
<p>I don't understand the attraction to Super Smash Brothers Maylee in this day and age. What? What is, why? Why are people going back to this? I know there's Evoke championship. I don't understand this fighting community.</p>
<cite>Eric:</cite>
<time>7:19</time>
<p>do I say it? It's like super fast, but also like and free flowing in a way that very few competitive games. I could gush for hours about this, but it's, game is your, like social life. It's like a tight-knit group of friends. It's like deeply emotional when you play and empathetic cuz you're trying to understand your opponent's psychology and trying to play around it. It's deeply strategic for people who are really into analyzing and breaking down optimal strategies. It's very creative and expressive for people who wanna show off that they can invent new things. It's, I don't know, people who play it, they call it a lifestyle game for a reason. It like, every aspect of their leisure time can get consumed by this.</p>
<cite>Phil:</cite>
<time>7:56</time>
<p>any, is it really that exciting when there's no new content, there's nothing new to master, it's people just mastering old things. There's no, not update.</p>
<cite>Eric:</cite>
<time>8:04</time>
<p>Yep. Yeah. People still play in chess, right? That game's super unbalanced. Fucking queens are broken. The other game I wanna shout out is a mini metro. It's a super chill mobile game. Released probably eight years ago, but I've been playing it when I watched the baby at night. It's very relaxing.</p>
<cite>Chris:</cite>
<time>8:19</time>
<p>Looks nice.</p>
<cite>Eric:</cite>
<time>8:20</time>
<p>It's about laying out a metro line in a city and getting all the passengers to the right destination.</p>
<cite>Chris:</cite>
<time>8:24</time>
<p>All right. Now I've got five games in my mobile load out. Sorry,</p>
<cite>Eric:</cite>
<time>8:29</time>
<p>It is fantastic On iPad.</p>
<cite>Phil:</cite>
<time>8:30</time>
<p>I can be the central planner. I've always wanted to be. Hi. Hayek is somewhere rolling in his grave in this game. Is it always a disaster? Is are things like always over budget</p>
<cite>Eric:</cite>
<time>8:40</time>
<p>They remove a bunch of that budgeting stuff. You can reroute lines with zero cost, even it's like digging massive tunnels. But yeah the game's designed to, it escalates to the point where you can't keep up. Some people have crazy high scores, so you know, they must know some urban planning strats that I don't.</p>
<cite>Chris:</cite>
<time>8:54</time>
<p>Phil, what have you admit up to?</p>
<cite>Phil:</cite>
<time>8:56</time>
<p>I have been playing a recommendation actually from you, Eric. Let's talk about, let's talk about this one instead. Because I've been absolutely obsessed with super auto pets, I'm understanding the name of the game correctly. And for those unfamiliar with super auto pets, it is of an auto chess game. You have a lot of the core auto chest mechanics where you'll start the game. You have a pool of manna, or what they call gold in particular instance. And with that gold, you'll buy from a set of characters that refreshes every single round, and you'll assemble a team from those characters. Then you'll enter into a battle phase in which your characters will battle. And this is really where the twist comes from for super auto pets, is that the battle mode is done in pair wise fashion. Pet will attack another pet from the opposing team. Simultaneously, the battle will resolve and then it'll move on to the next battle. And rotates in almost a king of the hill fashion until only one pet is left standing and then someone is declared the winner. It is absolutely addictive. It moves incredibly fast. You have this really tight cause effect loop where you can tap, and when you tap, it'll accelerate in action. And because it's pvp, they can go ahead and they can do that in many instances. And so you just keep tapping. And these, it's almost as like you're building those guttenberg machines where you have the dominoes and the dominoes will, have a roll down a hill and the marble rolling down the hill will enter into a cup and then the cup. And so what happens in the game design is that all of the really function from game design perspective on the order in which they're placed. So if the pet ahead of me takes, let's say, one damage, the pet behind me might gain one health on that attribute. So you're really trying to think about. how can I build this incredible team, this incredible machine where one thing happens, then something else exciting happens, and then you're facing off against someone else who's trying to screw with your strategy and that it's just, it's extremely tight. Just the game design is very And the other thing I just want to call out here that I thought was fascinating is that they don't leave a lot of the mechanics just for the battle. They also take it into that drafting, that economy phase where you're trying to assemble your team. So for instance, one of the pets, which is an otter, if you sell the Otter during that phase, you'll actually grant health to the rest of your pets. So there's a lot of interesting things they'll do in the economy loop, not just in the battle loop. And I am ready to see this on mobile. As a mobile free to play game. It is what, it's a free-to-play game, but it's very paid dlc. No one's really doing the live service twist on this game. And it's blowing my mind. This is ready to be Vampire Survivors 2.0</p>
<cite>Chris:</cite>
<time>11:22</time>
<p>See the way you describe it, I sa it's, it sounds like a Marvel Snap competitor. It's fast game play. It's a drafting game. Your deterministic outcome, like simultaneous play. What sounds like a card game to me, but per, it sounds like a, a DEOs of war card game.</p>
<cite>Phil:</cite>
<time>11:37</time>
<p>I'm gonna probably gonna throw up in my mouth by saying this, and I'm gonna use the word hyper causal, but I could say that it's almost a hyper causal version of Marvel Snap in some sense. Obviously there are a million different differences, but just that very quick mobile focus gameplay I think really comes through in this one. you have ruined so many dinners with my wife Eric. You've nights, I think. I think I'm finally done with super auto pets, like I milked it for all it's worth. But my God, that was an unbelievable three week period of my life.</p>
<cite>Eric:</cite>
<time>12:06</time>
<p>I'm glad I converted you. Yeah. The game is amazing. And to your point, it's so it's like vampire survival level, like simple, yet ingenious, and I'm waiting for someone to try to, take advantage of this genre.</p>
<cite>Phil:</cite>
<time>12:17</time>
<p>article time.</p>
<cite>Eric:</cite>
<time>12:19</time>
<p>Let's do it.</p>
<cite>Phil:</cite>
<time>12:20</time>
<p>Go for it.</p>
<cite>Eric:</cite>
<time>12:21</time>
<p>so first I'm gonna talk about just generally matchmaking, a little break, talk, have some then we'll talk about stability, and then we'll talk about strategy proofness. So yeah just I wanna try to chop it up so that, it's not just me, monologuing. All right, here we go. Yeah, so I'm here to talk about matchmaking in role-based team games specifically. and I think there's a lot of lessons that, game designers can learn from mechanism design, which is, know it's a subfield of economics, which is focused on if you have a bunch of PL users who report their preferences, how do you. Arrange outcomes in such a way based on their preferences or their bids or whatever, what have you, so that you get a good outcome. and I think two important learnings. One is that stability is an important thing. This is something that mechanism DI design cares a lot about. You wanna create a match that people want to stay in. And we can use incentives, unique incentives in these online games to do that. And the second is strategy proofness. That is you wanna make a matchmaking system that is unable where people can't lie about their preferences. And a lot of these role-based games have a support death spiral that gets created if they don't properly handle this. First quick primer on matchmaking generally. Matchmaking is super important in these social competitive online games cuz the quality of your experience is not just the game itself, but also who you're playing it with and who you're playing against. Broadly, they balance two things. One is cue time or wait time and match quality. You wanna get a good match, but you don't wanna wait too long. And match quality depends, varies from game to game, it usually includes the skill difference between the teams. You wanna have a close match maybe the connection quality between your own opponent, you and your opponents for an action based game. Maybe the skill variance within a team cuz you don't want your best endorse players to be miles apart cuz the worst player will have a terrible time. And among other things and one thing to note is that the matchmaker usually assumes the preferences of the users. So if I'm a user, if I really highly prioritize a high quality match and I'm willing to wait half an hour for a great match the matchmaker doesn't know that, right? Or if I'm just saying, get me on a match as fast as possible, the matchmaker also doesn't know that. It assumes some trade off between those two. There is some ability for users to differentiate, like often ranked modes place a higher priority on quality, whereas quick play places a higher priority on speed. But yeah, matchmakers assume preferences of the users, right? We're gonna see how that creates problems later. But yeah, a lot of great information about this. Apex Legends recently released a great blog post diving into this. There's an older article from awesome knots developers on game developer.com that talks about some of the proms smaller games have with matchmaking. Highly recommend, but today we're gonna,</p>
<cite>Phil:</cite>
<time>14:45</time>
<p>notes. We'll have that in for sure.</p>
<cite>Eric:</cite>
<time>14:47</time>
<p>But today we're specifically gonna talk about role-based games, like imagine games like World Warcraft, where you have tank damage dealers and support characters. Overwatch, also those three line up, league of Legends Dota, these MOBA games where there's five different roles. These role-based games often have hard role specialization and have role specific matchmaking, which creates, new challenges.</p>
<cite>Phil:</cite>
<time>15:10</time>
<p>So here's the thing that's of blown my mind, is that there's very little ability for a player to exp express any preference when it comes to matchmaking, which is something you were talking about earlier, that just because player expressing a preference in matchmaking just ends up expanding the pool, or excuse me, shrinking the pool at which I can match them.</p>
<cite>Eric:</cite>
<time>15:29</time>
<p>I think it's two things. One is the complexity of the UI side. Imagine your game had a slider that was like wait time versus match quality. how many people would actually know what to do with that. And the second is on the algorithm side. I'm sure it complicates things a lot.</p>
<cite>Chris:</cite>
<time>15:40</time>
<p>Methodologically, I think the most interesting part about this is all of the mechanism design decisions that the developer has to make are certainly interesting. And I think there are a whole slew of different options for that. But really the The ma the like computationally complex or the, the challenging part is actually just the score rankings. It's very fascinating how they go about calculating those because you can't just, you can't just take everyone and put them into the same pot and just rank them, from worse to best because you only observe a binary outcome of win losee or, kills versus losses. So you only have your unit of measure is like this transaction and that transaction is actually a battle. It's very difficult to measure the quality of a player, oh. They're faster or they shoot people more often. That's all great, but are they shooting more people more often because they're versing terrible players, or are they shooting people more often because they're just overall superior?</p>
<cite>Eric:</cite>
<time>16:33</time>
<p>Yeah, I think skill-based matchmaking, right? E ecosystems, true skill, mmr, gco it's a huge statistical problem, for the reasons you stated. Yeah, I'm not gonna dive too much into that. That's like stats land, not really my expertise. But yeah, it's interesting a lot of games, to your point know ideal the outcome is just win loss, but often that's not enough data or a signal to get the results you want. Some games, especially shooters, have started to incorporate in-game stats, in skill rating updates which is controversial cuz you can, play to optimize your stats rather than optimize your team's chance of victory and you don't want to misalign those incentives. Also like games with live matchmaking these online games are very different than tournament based games like tennis or fighting games where the, there's a huge selection bias in what matches are played. Yeah not my area of expertise for sure. Let's jump into stability. One thing mechanism design cares a lot about is stability. I think you can pull intuition from marriages, right? Imagine you have two couples in their respective marriages, but one member from each couple would rather be with each other than with their current partner. And so those two, run away with each other, That's an unstable marriage, And ideally your matchmaker creates stable matches where nobody wants to leave. And this is a big problem in online team games where people might intentionally leave their team. Sometimes they, a player might not get the role they want. They're like, oh, I wanted to play damage, but I got support. I'm outta here. Or they might look up their teammates and say oh, this teammate looks terrible. I'm outta here. Or I want to play with a tank support. And, you're playing a healer and I don't wanna play with you, so I'm outta here. So stability is a problem and obviously the matchmaker can improve that, but there's out of. incentives we can add to help this? So most games have a Q Dodge penalty where if you leave during the team formation phase, you might lose ranking points. If it's ranked you're might be locked out of matchmaking for 5, 10, 30 minutes. Yeah, there might be other penalties as well. That's all kind of designed to reinforce that stability of the match.</p>
<cite>Chris:</cite>
<time>18:23</time>
<p>I was gonna ask, is stability's important for player retention or just, what's the main. Because I think about the marriage example, it's those people have, those people don't really have the option to, it's much harder to leave a marriage than it is to leave a game. So the idea is you wanna maximize this experience. So does that involve making sure that everyone has this really. Tight, crazy, competitive experience or does it mean that you're supposed to pair, some really great players with some less great players so that those great players are like, oh man, I just, killed this. Yeah, I feel so good. I won. I find really interesting because there's this relationship between stability and retention. My first, I, I played Fortnite for the first time in a year a year or two recently. And I won and I was like, what the hell? Like, how did I win? Like I never won before as I was terrible at Fortnite. And I looked it up and I realized, oh, they word on the street is that they filled the first match with a bunch of bots. That's super un to me that's like unstable. The match was unstable, but I'm gonna stick around cause it felt good to win and I'm gonna go play another game. So how do you think about that and how do you yeah, how do you see that?</p>
<cite>Eric:</cite>
<time>19:28</time>
<p>so I think in this context, a stable match is a match that people want to play. So an unstable match is you get formed into the team formation lobby, someone leaves and you have enter Q again, and then you join the team formation lobby, someone leaves, you enter Q again. So stability is like the match actually executing. And you can imagine that's a terrible experience if people are constantly bailing on you and you have to go back into the queue again.</p>
<cite>Phil:</cite>
<time>19:48</time>
<p>Does stability include people leaving halfway through the match itself? Could we just think of it as the amount of players starting a match and the same number of players finishing the match?</p>
<cite>Eric:</cite>
<time>19:58</time>
<p>I think so. Arguably right, the first one, leaving before the match is not getting married. I think leaving in the middle of the match is getting divorced, but. it's, at some point during the course of the match information was revealed to the player where they said, I don't wanna play this game anymore. And they bailed.</p>
<cite>Phil:</cite>
<time>20:12</time>
<p>So those things matter, right? So you're claiming that those things matter. So if I'm designing a matchmaking algorithm or just a matchmaking system, and I'm like, okay, I think retention matters because, we don't want a player leaving in the middle of a match. That would be a negative experience. I think we've all played Dota, or at least have some experience of playing Dota Overwatch, where that happens. would be my next action? You said you can penalize, right? This has just become, a Gary Becker style, probability of probability, the cost of crime versus the probability of enforcement trying to get that right, or is there something more clever?</p>
<cite>Eric:</cite>
<time>20:42</time>
<p>I think that's exactly right. Most of the systems people use are penalty systems, If you leave in the middle of the match, you lose some currency or you lose ranking, or you're locked out for a while. it's tricky because sometimes. penalties are very easy to get around. You leave a match, you just log off for 30 minutes and come back later. think there's room probably for positive incentives here as well. Play through incentives. So one thing that League of Legends does is, so they have a role select, you can select your top two preferred roles, but if those are unavailable that you get auto-filled into a role that you probably did not want to play, and this might be an unstable match, you might say, I'm gonna leave and I'm gonna req and try to get my preferred role. But if you've play out the game gives you auto-fill protection, meaning the next time you queue you are guaranteed to get one of your preferred roles. And so this is a positive incentive for playing through the game. And yeah, I think games could probably try to do more things like that where essentially create between match incentives that encourage you to stay stable for the duration of that match. Another bit is on strategy proofness. I think this is a little more intricate, but imagine you have a game with three roles, right? Damage tank and support, right? This is pretty common in both like RPGs as well as in Overwatch. and they're usually that order preference damage is very popular. Tank is moderately popular, support is unpopular. How do you match make people, People are highly specialized. They wanna play one specific role and the way Blizzard does it in both Wow. And Overwatch is single roll select. You pick a role, one of the three, and you enter Q and you get a match. When you get a match. And what ends up happening is there's a massive disparity in wait times, damage rolls often are waiting 10 minutes in Overwatch, whereas supports, get matches almost instantly. Similarly in World of Warcraft raids healers are always in hot demand. So this isn't ideal. Dota two does something similar where, but it's multi-select. There's five different roles and you can pick the subset that you are willing to play and it will just wait until you get one of those, but you still have long Q times because the least preferred roles, the support roles are the hard bottleneck for this. League of Legends, like I mentioned earlier, allows you to select your top two preferred roles, but it will also force you into playing an undesirable role if it needs to fill it. And this massively improves Q times, but also, creates kind of these incentive problems. intuitively as you're designing the matchmaker, you might say okay, I wanna get as many people in their most preferred roles as possible. First I take all the primary damage, primary tank and primary support players and, they form teams. Great. Now, since we're out of support means we're gonna get some support secondary players in here. And intuitively you might say okay, let me get like specifically. support secondaries from damage is the most popular, right? So the most common support secondary group will be damage first, sec support, second tank, third. Lemme pull a bunch of these players in to fill it. All right? And so you optimize the number of, first, you optimize the number of preferred roles assigned, and then you optimize the number of second preferred roles assigned, right? The problem with this is that all of a sudden you're perverting the preferences, right? If I'm a damaged, let's say I prefer damage over support over tank. When I list support as my secondary, it increases the likelihood I get pulled into support, and it decreases the likelihood that I get my primary desired role damage. And if I had put tank as my second preferred role, it would've increased my chances of getting my number one preference. so all of a sudden the system is no longer strategy proof, right? I am lying about my preferences in order to get a better outcome. And this goes back to like arrows impossibility theorem, Where it's very hard to design totally proof mechanisms, but this perversion does happen. And we saw this happen in League of Legends when we released the role, role preference system where support used is the least popular, but it was around 16%. So out of five, it's at average of 20%. but as people realized that putting support second decreased their chance of getting their main role fewer and fewer people were putting support second to the point where it dropped down to as low as 12%. and they had to make some major changes to the matchmaking system. One solution to arrows and impossibility theorem is the random serial dictatorship that is like essentially randomly some people. Their first preference every time. In League of Legends, what they did was for the first minute of Q it will only search for your number one preference. and after that minute is up, then it'll expand the search to your second. And so that this helped things a bit. There's still a little bit of perversion. And another thing that did was just try to design support to be much more desirable, but it's the danger of this scenario is that the popularity support goes down. People lie about not wanting to play support, but at some point, if I put support sec if I stop playing support altogether, my preference for playing support drops and this role gets even more unpopular, which means even more people have to be auto-filled into it. And it creates a deaths unraveling spiral.</p>
<cite>Phil:</cite>
<time>25:16</time>
<p>So the idea of creating a strategy proof, or at least resilient matchmaking preference system makes a lot of sense to me. Do we need to go to that level of sophistication? Is it possible just to provide incentives? I look at what Overwatch is doing right now. If you play something that has, if you play a role that has longer wait times, they just give you more experience points. Is that just like a more straightforward solution?</p>
<cite>Eric:</cite>
<time>25:40</time>
<p>I think it can work, but the risk is that the, it's of apples and oranges. These two things are categorically different. And players' preferences might swing very hard one way or the other. They might not care at all about xp and so your system doesn't work or they might care too much about it and they're optimizing for farming that resource. The nice thing about league's auto field system is it's within category, You're giving role preference rewards for role preference harm. But yeah I do think people should be trying more things in that territory. I remember you, you pointed out something interesting Apex was doing about like wager more ranking points on their matches, I think that's in this category, blurring the lines between ranking and resources.</p>
<cite>Phil:</cite>
<time>26:18</time>
<p>I think that's one alternative approach that we should talk about here because the answer to almost all matchmaking problems is always more players, More players help solve every problem. Rising tides lift all boats, so to speak. And what I really loved about Apex approach to rank design was that you could have a battle royal match. And of course matchmaking is much tougher in Battle Royal when you have 6,100 players. That's huge. Strange on psu rather than, five to 10 players. An Overwatch match is that you are able to blend these players together because they set different bars for a given player to progress when you are blended into a match. So if you were a much higher rank, the bar for you to be able to move forward and. Might have been, let's say, the placing in the top five. Whereas if you were a much lower ranked player, the bar to move forward might be ranking in the top 50%. so you can still have a great experience. There are actually, I think, really interesting consequences to this of okay, what does it mean to have fun in a game? Is it okay that I still get pub stomped as long as I come in the top 50%? Is that okay for me to be able to move forward? Is that just a fun, enough experience? And if I'm a really top ranked player, is it okay that really my ability to move forward is based on like how many noobs I can kill rather than killing people, let's say more fair level skill. But nonetheless, this a whole idea of like wagering points, which is what you have to do. You have to put up your points when you enter, and then you almost get a refund backed based on how well you do. And of course, if your refund exceeds what you originally put in, then you're moving forward. And if it doesn't, then you're moving out. But that's just another way they've thought about being able to move forward as well, is that it's not winner loss. There's a little bit more to this. There's. the margin. Thinking here that I think is really compelling and I haven't seen a lot of games experiment with this.</p>
<cite>Chris:</cite>
<time>28:01</time>
<p>Do you guys think that these like points system so we're talking about almost like a bidding system where you say I want to play this badly. I'm willing to wager this much. Do you think that you've got like coins and Wow that you've gotta buy or, specifically for a game like like Overwatch or Apex or something like that. Or actually like a mobile where you have these really distinct different types of characters and some people wanna play some more than. They have this little coin and they're able to say, I want this one, this bad. It's like the solution to like a line is to just have people pay for their position in the line. So if you've got a thousand people, like the people who want it the most will be willing to pay the most. It's almost it almost goes back to this concept of the land value tax where, it's obviously not a tax, but it's oh who wants it the most? Who's willing to pay the most for it? And just and I feel like these days everything just comes back to an auction. Like you said earlier, Phil, I'm like vomiting in my mouth as I say this, but how does something like that work? Is that a monetization strategy for the company?</p>
<cite>Eric:</cite>
<time>28:59</time>
<p>I think it could be, it's a little risky because then it gets into oh, there's the upper class gets to play damage and all the, all the Venezuelan farmers are all playing support. But I do think that using prices to clear the market is something people haven't tried enough of.</p>
<cite>Phil:</cite>
<time>29:12</time>
<p>Have been thinking about on the supply side, It's the supplier paying you to clear the market versus the player paying you playing to express their preference. That's an interesting dichotomy, but I think you're right, Eric. Of course there'd be revolt</p>
<cite>Chris:</cite>
<time>29:23</time>
<p>what if the points are allocated instead of, you have to buy them, you have to earn them. The better you do in your in your match, the more points you get. The more easily you can buy your place in, in, in the queue, then it's skill-based. I still think people would revolt, but.</p>
<cite>Phil:</cite>
<time>29:38</time>
<p>there's a centripetal force there that I think will drive people nuts because the other consequence of this, if we think about it in terms of roles, is that now I'm gonna be stuck, let's say playing support it's almost like a sub ranking system where like I have do well enough as a support to be able to play mainly, support might just have lower ACA ratio. There might naturally be slower at progression. some sense, I don't know, maybe there's a twist on this where you can make it mechanic. Chris, maybe this is where we go from being economists again to game economists. this is a subsystem where you can move up the chain in terms of roles.</p>
<cite>Eric:</cite>
<time>30:10</time>
<p>I think setting up the roles in kind of a price system can create this feeling that support is the inferior role. That oh, I'm upset that I played support cuz it's what the poor Venezuelan farmers play. It's what the badged players play until they win enough games to play damage. And it can turn a role that they would've been happy playing into one that's seen as like lower cast, so yeah, that is, I think some of the risks of some of these market clearing mechanisms is that it adds this social kind of psychological layer. Phil, you mentioned that large player populations solve most matching making issues, and that is a hundred percent true for skill-based matchmaking. You can have a terrible stats algorithm, as long as you have a large enough sample size, it'll usually work. But for role-based matchmaking, actually it doesn't because it's all percentage based, right? If you know you need one-third damage, one-third support, one-third tank, but you only have 15% supports, right? It doesn't matter if you triple the audience size you're still gonna have that Q time issue.</p>
<cite>Phil:</cite>
<time>31:03</time>
<p>There's also this time series problem, which is really interesting in matchmaking, which is that you have survival bias in which players who tend to be highly skilled, retained better than players who don't. And so over time you end up having a larger and larger pool of highly skilled players. And so when you have a new player that enters the cohort, they might actually face a radically different matchmaking experience because the pool has changed in such a dramatic way. And so in the Apex Legends post, they talk about the fact that at a particular rank, if you were looking at the launch of the game versus that same rank now, so almost what, five years later, that the skill level of that individual player now at that rank is much higher than it used to be just because they've had the survival bias. And so that's another, it seems like wrench in this kind of model. Oh shit, how do I deal with time series problems?</p>
<cite>Chris:</cite>
<time>31:49</time>
<p>Bots,</p>
<cite>Phil:</cite>
<time>31:49</time>
<p>no, Chris don't give into that. Bots are wrong.</p>
<cite>Chris:</cite>
<time>31:52</time>
<p>throw some bad bots in there.</p>
<cite>Phil:</cite>
<time>31:55</time>
<p>Conti and ethic violation</p>
<cite>Eric:</cite>
<time>31:56</time>
<p>The the Koreans have a term for that. It's called it translates to essentially like stale water, like a standing pool of water that's infested with like bugs and algae. It's those games that are so old that everyone playing it is incredibly high skill. And if you go in there, you just get poisoned to die. on the Apex Legends though, one thing that they didn't not talk about at all was Smurfs probably, cause they don't wanna touch that thing with a 10 foot pole. But a lot of the matchmaking adjustments, like I'm looking at'em like, oh, this is actually to deal with Smurfs, but they don't wanna say it</p>
<cite>Phil:</cite>
<time>32:24</time>
<p>Oh, interesting. I didn't think about that in the going through the</p>
<cite>Eric:</cite>
<time>32:27</time>
<p>They use the max of your team skill instead of the average. Or like some, variance adjusted. And that's because the lower level players on your team are likely to be SMUs. And so that's why they use the max.</p>
<cite>Phil:</cite>
<time>32:37</time>
<p>So they're purpose, they're purposely adding, they're purposely creating new accounts to drag down the average.</p>
<cite>Eric:</cite>
<time>32:43</time>
<p>or if you're getting boosted by your friend plays on a pretend bad account, but they're actually good. The other thing was they mentioned that pre-made groups get a rating boost because they, they say because pre-made are more coordinated, but also pre-made are far more likely to contain smu. It creates a really bad experience for honest pre-made actually. Using intra like intergame incentives to try to improve matchmaking through stability. Like the auto field protection, the penalties, that kind of stuff. As well as thinking about making's impact on the meta where if you set up your matchmaking in a certain way, it might encourage or discourage certain roles from being selected which can, have knock on effects.</p>
<cite>Phil:</cite>
<time>33:20</time>
<p>do you think systems designers should take a stronger look at how the matchmaking algorithms are designed? Because the way I see it in most organizations, it's been interesting. It used to be an engineering task, and I've seen it move more and more to the data scientists who now control matchmaking, but the designers seem to always be in another corner and seem to be maybe listening in on what the data doing</p>
<cite>Eric:</cite>
<time>33:41</time>
<p>The problem with the designer led stuff is sometimes, it's like folk art statistics where they like cobble together an algorithm using qualitative descriptions. but I think thinking about the, yeah, the behavioral implications when once it becomes an adversarial search problem and the players are optimizing against the matchmaker sometimes the engineers don't just assume that players will behave honestly.</p>
<cite>Phil:</cite>
<time>34:02</time>
<p>Marvel Snap. It's been living rent free in my head. Second dinner continues to invade my life with this game, but there's more to Marvel Snap than I think just what's going on inside the game. I think there's a lot more going on about what Marvel Snap represents and the questions it poses to us as an industry. I published a post asking why Marvel Snap is struggling to scale. It seemed to have a really great launch, but as with all game launches, it's really hard to know if the launch is because the game is doing really well and it's bringing in a lot of organics. Or is it because the publisher had some sort of minimum guarantee and they went out with a brand marketing or chest and unloaded millions and millions of dollars without the I, the knowledge that they were spending profitably acquire users. Does L T V exceed C P I? That is always the equation that we are. to there is no, there is nothing else. That is it. And so a lot of the times when you look at these game launches, it's unclear if they're spending pro profitably. And when we look at some of the data that has come out of app, a e, or sensor tower, when it comes to Marvel Snap, we can see that downloads have been in complete and utter free fall since they launched. Again. That's a normal thing. Usually if you have some brand marketing war chest, but it continues and continues to fall. They've gone from over 800,000 downloads a day, so I believe, 50,000 or even 10,000 daily. And it's continuing the client, it hasn't reached equilibrium yet. This is concerning stuff. This is not what you want to see from a game that at least in my opinion, should be a billion dollar a year franchise, or at least a half a billion dollar a year f. Again, the benchmark, I would argue for Marvel Snap isn't these other CCGs, which is nonexistent on mobile. There are no cgs CCGs that matter in mobile. It really needs to be go after Clash Royale. You know what? You might consider a card battler, but putting all those genre labels aside, it doesn't matter what matters as profit. Why is Marvel Snap not str? Why is Marvel Snap not scaling in the way it should? And so what I wanted to argue in this piece is that there's this problem with the horizontal progression. And so what you tend to find in the games with horizontal progression in which you aren't necessarily getting more and more powerful, but instead what you're doing is you're collecting more and more, let's say, different experiences or different things to master. think about getting a new operator in Rainbow Six Siege. You have to master their abilities. You have to master how they move, how they fire, how they work with other players. The design intent at least is not that one operator is more powerful than another. And so in many card games, this is also how they tend to design. So you'll have a new card or you'll have a new expansion set. New expansion set or card isn't meant to be more powerful than another card. And if you look at a lot of the magic, the Gathering Head Designs talks at GC explicitly talks about this. Each card is supposed to have some sort of unique and understood purpose, supposed to be horizontal progression. They offer something different, not necessarily better or not as something worse. And so this is something that Marvel Snap incorporates into their game design. Each card has its own place, its own purpose, not necessarily better or worse than another card. That is pretty radical departure from where we've been in mobile game design, which is that players tend to have more and more vertical progression. And a lot of this is due to monetization. Players like to win. And so if I offer vertical progression, usually I can have some sort of tie in to monetization. But what I'd also argue, and this is the most important part of the piece, is that vertical progression also guarantees that a player could just invest more time into the game. If a player invests more and more time into the game, they can progress without having to solve the puzzle that's presented to them. So in Marvel Snap, there's a ranked mode, and for me to be able to move forward in that ranked mode, obviously I need to win. But for me to be able to win at successive stages means I need to solve the meta at those successive stages. So a player having to solve that is a huge cognitive task. It is incredibly draining for a player to solve that. It requires a lot of complex thinking. They might even have to go on the internet to figure out what decks are in the meta. That's not something that scales to millions and millions of players. Maybe for people who play on Reddit, maybe for those hardcore gamers, but not for the mass general audience. And when you compare that to a game like clash Royal has cards, again, you deploy them, you build decks. But what happens in Clash Royalle is that you open loop boxes. When you open those loop boxes, you get copies of cards. when you get enough copies of a particular card, you can upgrade it and when you upgrade it, that card gets more health. It might get more at attack speed. It gets a variety of things that make the card better off without not making, without making anything worse off with that particular card. And so in Class Rail, even if I'm not good at solving the meta, even if I'm bad at deck formation, if I play long enough, I will open enough loop boxes and I will upgrade the cards that are in my deck to a high enough power level in which I'm guaranteed to progress, or my probability of winning any particular match is near 100%. Now of course, you're gonna reach equilibrium at some point in which you're gonna come back down to 50%, in which case you'll need to open more loop boxes and need to continue to upgrade your cards. But again, there is a positive relationship between the amount of time I spend in Class Rail and how I progress in the ranked mode. No such relationship like that exists in Marvel Snap. If you continue to play Marvel Snap, there is no guarantee that your win probability will increase. And this to me is a huge problem. And when I think about a lot of the empirical evidence we have around this, it seems to suggest that without vertical progression, or excuse me, without the ability for players to progress, without any sort of guarantee that a player will progress, they will churn. And so I find this to be one of the reasons that I think Marvel Snap is struggling. Now, I wouldn't suggest that Marvel Snap moved to this paradigm of incorporating vertical progression. It's impossible. It's impossible for Marvel Snap to do it. When you have a card that has, let's say one health and one defense, it's very hard to micro-dose progression into that. So clash reality, you might have a hundred health and a hundred defense. So if I want to increase that by 10%, it's very easy to add those units to the card. Whereas Marvel Snap, there's no decimal points. there's the integers are very low. So there's almost this Nu Am numeracy problem, and only that there's a stats. Marvel Snap Card only has two stats, whereas Clash Rail, a card might have, 10 stats, potentially a attack speed, health defense, armor, resistance, magic resistance, all these different things. And so I think Marvel Snap is really at a crossroads for how they solve this. Second dinner has come out with an in compelling game design. It completely changed the mobile design paradigm. Their HD designers, they said, screw, screw what's going on in mobile? I don't care about match three, I don't care about four x. Fuck that. I'm gonna make an innovative game. And they went out and they did that. But I think this is really the revenge of the paradigm. This is where they're starting to struggle, and I don't quite know what the solution is. I think the best I could offer is the same one I offered beforehand. When it comes to monetization, double down a horizontal progression, create different experiences in which different players succeed. That to me is one way to expand your player base, is let them progress in different types of mode. And to me, the best way to do this is to incorporate drafting or for players to be able to enter a mode, they pay a fee. And in this mode, they are presented with a unique collection of cards and they form a deck on the spot, and they will take that deck through a series of matches. At some point, when they lose enough matches, they will be out of the tournament. But you can continue to do this. It's an infinite money sink and it's something new to master almost every single time a player steps into it. This is the best I can offer for Marvel Snap, but to me, no vertical progression means churn. And churn means your player base is not gonna grow. I think this is a big problem for Marvel. Snap.</p>
<cite>Eric:</cite>
<time>41:34</time>
<p>So on the power creep note, they could power creep, right? They could print new cards. I guess that's a lot of content to create and also break some of the trust, but</p>
<cite>Phil:</cite>
<time>41:42</time>
<p>if you created new cards, are you saying like the way you would imagine that is Marvel stamp creates new cards? The cards are more powerful, but here's the thing. If you do that, you still have to make sure that the mainline progression is giving you more and more powerful cards as you move up the mainline progression. so if the card you get at each level is random and I add more powerful cards to the pool, that doesn't solve the problem, right? Because I need to be getting increasingly more powerful as a time series if that makes</p>
<cite>Eric:</cite>
<time>42:14</time>
<p>So magic does that, Where I know they say they have totally horizontal design, but no, there's just cards that are straight up better than other cards. And they justify it with drafting, but. You get random packs and the average, your top 60 cards on average goes up as you buy more packs. but yeah, to your point, like the numbers are so small and they few attributes and they can't have complicated rules, text as complicated as magic. So it's hard for them to really creep up.</p>
<cite>Chris:</cite>
<time>42:38</time>
<p>Royal Claro I'll be honest, I've not played Clash Royal very much at all. They are more heavy in stats, less heavy in car text. And Marvel Snap is more heavy in car text, less heavy in stats. Is that fair?</p>
<cite>Phil:</cite>
<time>42:50</time>
<p>a very that's a very interesting way to put it. Yes</p>
<cite>Chris:</cite>
<time>42:52</time>
<p>It's like when you're trying to do this choppy, fun, exciting, explosive gameplay, most people don't wanna have to read a lot of effect text. And it sounds like the way that they can go wide, the way that they can or, go, add some verticality to the play styles and to the cards requires them to get very complicated with the card text because they already have some really unique, interesting, creative effects from in their card design right now. So it's like you just have to make it more complicated. I played Yugi o the other day on Xbox, and I was like, every card has two paragraphs of thick text. It's untenable.</p>
<cite>Phil:</cite>
<time>43:28</time>
<p>So would your solution to be just like double down on something new to master, create more and more depth in terms of the core gameplay so that there's more content almost per card.</p>
<cite>Chris:</cite>
<time>43:38</time>
<p>I haven't churned yet, but I wonder if that's because I like Tcgs I like to play CCGs and Tcgs. I just like cards. I wanna ask somebody who, I don't like, not all of my friends, like my friend group is pretty heavy in trading card games. We, we all to play competitive card games. So I wonder I wonder what it feels like when they churn. What does that feel like for a casual gamer, for somebody who's not really like a hardcore card game player?</p>
<cite>Phil:</cite>
<time>44:02</time>
<p>So let's put it this way, Chris. When you play Marvel Snap, you can only play the ranked mode right now. What rank are you?</p>
<cite>Chris:</cite>
<time>44:08</time>
<p>That's a good question.</p>
<cite>Eric:</cite>
<time>44:10</time>
<p>The rank check</p>
<cite>Phil:</cite>
<time>44:11</time>
<p>Cuz so here's what's gonna happen, Chris, right? At some point you're gonna get to a place where you're gonna reach equilibrium in. let's say it's a rank 30, to be where the bots start to decline in Marvel Snap. So that place it's a true P game. So you're gonna get to that place. And when you're at that place, you're presented with an option, So you can continue with the strategy you've been using and the strategy you've been using will continue to put you at rank 30. Or what you can do is you can start to make changes to And that's what most players are gonna do when their win weight starts to drop or goes to 50%, how do I progress? And so that's the puzzle. And at that point that's the exact point I'm talking about here is oh shit, I need to change my deck. So what do I need to do to win? And that is a very challenging puzzle And to me, there are very few people that are up to that task of being able to solve that puzzle. And especially for a mobile game that's trying to hit the scale they're trying to hit, I think that's gonna be a huge problem. It is a</p>
<cite>Chris:</cite>
<time>45:07</time>
<p>Yeah I have not lost a game yet. And I'm so I'm 15 yeah, I haven't. And I like literally downloaded the mobile version like a couple days ago and that's when I was like, oh shit, I gotta play this. But know, it's like I spend a lot of time in the menus looking at cards, messing with decks. I don't think, I think you're completely right. I don't think the regular player does that. I don't even think they know that they can adjust their deck. And quite frankly, like in I've played, I've actually played through the beginning twice. I created two at different accounts on accident. And I was, both times I was frustrated with the way that they were stepping me through oh, go into your deck and swap this card out. And I was just like, I don't care. Like I, I just wanna keep playing. And that would be interesting to see how that once I do actually need to adjust my deck like how I like that. It sounds like people are churning out before they reach those really competitive ranks.</p>
<cite>Phil:</cite>
<time>45:53</time>
<p>I don't know if we know that yet. I don't think we know when people are experiencing churn. My thesis is that people are experiencing churn. When they get to the place the bots are gone, they reach equilibrium and they're not quite sure how to solve the puzzle because that, that, that's the place where like their core, their mainline progression really comes into play. And so if you remember in Snap, the way you progress is that you end up playing, what is it? How do you end up moving? You move up your collection level by upgrading, by frame, breaking your cards, and you frame, break your cards by getting more boosters, which let you frame break, That is their core progression loop. And the way you can solve that puzzle of you being at rank 30 and not being able to progress is continuing to grind and continue to play these matches. And then you get new cards and then you it's let's put this way it's almost like a Bayesian approach, right? Like you have your priors and then you get new information, which is a new card, and you try to update your deck with this new information and then you test it, Whether or not your theory works, and so you're constantly tweaking it. It's exactly the loop we were talking about in super auto pets, right? You're trying to build these berg machines and see if they succeed. I just think it's brutal. It's just a brutal fucking process. In a PVP environment.</p>
<cite>Eric:</cite>
<time>46:59</time>
<p>Do you think this is a failure of the deck building experience? Comparing again to Clash Royal. Clash Royal has like what, 30 cards and you have an eight card deck, so it's much quicker and simpler to build. Whereas in if it's full horizontal, like I'm sure Snap has a hundred cards, right? And if the deck and deck building on mobile, looking at a hundred different cards in your collection on mobile seems awful, Like maybe it's that deck building to try to improve that part of the loop is broken.</p>
<cite>Phil:</cite>
<time>47:23</time>
<p>So I think that's a really interesting point you make, Eric, is that crafting a new deck is expensive. Testing a new theory. is expensive. Testing a new theory about, okay, I think I'm gonna win if I include this card in my deck. It requires a lot of trial and error cuz you have to get rid of the randomness, right? As n goes to infinity, you can reduce that error rate of, okay, this is just a random match that isn't representative of the long run win value. Because that's what you're trying to understand is the current deck I have in the long run is gonna have more than a 50% win rate, right? You want to milk that for as long as you can because at some point it's not going to at some point you're gonna reach a new place in your ranked progression where you're back down to 50% in which you have to solve the puzzle again. So you get that new card, you're incorporating it into your deck, and you're saying, okay, now I have a greater than 50% win rate. I need to keep playing. I need to keep playing. So the question is like, how can I reduce the cost of that learning loop of whether or not this deck is gonna work or whether or not this deck isn't going to work. I think sometimes the answer is just in better deck building ux. How can we make it as easy as possible for players to select new cards and incorporate them into their deck? I think the other opportunity here is perhaps some sort of suggested new card or maybe a bonus going back to our matchmaking con conversation. Maybe there should be a bonus for a player including a new card into their deck. Perhaps it's a one-time XP boost if you use it in the deck. But there has to be better ways for this. I don't, the presets are cognitively draining from a UX perspective. Maybe it's a pre-made deck. Maybe it's the opportunity to copy and paste decks from the internet I into your pre-made deck set. We've seen, I believe it was Hearthstone and Magic the other arena, make it easy for you to go online, copy it, deck, paste it somewhere. So I think to me, like a lot of it comes down to UX and incentives. Again, like how can we make experimentation cheaper for players? How can we make solving that puzzle less expensive and less draining, if that makes sense.</p>
<cite>Eric:</cite>
<time>49:14</time>
<p>You mentioned the only mode is ranked</p>
<cite>Phil:</cite>
<time>49:16</time>
<p>They only have ranked,</p>
<cite>Eric:</cite>
<time>49:17</time>
<p>So there's no safe space</p>
<cite>Chris:</cite>
<time>49:19</time>
<p>It's super uncomfortable for somebody like me with a traditional background in, in trading card games. I seldom play in ranked unless I'm like, it's it's 9:00 PM my day is over, I'm locked in. I'm like, okay, it's time to sit down and play some ranked for real serious for an hour and a half or two. It's so uncomfortable when I'm on my phone and it's like, like, do you wanna play a ranked game? No. It's just my stupid little phone. I need to be in front of my computer to play a ranked game. But I don't think though that the a, I don't think the average casual gamer cares or notices about that. They're winning every single game anyway. So it feels like they're playing the bots that they would normally play in an unranked game anyway.</p>
<cite>Phil:</cite>
<time>49:54</time>
<p>I do have one fucked up theory, Eric, this is the most fucked up theory I have about this game and how Ben bro designed mainline progression. But if you were, if you remember how the collection level works is that you get a mystery guard. So your progression is almost random. Now, they tried to segment this on the back end with pools. So technically there is some pool in which you're drawing from, which has a certain collection of cards. So they do have some level of control on the type of cards that you're having at what particular collection level. But it's extremely broad. and you can't pay, you can't pay to get more of these cards. It's extremely limited in your ability to acquire new cards by And so my fucked up theory for this is that Ben Broad wanted to prevent people from going onto the internet, copying a necklace that has the highest win rate, importing it into their phone. I think what he wanted players to do is to be able to truly craft a new deck from the collection they have independently without looking on the internet. He wanted this to be a part of and by randomizing the cards in which the order in which you're getting the cards, you're forced to continually solve this deck building puzzle. Again, I think it's just fucking brutal</p>
<cite>Eric:</cite>
<time>50:58</time>
<p>It's a card game player designing what he thinks the casuals want.</p>
<cite>Phil:</cite>
<time>51:02</time>
<p>Yep I would that</p>
<cite>Eric:</cite>
<time>51:03</time>
<p>I think the drafting idea you had is, I love drafting games. Yeah. I love super auto pets. I love drafting and magic and asynchronous digital drafting, I think, fits the mobile paradigm pretty well.</p>
<cite>Phil:</cite>
<time>51:12</time>
<p>And we don't have to deal with matchmaking anymore. Can avoid that whole conversation. You just it's Bob's next</p>
<cite>Chris:</cite>
<time>51:18</time>
<p>what about a model where there are, there, there's 300 cards or 200 cards, and every game is how many cards are in each deck? 10. Like nine cards. Eight. 8, 0, 18. Okay. It's a little more than I thought.</p>
<cite>Phil:</cite>
<time>51:30</time>
<p>Yeah, I think it's 18.</p>
<cite>Chris:</cite>
<time>51:32</time>
<p>There's six turns. You draw one card per turn, right? You start with five cards in your hand. How could there possibly be? Unless you don't draw your whole, you don't draw your whole deck</p>
<cite>Phil:</cite>
<time>51:41</time>
<p>don't enjoy your whole deck.</p>
<cite>Chris:</cite>
<time>51:42</time>
<p>Okay. But like a mini draft so you don't have to do that deck because the, I think the, like the critical component that we've identified is the deck building part of this. That's the part where people once they hit the point where they have to do the deck building, maybe that's where there's a lot of, churn potentially that we're that's a hypothesis. How do you drafting is not deck building. I mean it's deck building. Sorry. It's not deck construction. It's not constructed deck building. It's drafted deck building. Is that like a, is that a gameplay mode that we think would help to, in improve churn or improve retention? Because now I have this outside option to go play the draft instead of the, I'm sure they're working on draft.</p>
<cite>Phil:</cite>
<time>52:21</time>
<p>I think so that's what I, that's what I want to see, not just for the reasons that you identified. I think again, it's a safe space for you to explore new cards. In fact, you're actually forced to explore new cards when you go into a draft. And to me, like the third reason to include drafting is that it's a limitless money sink. you can go into drafting and because you're having an experience in which you're paying, let's say two to$3 per draft run, you can continue to press that two to$3 button over and over again. And that's something that I've pressed over and over again in matching the gathering I go on arena runs, it's extremely compelling for me. I the puzzle solving and the creativity that comes outta that. Cuz you can't go on the internet. You can't go on the internet really and solve this. Now there are actually some tools</p>
<cite>Chris:</cite>
<time>53:04</time>
<p>Yep.</p>
<cite>Phil:</cite>
<time>53:04</time>
<p>and some people try to do it, but it's much more difficult.</p>
<cite>Eric:</cite>
<time>53:07</time>
<p>You don't have that analysis. Like with deck building, you just hit a brick wall instantly. I've got 50 cards, I gotta pick 18 of them. What do I do? You freeze. Drafting is here's three, pick one. If you want to go to the heart stone route. So it's way simpler to make those decisions, keep them puttering along until they finish the deck.</p>
<cite>Phil:</cite>
<time>53:23</time>
<p>All Right. That's episode four in the Can Game Economist cast. Peace out all right, peace. I'll talk to</p>
<cite>Chris:</cite>
<time>53:29</time>
<p>See you guys.</p>
