Game Economist Cast

E28: B-A-N-A-N-A!

Phillip Black

Send us a text

Play to earn hits Steam, and the crew is here to dissect the phenomenon. Why now, and why bananas? 

Phil is back into the economics of social casinos, while Chris develops a progression model for Darts. Eric and Phil debate the externalities of bots, while Chris solves for equilibrium. 

Speaker 1:

I've seen so few actual faces on. Web3. I think they're all dogs.

Speaker 2:

Yeah, the only reason we don't have our visual on is because, I think, because of the limits of Discord. But I've gone on with my face on multiple different community events and stuff like that. I'm not. You look at my Twitter account. I look like some dude. I could just as easily be a whiskey influencer as a game economist.

Speaker 3:

Is this your?

Speaker 2:

alternate life. For a while I did reach out to a few whiskey distilleries and be like hey, you guys.

Speaker 3:

How many whiskey influencers do you follow?

Speaker 2:

I don't because I find them annoying, but I know of a few and they're all. Anyway, I don't want to talk about whiskey influencers on this podcast.

Speaker 3:

Let's start with utility.

Speaker 2:

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. In fact, there may be a fundamental problem in modeling that we don't want to model. Economist Cast, episode 29. We are in video in HD 4K maybe not 4K. Bits are cheap, marginal costs are down and we have three wonderful topics to talk about. I know everyone's just been burning to talk about the banana game. Eric, you had an amazing piece on the banana game on your sub stack.

Speaker 3:

Yeah, I mean it's over. They're down Banana's number four on Steam right now. Does anybody know what the market for banana looks like?

Speaker 1:

I don't Where's transaction volume for Steam. That's what I haven't seen. I don't know if you can.

Speaker 3:

Yeah, I had to estimate it off of something that sounds good.

Speaker 1:

I'm joined today by my two other regular co-hosts. Eric, how are you?

Speaker 3:

Doing alright. Been at second dinner for about a month now. I'm enjoying it.

Speaker 1:

Past the aperitifs. I don't know, is it a meal like that? That is, aperitifs before or after. That's an appetizer appetizers yeah, dj steve is just after. Drink after dinner. Drink the honeymoon period.

Speaker 3:

Yeah, it's definitely still honeymoon period, like no one's like demanding deliverables for me yet, but definitely enjoying it so far. It's nice to work on a game that you actually enjoy playing and think about a lot in your free time do you just walk around thinking about marvel snap?

Speaker 3:

yeah, like, how do you define like character archetypes in a card game? Right, you have to like cluster decks together, but then is there any true definition for what is a discard deck? Because it keeps changing all the time and anyway I do a whole rant on that. But, yeah, a ton of data problems and it's great that I think this is like my sweet spot for the type of game I like to work on, which is a game that's live, it's got enough players and activity that there's real data to work with. It's an interesting competitive game with enough stuff in the game itself to analyze and also the company's not so big that it's full of bureaucracy and political bullshit.

Speaker 2:

How is it compared to working in Web3? Is it a bit of a? You worked, obviously on League of Legends, so you're probably used to it.

Speaker 3:

The huge difference for me compared to Web3 is that there's actually a product that's live and I can look at the data for it. Chris, you've got a real product, but most of these things it's 99% marketing. And did the price go up? What made the price go up? There's no product yet.

Speaker 2:

There's no product yet, there's no users yet you got what? Like a thousand users on sandbox. That's not like fucking static for anything. That's good to hear, just they're just dreading when you're going to be like you know what we should do with these cards. We should make a marketplace for them.

Speaker 3:

There's actually a decent number of ex web three people who are just like yeah, I'm over it.

Speaker 1:

Okay, chris, how are you?

Speaker 2:

We have just had our second star Atlas Economic Forum, which is basically something that the cool thing about being a pioneer in an industry, regardless of whether that industry will be here in 10 years or not is that you get to invent new stuff. I found that directly addressing the community and like having these online discussions was just not a sustainable practice for me in particular sustainable practice for me in particular. So we basically decided to do like a monthly community chat almost a fireside chat, but I'll have a presentation where we go over changes that we made, updates to the game, any like patches that we did, and like the overall what's going on, what happened this quarter or this month, and then a q&a session with the community and it's been a big success so far. Highly recommend any web three people who are listening. If you are struggling to manage your community from the economic side, just allow them to get into a room and talk to them like they're humans Half the time.

Speaker 2:

That's what people want is. They want to be heard. So it's been a huge success and I've been having a lot of fun. Had that just before this. Yeah, things are good.

Speaker 1:

What do the people want to know about? I guess this is like a quarterly earnings call. What do the investors want to know from you?

Speaker 2:

Whoa whoa. I don't like that language.

Speaker 1:

They forgot there's lawyers everywhere.

Speaker 2:

What do the players want to know?

Speaker 1:

about? What do they want to know about their assets?

Speaker 2:

in the future? No, it's not. The players want to know. So typically the questions will be like what's going to happen in the future, or are you going to make a change to this thing that I think is broken? So half of the reason we got this forum together was to try to better aggregate opinion, because, especially in a project like this, where if 50% of the player population dumps their assets, it's not good for the overall health of the project. We want to support people and we want to make sure that we're addressing concerns, because it's important to address concerns, but if you have one person complaining about one thing, that's not that important.

Speaker 2:

So, for example, people will ask forward looking questions hey, chris, once all the star bases are upgraded in the game and we lose this, once the crafting loop is exhausted, because it's all on chain stuff, we can't just reset it, which is a very difficult limitation to deal with. What are you going to do? And so I'll answer that. So sometimes it's a little bit more theory, crafty like that. And then sometimes it's hey, this mechanic is broken, it's killing the small players, what are you going to do about it? And I can either answer that or we can go over. Hey, we made this change last year or last month and so it's mostly like I said, it's people wanting to hear, get their voices heard. Typically, most of the comments, especially when you're face to face, most of the discourse is very positive and it's very respectful. It's a lot easier for discourse to become like ugly when you're just behind the keyboard.

Speaker 1:

We have two wonderful topics to talk about today. The first PvP bots. We have a lot of takes. This was sparked by Eric's substack, which is on fire, causing so much controversy right now, and also internal competition gone wrong, where you're going to take the traditional game economist shit. On pms now, we love pms, it's let's just say we're skeptical of some of their methods and sometimes there are bad incentives that come out of that. But before we do, let's talk about the banana games we've been playing got a selection of good things on sale.

Speaker 3:

Stranger banana how many bananas you guys? I literally have not got a fucking banana yet.

Speaker 2:

And I've been clicking in the background. I've been clicking. You just leave it open and it drops a banana every couple hours.

Speaker 3:

How do you check, if you?

Speaker 2:

have a banana. Does it say you've got a banana? No, the game tag gives you absolutely no feedback. Okay, so maybe I do have bananas, it but yeah, I guess.

Speaker 3:

To summarize I think Chris put it really well this is the most minimalist, reduced form of a play to earn game possible. So banana, as you may have seen, topping the steam charts. That was like number one for a week, above Counter-Strike and Dota. It's a picture of a banana and you can click on the banana and there's a number above it that goes up. That's it. But it's not really a game. Really. What it does is, if you leave bananaexe running on your computer in the background, every few hours it'll drop a tradable banana into your Steam inventory which you can go to the marketplace and speculate on. Most of them are common and worthless. Most of them sell for one cent, but some of them were selling for like $100 or so.

Speaker 3:

This game was trending on Steam because of the way they manipulate their stats and it was the number one game for a while drew a bunch of attention, all the sort of gaming influencers talking shit on it. A couple copycats are out there. One of them, the most successful one, is called Cats. But yeah, that's it. It's a game, it's an executable. That's not fun at all to play, gives you tradables and then people speculate and you can stimulate this trading market.

Speaker 2:

If anyone's looking for a screenshot of a 10,000 banana, please just let me know. I'm selling it for only $1 per click. $10,000. That's all I ask. Just throw it out there.

Speaker 3:

Was this labor theory of value? We don't believe on that here.

Speaker 1:

The thing that I was most impressed by in your piece, eric, is how you got across the fact that this is actually just a bad game. It's not even a good clicker game. There's zero feedback to clicking. You don't even have any haptic feedback, while you even have haptic on pc. But it's not like the screen shakes, there's almost no celebration moment, the number just goes up. Like this is something you could have put together. This is like the napoleon dynamite of games. Actually that's not fair, that's not fair at all, but it's just. It's literally just a number that goes up.

Speaker 1:

And I think the confusion is like why did it take this long for play to earn economics and like this animal spirit, to reach steam? Like this should have happened a long time ago. Like we see this all the time in web 3 and like the difference between this and web 3 is that you can take the banana but you can't exit the money from steam. You can get steam credit and steam credit ultimately can be used to redeem against steam games, but you can't get real cash. So that is an important differentiation. And their price caps on the Steam marketplace. I think that's the other important point. There's also no ability to circumvent it. I guess you can do off-platform trading, that's true. So that is true. You could avoid royalties like you can in Web3. That is an issue, but I guess I'm just confused on why now. Why is this blowing up now? On why now, why is this blowing up now?

Speaker 2:

Honestly. I think it's because play to earn is finally at least in the Web3 context, has finally fallen out of the mainstream rhetoric. Nobody cares about it. I think if you had tried to do something like this, maybe at the heat of the moment, maybe it wouldn't have been as successful. I'm not surprised it took this long. I think this is a horrible bid in the play to earn hitting the mainstream, but it's a really interesting experiment.

Speaker 2:

I, like Eric alluded to, I think that this is a necessary experiment.

Speaker 2:

It's not something that's going to, it's not necessarily good for the image of play to earn and it's not as just like a product that's probably going to have some staying power. But for me, I literally have thought of banana in my head as what is the minimum like viable product that you could call a play to earn game. And I think banana is that you, you have to put in effort, minimal effort, but you have to put an effort and then you get this random reward and I think the randomness is really critical to this game success. Right, if you were, if, let's say, you were building up these collections, I don't think it would be anywhere near as exciting and I definitely don't think that the prices would be as high as they are because they're delivering random items. There's like kind of this gambling aspect to it. That's what gives it its oomph, and I think I'm starting to think that this kind of randomness, excitement is important for these types of games, some sort of surprise and suspense, as Eric has talked about in the past.

Speaker 3:

Yeah, I try to calculate the payout rate and it's under one cent per hour. But if you go to the marketplace you look there's oh, there's $100 banana right. So there's this belief that there's a small chance I hit the jackpot, but you know, might as well keep it running.

Speaker 2:

Which, when you consider how stripped down it is, what happens when you make it more interesting. What happens when, instead of this being banana, it's runescape. Does that per dollar value or per hour value increase? Does it increase to 10 cents an hour? Does it increase to a dollar an hour? And then you start to get into the territory of, oh, actually, this could be a real market. So that's why I think it's valuable to your question, phil.

Speaker 3:

Why did it catch on? A lot of it is memetics. Like just this was super memeable. First of all, the experience playing the game. You're just like shocked at how bad of a game it is, like it makes you want to tell people about it and the other. And also bananas are funny monkeys like bananas, they look like penises. It's great meme material.

Speaker 3:

But I think on top of all that they hacked steam's user base charts. Like game. To get the bananas you have to leave it running in your background, which boosts the game's active player count. People like me would check Steam. Oh, what are the top games right now? Banana, what's that? Click on it and I think by hacking the engagement stats they get to the top of the charts and get a bunch of attention, similar to how other games will hack, like Dapp, radar or other Web3 metrics to try to draw attention. Doing wash trading. Yeah, I think that's a big part of banana success. Call it success. It's pretty much dead now. People are abandoning the game. But there were previous games that did this exact same thing. There was one called Egg that this is a blatant ripoff of. That was just the exact same thing, just eggs instead of bananas and that game did not do nearly the same degree of success and I think it's all the virality aspect the same degree of success.

Speaker 1:

And I think it's all the virality aspect that makes sense to me. I buy that. I just don't understand. I guess what I want to get to is like how we can control the animal spirits Like I. How can we direct this idea of speculation and get it to work for us, Because right now it just takes off that these bananas are worth anything, and they're worth anything because other people think they might be worth something. It's it is a bubble. It's a classic definition of a bubble. We've seen this reproduced in labs, but we don't know how to control it yet and ultimately, like how to use it to our own benefit. Yet this is what I think Web3 has struggled with is like how to actually capture the influx of people.

Speaker 2:

How do you know Bonk is going to be a success and Bonk 2 is going to be a success? It's very difficult to know success. It's very difficult to know you could come up with a mean coin. It doesn't mean that it's gonna take off just because it's memeable.

Speaker 3:

Do you think it's fundamentally unreproducible? Like it might just be a chaotic, random process that, like you, cannot replicate the success there's got to be?

Speaker 1:

indicators right there's already cucumber, though let's put it this way are we predicting that there'll be more of these? Now on steam, are there going to be these random animal spirit hype cycle games that come out like what's our model for understanding this? I'm sure periodically there will be a game with a trading economy that pops off there there has been like there are every game that I think has a steam marketplace could go through. I don't know. Do they go through periods of speculation?

Speaker 2:

I don't know, sure, yes, goes market, different game.

Speaker 3:

I've had moment. Yeah, I think those legacy crates are trading for crazy prices.

Speaker 2:

I think the best case scenario and in my opinion the likely scenario is that you end up with somebody who sees the virality of the prices and tries to replicate that part, but hopefully not the gameplay component. So they say, oh, I'm going to release like Power World 2.0, but it's going to have like random drops that are going to be sellable on a marketplace, something like that, where it actually has gameplay. So they'll try and copy one part, but they hopefully won't try and copy the other part, because what worked about Banana Like I could be wrong, but I think it's the Bananas, it's the random Bananas.

Speaker 1:

Eric, you're like our resident memeologist. There is an aspect of internet culture here that I think like this captures in a very interesting way. But then I want to see, like the waifu version, like why, or I just imagine how many ways you can play in esteem like why can't I, why can't you get like those girlfriend pillows for those girlfriend files, and imagine someone making like some waifu version of this. I just I want to capture that. What is the?

Speaker 3:

waifu material way that I'm. What is that? It's like you see those gacha games. Oh for the uninitiated waifu basically just means like a hot anime girl that is sold in these mobile games. Yes or yes, kitchen impact now that I've seen it.

Speaker 1:

Definitely, I've seen that there is some footage on twitter about a waifu third-person shooter and when you dodge a bullet, you actually become like 2d to dodge a bullet, which is really cool that's cool like not. There was an announcement that the anime horse racing game from Japan is coming to America, so I'm all about this right now. This is a game that's generated billions of dollars in Japan. You are a. It's a horse girl. She has a tail.

Speaker 2:

Anime horse racing girl.

Speaker 1:

It's a normal body. There's an anime tail.

Speaker 2:

Okay, pretty standard. There's an anime tail okay, pretty standard and you collect them.

Speaker 1:

It's very similar to pocket.

Speaker 3:

It's a horse body with a. It's anime girl's tail.

Speaker 1:

It's a human, human, it's a human body with a horse as a tail, a horse tail does it have horse legs?

Speaker 3:

no human body.

Speaker 2:

So she's running on all four?

Speaker 1:

no, she's running on twos like a normal person. She just have to have a tail and furry ears. But it's clearly so. They're not centaurs. Okay, now that we've cleared this trend of anime?

Speaker 3:

why, like the anthropomorphization of anything? All right, there's that like tank game where all like the tanks and planes and like battle cruisers and stuff are just anime girls. Right, there's like the nike, which, like the guns, are like anime girls and they always have the mobile ads where, like, they show the jiggle physics on their boobs, like, yeah, I don't know, it's definitely a thing, they're trying to take everything and turn it into hot anime girls and yeah, I don't know, there's a lot of space to plumb there, I suppose.

Speaker 1:

So uh we're not ready for linkedin banana and jiggle mechanics. That's what you're saying this is not combined banana with jiggle mechanics and we're not streaming on linkedin anytime soon we're definitely an x show, for sure.

Speaker 2:

yeah, we, yeah, we're not professionals.

Speaker 1:

We're more Joe Rogan than Scott Galloway. Eric, what are you even playing?

Speaker 3:

I just want to say my daughter beat Yoshi's Crafted World with a little bit of assistance from me, but she completed her first video game, which I'm very proud of her, and as a reward I bought her another video game. That's how incentives work. She wanted to buy Kirby and I was like game, that's how incentives work. Yeah, she wanted to buy kirby and I was like you got to beat yoshi before I'm going to buy you another game. Oh wow, I was super motivated to beat you finish your food basically it's a good lesson.

Speaker 2:

She'll never be one of those people with a 10 000 game steam library yeah, there's just an interesting report about like how much that's generated in sales.

Speaker 1:

Was it like 13 billion? I saw in the report lifetime sales of games that have not been opened yet oh right players it's, but this is the whole point of the sale.

Speaker 1:

Strategy is to be able to generate behavior like this. The steam has turned game collection into meta collection. Like one of the things that phil spencer, who is the head of xbox, said recently was that they missed out on the most important generation, which was around the 360 generation to the the4 generation. That's when a lot of people built quote their digital libraries. That's what he says. It's as though you're building a library and if that library is split across different platforms, it doesn't feel as conclusive, and all of these things generally have backwards compatibility. So over time, you're building a digital catalog for yourself, and Steam has mastered this like no other, in my opinion. I like it. Everyone wants to have their games on Steam. Like they were able to bring Blizzard back into the fold, they were able to bring EA back into the fold. You can play Apex on the platform. It is impressive what they've done. Being a monopolist is fucking great. Monopoly power is awesome. Like owning distribution is great. Like they have user demand, they own distribution.

Speaker 2:

That's all you need yeah, and people want to have their stuff like they want to have it in one spot. I think that's a. It's something that people scoff at when they think about web 3, but it's true chris, what have you been playing? Yeah, in usual fashion, I have been playing darts, real darts like not not video game darts man, I love this group.

Speaker 1:

You're not gonna get that content anywhere else. I've been playing, actual I've been playing. I'd give us the scintillating thesis so what's wrong with that?

Speaker 2:

I think darts is a perfect game. I think it's a beautiful game. The only thing that I think is funny. And so I wrote a piece on darts because I was bored last week and didn't have a good topic. So I decided to write about darts and what I did was I just I took every single half split. So like I tried to think about what's the easiest target you can hit on the dartboard. That's not the whole dartboard, one half of the dartboard. So I said, okay, let's assume that you have enough aim to hit one half of the dartboard and not the other half, but you don't have any more aim than that.

Speaker 2:

So this is like a very rudimentary version of a really fancy physics paper that are physics article that somebody wrote where they were like, you have your like skill distribution, your mean and your standard deviation and here's where you should aim if you're really good, anyway. But I basically said, okay, which part of the dartboard is the most value? And I just ran a Python loop to add up the different sides and you would think as like a game designer, you would think, oh, I'm sure that if you were to split the dartboard into halves, every half is worth the same amount of points Like a die. The opposite the two opposite ends, are always going to add up to seven. I think it's seven. Yeah, you would expect the same type of thing for a dartboard, but it's actually not the case. The left side of the dartboard has a higher almost a one point higher average score than the right side of the dartboard. And if you're aiming for the highest point target on the board, which is the triple 20, the two p, the two panels that are right next to adjacent on both sides of the 20, are the one and the five.

Speaker 2:

I'm very new to this. If I got that wrong, one of them is the one, I think with the other one's the five. So you've got these super low scoring tiles that are right next to the high scoring tiles. So it's a really interesting thing.

Speaker 2:

Where there's this phenomenon, where the better. So if you're really bad and you just randomly hit the board, you're going to get an average score of 11 every single throw. If you hit the board versus if you aim for the triple 20, which is the highest scoring place on the board, and you're really bad, you're going to get an 11 because you're randomly hitting the whole board. If your precision gets better, you're going to start getting closer and closer to that 20, but you're going to start hitting more ones and fives. You just got to really drop your average score to nine and a half. So you end up with this situation where you're really good and you have a higher average score and the better you get, the worse your score gets, until you get really good where you're only ever hitting the 20 and the triple 20.

Speaker 1:

Doesn't your variance increase? Isn't that what predicts Like you have a stable, good score and then your variance goes up? But I guess your expected score would go down?

Speaker 2:

Well, your variance technically is getting tighter, so your variance is going down. The distribution is getting tighter for your throws, so your variance is going down. The distribution is getting tighter for your throws, but the expected score is decreasing because you're only hitting. It's an all or nothing once the closer you get, so you're either getting one or 20. So your skill is improving.

Speaker 1:

Your skill is improving, your score is dropping, increase in skill level, like when I get to that inner circle where there's the alternating point values. That feels like it's such a big jump between whether or not I get it in this margin or that margin that when we expect greater variance in my average score.

Speaker 2:

I see what you're saying. Okay, a variation in variance and score increases. Yeah that's true I'm talking about. If you think of precision, like a classic, normal distribution, like precision, as being the variance of the distribution, the skill distribution. As that variance becomes smaller, your average score drops but, like you said also, the variance of your score also increased.

Speaker 1:

So you accept balances because the all or nothing mechanic for people who are in that inner circle of skill level can make it so that you're worse off than the beginners at the end, despite the fact you're on mastery curve. That is really fucking cool.

Speaker 2:

It's super interesting. Like I couldn't believe it, because I was looking at all the scores and I was like they weren't distributed the way that I would think they'd be distributed, just like evenly over the entire thing. So it was really interesting. I'm enjoying it a lot. I got a free. Where the hot spot? The hot spot.

Speaker 2:

If you were to do a heat map, yes, oh, we'll definitely link the article that I was reading, the physics article, in the show notes. I'll post it in our show notes. But so if you're super good, super low variance, you should aim for the triple 20, because if you miss, you're probably going to hit a 20, which is still pretty good. If you're not super, super good, you should aim just left of center, because you're going to randomly hit a bullseye every once in a while. There's a lot of high scoring tiles on the left side versus the right side, so you're best better off aiming just left of center.

Speaker 2:

If you're really bad, or if you're the worst you get from not being super good, the worse you get, the lower left you should aim. So you should. Basically, if you're me, you should be aiming for the lower left of the board, and I found out this is called like the fisherman's square or something like the fisherman's angle and it's like for casual guys at the pub in england they would aim at that score at that part of the board. Really fun, I've been enjoying it like hell. I got a free dart board and so far I've spent about 200 on dart supplies, doing really well with my new hobbies it's an art supplies even no.

Speaker 1:

No, I was trying to give you an out, oh no, oh man there you go yeah, so this is this guy's going for the all or nothing.

Speaker 2:

This guy's that's that's what like a sweaty player would aim. Oh, you know what? Can I share my screen?

Speaker 1:

please, please. This is not going well for me, as you could see the outer rim triple or is the inner ring triple? Dude once you get into this visual medium.

Speaker 2:

You can't go back inner ring is triple with visuals, the inner ring is double, the outer ring is double, and then anything in the middle is going to be single. So it's funny because I think a lot of people think of the double bullseye in the middle as being the best spot on the board, but that's actually only worth 50 points, and the triple 20 is actually the highest scoring dart space.

Speaker 1:

And it's interesting because they have the points on the outer rim of the dartboard dart space. And it's interesting because they have the points on the outer rim of the dartboard. You could rebalance this fairly easily.

Speaker 2:

Yeah, so you could change the orientation of the dartboard to completely change the game.

Speaker 1:

Just replace the numbers. You could make the numbers. You could make it UGC.

Speaker 2:

Alright, how's this?

Speaker 3:

Yeah, it's a legacy game, though those boomer fishermen don't want to change it.

Speaker 2:

I'm sure they did that right, you just need to post it now. So these rings are like the variance. So he's got this this normally distributed model.

Speaker 3:

He uses a Python package that I've never used before called. Is that the optimal? That diagram right there, Is that the optimal place to aim? That's exactly right.

Speaker 2:

So you can see, the lower your variance or the higher your variances, the more left of center you should aim, the better your variance you should aim at. I believe this is the 19. And then, once you get low enough, then you should start actually aiming at the 20, because you're not going to hit the one in the five. Wow.

Speaker 1:

That's a crazy progression curve that just comes out of this.

Speaker 2:

It's pretty sick and then they do some skill assessments here. But yeah, he says I really should just aim left of center. Congratulations to Alemi for this post. This was a lot of fun.

Speaker 1:

First of all, you made darts exciting. Thank you for that. Darts are cool. The game I've been playing is called Jackpot Parties.

Speaker 1:

I've been getting back into slots ever since we talked to Matt a while back about slots. I wanted to dive back in and I think the success of Monopoly Go also made me just pay attention that slots operates in this very distinct and unique way and normally we think about the fact that, hey, it's just an EV calculation and really that's the only lever that you can pull. Pun intended for slots is that you just basically change EV and that like changes. That's basically how you run all these models and I would say that's not true. There's actually a lot more optimization you can do. There's a lot of different ways that you can play with the metagames, as well as the slot payouts the rate at which they pay out, and as well as the slot payouts, the rate at which they pay out.

Speaker 1:

And slots has just very unique set of spending patterns. We keep talking about like this turn style monetization, or I guess that's not the right way to describe it, but spend velocity is really high because you can sit down and you can say, okay, I'm going to have 100x multiplier and the cost per spend can increase significantly. So you can really change how much money you spend in a given unit of time to extremely high degree. And what these games do over time is they tend to inflate, and what I mean by that is they end up giving out more and more units of currency per dollar spent. But the thing that they do to countervail that is that they'll also increase the minimum bit size over time, and so you have this small kind of like 2% 10% inflation per year, and so when you look at old IAP skews and look at recent IAP skews, you'll see two changes in the amount of coins at a given dollar, and so that, to me, is actually the effective way to inflate an economy. And, eric, you've actually talked about this before and about like effective inflation, because what it does is incentivizes you to always have these zero account balances, get down to zero, spend everything you have and then re-up, and so one of the mechanics that they developed to take advantage of this is this thing called the safety net, which I thought was a really interesting and unique effect, and they sell it for $5.

Speaker 1:

I ended up converting on this. So what happens is when you reach the zero balance. You've done all the spins you can do with the money you have. They give you 10 million coins and then you have a 30 minute cooldown timer, so you get this bounce effect. When you spend down to zero, you get almost like this little rebound up to 10 million and you can do that every 30 minutes. So basically, you want to go in and you want to clear your inventory every 30 minutes.

Speaker 1:

I think it's just like this, really brilliant design. I just wish they had higher priced options, because I want an even bigger rebound. Like this gives me a 10 million rebound. I wish I had a hundred million rebound and you pay for that. So many mechanics have come. Yes, like this is. This is actually a stream of benefits. So I pay one fixed fee and I get the rebounds up until level 100, in which they'll stop. So you get a little bit of a boost as well, but it delivers an XP boost in a really interesting way and gets you to spend down to zero. That's a very particular set of incentives.

Speaker 2:

Are there any traditional free toplay games that use this mechanic?

Speaker 1:

I have not seen it, but I think it's totally viable. I think about squad busters, which I know eric and I have been playing. Why can't I have a little bit of a? We can have a mechanic like this where I get a chest ticket when I go down to zero, or I get a chest doubler and then there's a cool down and then I get another one. I think there's a lot of really fun ways you can layer this on to monetization the.

Speaker 2:

when an account, let's say somebody's got 10 million coins stuck in their account and they leave, does that not apply. What would that be? Deflationary pressure onto the? Well, I guess inflationary pressure, the token kind of like price. There's no trade here, though.

Speaker 1:

There's no trade here. I think that's important to remember. It's all individual economies. Zynga Poker has that problem because there's trade, because players are bidding against one another. So you're taking other players chips, which is how they get all the play to earn economics and zinga poker or any poker game.

Speaker 2:

But slots doesn't have this problem because you're always playing with the machine huge slew of skews, uh to you know, huge for all the different types of players there's one other mechanic that I want to talk about, which I also thought was compelling and different.

Speaker 1:

It's a mechanic called star power, and what star power does is it gives you an extreme number of benefits on a per spin basis, and the benefit scale with the level your star power is at. The way you increase the level of your star power is by increasing your average bet size. So you actually have to grow. You have to grow your average spend over time to be able to maximize the benefit of this reward. Now it resets so you can, like, basically go up on a ramp every day and it basically incentivizes bidding ramps. So I want to bid here. I want to bid here. I want to bid here because I want to keep increasing my average every single bid, so you get a natural scale effect which is really freaking cool.

Speaker 2:

Do you think that there's like these slots games, these casino games, the more like the games that are obviously in the gambling realm? Do you think they get off the hook when it comes to something that a traditional video game player would call manipulative or I forget the favorite words like predatory monetization practices?

Speaker 1:

They love that. Do they get off the hook?

Speaker 2:

because they're in that industry and if I'm going to a casino, I know, that I'm going to get screwed.

Speaker 1:

No, they get off the hook because the people who complain about them are different than the Reddit audience. Like, the Reddit audience is not the gambling audience. This is an older audience. You'd be surprised how old this audience is. It's not the, I would argue, probably the 20, 30-year-old male that reads, watches YouTube influencers all day and is joining the latest subreddit, witch hunt. I think that's why they get a pass. It's just based on audience. Same thing with Roblox, right?

Speaker 2:

Roblox has so many pay-to-win mechanics and loot boxes in a game that's for young children and has received very little criticism or attention because reddit does not play robo. I don't think it's fair to say that nobody wants x like, nobody wants play to play to earn, nobody wants web 3, nobody wants pay to win. Gamers want that. What they mean is this cohort doesn't want that. What we know is that, like free to play games, people do want that and there's a huge number of players who want that and there's a lot of money that's spent for that stuff.

Speaker 3:

So I think this just reinforces that point you might say nobody wants vegemite, but clearly you know what vegemite is yeah I know vegemite.

Speaker 2:

Disgusting substance is it. Is it a?

Speaker 3:

vegetable for this art it's like a brown vegetable based paste that people in australia put on top of stuff.

Speaker 2:

It's like kind of bitter, but oh it almost.

Speaker 3:

I thought it was nutella yeah, it looks like it tastes much worse. Anyway, topics I've been working on marvel snap. Marvel snap, like a lot of other games these days, sticks bots in the pvp matchmaking. It's set up where you click, play. You think you're playing against a human, but it's really a bot. The bot loses to you and you feel good about yourself. This is a trend that seems much more common lately, especially in mobile games, where it's harder to tell whether someone's a human or not, as well as in battle royales like Fortnite, because they need to fill out their lobbies and add players.

Speaker 3:

Populate 100 is a lot of people and yeah, they've been pretty popular. The big difference from historically is that these bots are ambiguously labeled to look like humans, right, whereas in most games traditionally the bots were explicitly labeled as bots, and I think it gives people a bit of a rush where they feel like oh, a lot of people have that experience their first game of PUBG, mobile or whatever. They're like, oh, I won. Oh, my god, I got number one out of 100, and then they're friends like, oh no, those were all bots, but the emotional high is still there. High is still there, right, and what we've seen is that letting people get early wins increases engagement, increases retention, etc. And yeah, it seems to be a much more common practice these days. Of course, too easy is also bad, like anecdotally. The optimal, like easy win rate is maybe around 75. That varies by game and player preferences and whatnot. Bots and pvp bots masquerading as humans and pvp is like a pretty common thing these days. Phil has strong opinions about it.

Speaker 1:

I wrote a piece. Is it okay? If should I jump in here? Yeah, go for it. I actually do have an ethical framework which may surprise listeners. It's not just about dollars and cents, I think if you take a Kantian approach to this, so basically to I would argue that the economist Kant approach is basically to say look, if you universalize one of these precepts, does industry GDP increase or decline? And so one of the things I would argue is that when you have bots that are unmarked in PvP games, it can create the impression that there's a non-zero chance that the opponent I'm playing is human, even in games that don't have this. And so if you universalize this across all the gaming, the entire gaming industry, I think it would make gaming a lot worse off. I think it would make everyone worse off.

Speaker 1:

It's a form of fraud, in my opinion, because it cheapens the experience for everyone. It's almost counterfeiting. It reduces the value of Louis Vuitton bags just because they're counterfeits, because there now is a non-zero probability that the bag you're buying could be a counterfeit. It's a huge problem and, of course, if you eliminate more of the counterfeits, what happens is that the value of any individual counterfeit actually goes up. So you have this weird circular drug enforcement elasticity problem.

Speaker 1:

But putting that stuff aside, I think it's wrong not to mark bots for players. They should not Every bot. I've seen in a lot of the recent PvP mobile games. By all intentions they are impersonating people like they're given names, sometimes they're given emojis. So there's clear intent to act here and that, to me, is not a good outcome for the gaming industry, and I'm like pro gaming. And this is why I think fake ads are also wrong, because if we universalize fake ads, we say that's okay, that's acceptable, which we've gone on for five years. The market has not taken care of this, platforms have not stepped up and enforced this. Then when a user sees an ad for a game, they don't actually know if that's the game that they're going to get when they download it and ultimately, cpis go up for everyone because they don't download games, because they don't know who's the truth seeker and who isn't. So that would be my rant, eric. I don't agree with this. They have to mark them, which I think you were surprisingly open to as well yeah.

Speaker 3:

So I have two counterpoints. One is, if you want to be the game that is explicit, you can just label your bots. You talk about this negative externality, about people not trusting things if there's more counterfeits in circulation, but games are game-specific and if your game labels your bots, people know or at least they should know that your humans aren't bots, so you can get around this counterfeiting problem. The other point I would make is that I think it's probably good for the engagement of the game when I've been teaching my five-year-old daughter how to play a lot of different games and very often I will intentionally lose or sandbag on the first couple games to give her a fair experience, make her feel good about winning. And so, as a question I would say imagine, instead of bots, lost bots. The game paid a bunch of people in Indonesia to purposefully lose when they played against new players and they were labeled as human. Is this more ethical? Is this an acceptable thing for them to do, Ignoring the cost side?

Speaker 1:

So I think it still validates the Kantian precept that I suggested, which is, if this is universalized, it's still faking it. I don't think it's the idea of it's. The idea of this isn't a truthful experience, and that cheapens it for people.

Speaker 3:

So is it unethical for me to intentionally lose to my daughter in tic-tac-toe because I'm like cheapening the experience, I'm making her believe I'm trying when I'm really not?

Speaker 1:

I think there's a different expectation between your daughter and someone who wants to be like the best, so like someone who wants to climb up a leaderboard. The fact that they're like climbing over bots, I think, creates an artificial sense of accomplishment when, in case that wasn't real, like we could put it other way it's almost like lying on like a calorie tracker. You're a truthful seeker and you're not providing me accurate information that I can assess. I don't think your daughter's after accurate calorie seeking. I think she just wants to have fun and play chess. I think you're allowed to lose to her, but I think for PvP players there's a clear intent.

Speaker 3:

So it's the combination of this with some kind of ranking metagame.

Speaker 1:

Yeah, I think there's an expectation that when a player enters a PvP game, if we were to survey these players, I think a lot of them would be under the impression that these are. I think, over time they have a natural expectation that there are bots, and that's exactly the problem. But I guess what I would argue is that if you surveyed people in games that didn't have bots, over time you would see an increase in the amount of participants that are saying that they think the game has bots, even though it doesn't, and that ultimately reduces the experience of PvP in those games. And that's my problem with this, because that's not aligned with their expectation and that there's a retentive effect there. Do you think that's a reasonable? There's these sudden, there are these almost shadow taxes. Everyone else pays because games like marvel snap won't label their bots, because squadbusters is doing this too, by the way, and call of duty mobile, like a lot of eastern games, are doing this there's a deadweight loss and it's a function of the number of what I call hardcore players.

Speaker 2:

So I don't know if you guys want to hop into the. I think this is exactly like yeah, go for it.

Speaker 3:

Yeah, Chris, the economist thing where he was like there's this controversy.

Speaker 2:

I'm just going to make a model and parameterize it and the model basically says it agrees with everybody. And I think, to preface the finding, in my model it's like it's probably better if everybody's just transparent about whether you are or are not a bot game, because then you can avoid a coordination problem, because there's a coordination problem between what the firm should do and what the consumer should do, and you don't want them to do different things. So the whole model basically looks at the profit from the firm and optimizes profit. And then it looks at the consumer welfare and tries to optimize consumer welfare. And the important thing is that there are two different players. There's two different types of players. There's hardcore players who don't like bots. So basically the way it is, if there are no bots, then both the hardcore player and the casual player are both better or both get one utility. They get the same utility. They're perfectly fine. However, if there are bots, then the hardcore player gets negative one utility and the casual player gets two utility. So basically what this is that hardcore player that Phil was just alluding to is super angry, hates it, gets pissed off, doesn't want to play the game. So if the hardcore player sees that there's a bot in the game or thinks there's a bot in the game they're not going to play Versus the casual player who's oh fuck, yeah, I'm winning more games, I'm going to play this, this is awesome. So that's Eric's daughter. So you've got the hardcore player and you've got the casual player.

Speaker 2:

So the model basically just says look at the share of people that are going to play the game. If there's a bot, it's just the share of casual players in the game, and if there's no bot, then it's going to be one. So we use this relationship. What in the model is 1 minus H, which is the share of hardcore players, or the share of casual players is equal to 1 minus the share of hardcore players. We use this to create a condition.

Speaker 2:

This condition says that bots are welfare-enhancing only if the share of hardcore players is sufficiently small. So the majority of people are love bots or sorry, don't care about bots, or get more utility, and it's actually better for overall welfare for there to be bots, even though there's a share of hardcore players who are harmed, so to speak. So this is welfare enhancing to have bots even though there are people that hurt. So this is the whole entire social welfare kind of framework.

Speaker 2:

So we're taking a Pareto approach and we're saying, if we're here, if we're at point E and we apply and we try to get to point E star, are we better off at E star? And we try to find the point where we're best off for everyone. So we sum up the overall utility and we say, okay, whichever equilibrium, whichever strategy is, the one that optimizes the sum of all utility is the best one. And now those utility values I gave earlier, minus one, one and two, with those utility values assumed without loss of generality, you end up with this condition that says if the share of hardcore players is less than one third of the population, bots are good. So that's condition one that's a really important condition. It says that if the share of hardcore players is small enough, bots are good.

Speaker 3:

If it's big enough, bots are bad, so one way to interpret this might be that. So for COD mobile, if COD mobile has more casual players and COD PC has more hardcore players, then it makes sense that COD mobile has more bots.

Speaker 2:

That's exactly right. So you would employ so and we're about to get to the firm's problem but you'd be better off employing bots in a more casual atmosphere than you would be in a more hardcore atmosphere. So bots are good in one context for consumers and bad in another context for consumers.

Speaker 1:

My issue with the model from the way you described it, is that I agree on the benefits of bots, but what I don't understand isn't lying to a hardcore player about a win super valuable. The person who is hardcore values win rate quite a bit. They love to win, they have a thirst to win, and so wouldn't me being able to supply more of those, even if I'm lying increase net utility.

Speaker 2:

So you're saying that the hardcore player is even happier.

Speaker 1:

So I think so. To me, the issue that we're trying to get our way out of the whole reason that bots have been wildly successful, and I can say that with extreme confidence that introducing bots into a game does wonders for KPIs. There's no doubt about it that Marvel, Snap and Call of Duty Mobile experience huge gains in retention because kill death ratio goes up when you have bots Because, remember, they can suck up a lot of kills. There's not an equal distribution and so that means everyone's KD goes up. Everyone feels a lot better. I have zero doubt about that effect. I only care about whether or not lying is useful, whether or not concealing this information, and I think if you lie to high value players, that they value it the most.

Speaker 3:

Phil, I think one way to interpret it is so. First off, I think casual players care a lot about winning. My daughter definitely cares a lot. If she wins, she'll cry if she loses.

Speaker 1:

I think it's more that but she's competitive, though I guess maybe that's the right persona. She's a competitive player.

Speaker 3:

Yeah, I think one way to think about Chris's model is that the hardcore players can figure out that they're bots and the casual players don't realize.

Speaker 1:

Yes, okay, okay, so it's. You can't keep the lie. Exactly so it's like an open information environment.

Speaker 2:

So there's no, we're not using Bay's rule. Yet I have another one, another article that I'm writing that does use Bay's rule.

Speaker 1:

Okay, this makes a ton of sense. So, this is this is.

Speaker 2:

You discover the this condition that says if the share of hardcore players is high, bot's bad. If it's low, bot's good, that's for the welfare. So then we're going to get what is the condition? When are bots good for the firm and when are bots bad for the firm? So we set up the firm's profit equation, not going to talk about that, just going to say for the firm what's better, a bot or not a bot? And we basically say, okay, if there's a bot, this is how much the firm makes. If there's not a bot, this is how much the firm makes. And that gives us condition number two. And economists love coming up with conditions because they allow us to say things about equilibrium. Condition number two. So the condition number one for the consumer how do we maximize consumer surplus or how do we maximize consumer welfare? And condition number two is how do we maximize profit? So condition two once we optimize that profit function, we come to the conclusion that the share of casual players must be greater than bot elasticity of demand. So I made up this term, bot elasticity of demand, which is basically what's the gain in profit from introducing bots into your game, which is basically, in my simple model, the ratio of the, the bot, the no bots price over the bots price. So this is basically if the increase to profit is 200% when you introduce bots, that is that bot elasticity of demand, so the amount of profit you get when you introduce bots. And it's assumed in this model and Phil you, just you already spoke to this the assumption is that introducing bots has a higher marginal product or, sorry, a marginal price than no bots. So you're going to get more money per user in the bot model than you are in the non-bot model. So you end up with this condition that basically says as long as the casual share of players is high enough relative to the punishment or the penalty that you would get from not using bots, then you're going to use bots as a firm. So if casuals are too high or are high enough, you will have bots. If the casuals are too low, you will not have bots. Likewise you could also say if the body elasticity, if the amount of money you're going to get from introducing bots, is great enough, then you'll introduce bots. Relatively straightforward condition.

Speaker 2:

And then we try to solve this in a general equilibrium cases like way. So we end up with a bunch of different cases. So you have cases where the bots are good for the players and the bots are good for the firm, and then you have a situation where the bots are bad for the players and bots are bad for the firm. And all of this is defined by the share of hardcore players Because, remember, both conditions have eight in them. They both have the share of hardcore players in them.

Speaker 2:

So when you look at those two conditions, you actually end up being able to plot a number line and on that number line is just from zero to one share of hardcore players and, depending on the bot elasticity of demand, there's a segment on that number line where, if there's a low number of hardcore players, you end up in what I call the good equilibrium, where both people are, or sorry, the true equilibrium, where both the firm and the consumer wants bots, and the false equilibrium, which is where both want nothing. So a bad equilibrium is one where we want opposite things and a bad equilibrium is one where we want opposite things and a good equilibrium is one where we want the same thing. So I basically plot out, based on the body elasticity of demand, where on this number line is a bad equilibrium located, and I'll just share this image because I think it'll make a little bit more sense.

Speaker 1:

We got visuals now.

Speaker 2:

So this is. I spent a lot. I'm very proud of this visual because I was able to. Basically, I say, okay, we know when the equilibrium is good, it's when both guys are doing the same thing, the firm and the consumer both want the same thing. So we don't care about these, we care about bad equilibrium. When are bad things going to happen? And it happens when either one wants bots and the other doesn't want bots, or vice versa One doesn't want bots and the other does want bots.

Speaker 2:

And you can see, there's this little segment here between this threshold one third we talked about earlier and one minus the body elasticity.

Speaker 2:

If you're in this range, if the share of hardcore players is in this range between these two numbers, you're going to end up in a bad equilibrium. And if you're outside that range, you're in a good equilibrium. Everybody's happy, everyone's better off with bots or in the case of or without bots. The important thing is, the closer these two numbers are together, the fewer cases you're going to have a bad equilibrium. And the greater is this number, in other words, the greater is the incentive for the firm to use bots, the more likely you end up in a bad equilibrium. You can almost imagine randomly drawing a number from a number line and your likelihood of ending up in this bad region increases as the profit that is to be gained from the bots increases as this point here, this one minus the bot equilibrium, extends all the way to one. So that's the model. It basically says there are situations where net, so overall welfare. You take this Kantian approach, for example overall everyone's worse off because you introduced bots or because you didn't introduce bots, and so that's the model.

Speaker 3:

If I'm interpreting this number line correctly, if the share of hardcore players is very low, then you're going to get players want bots, firm makes bots. If it's very high, it's going to be. Players don't want bots for exactly bots but there's this sweet spot in the middle where your player base is divided. Right, you've got a big faction of both hardcore and casuals and the firm the game might choose the quote-unquote wrong or bad equilibrium that's super well said sounds like squad busters, to be honest with you really, where do you thinkbusters lies in this?

Speaker 1:

I think it has a bad equilibria, where it has a mix of hardcore and casual players and they use bots.

Speaker 3:

And are you the hardcore player who's upset that there's bots in there?

Speaker 1:

The thing that's different is that the bots are tuned down, though, so that they increase your power, like it feels good to defeat them, but I'd rather defeat a real player, because actually it frustrates me. Actually, it does frustrate me, because it basically means that if you spawn close to a bot, you get a free kill. That's really frustrating to me, because it's zero sum.

Speaker 1:

It's zero sum because they basically die out so quickly because it's Battle Royale style. They don't survive, so they're basically food for whoever spawns next to him. It's basically like spawning next to a free piece of Tomahawk steak.

Speaker 3:

So I know some hardcore Fortnite players and they say the same thing, where there's bots in the games and they're like free kills to get loot. But for them they think of the bots as just another loot source and the skill is in determining whether a player is a bot or a human. And if they're a bot you go kill them really quick, get all their gems. So they're just yeah, for them it's a skill test and they don't really think of them as competitors.

Speaker 1:

It has cheapened my entire experience, though, to know that I could be beating bots.

Speaker 3:

That's true, because sometimes you're like, oh, those weren't human.

Speaker 1:

Or I'm not in the real competition yet.

Speaker 2:

Now going back to if the firm is transparent about whether it is using bots or not. Even my model suggests that there could be a situation where you're in a bad equilibrium. Even with perfect information, one could argue that bots are bad.

Speaker 3:

Phil, this discussion so far has all been about lost bots, like bots that are intended to offer easy win. Does any of this change? If it's, let's imagine it's a bot that's a perfect simulation of a human. It passes the Turing test and it has 50-50 win rate, or whatever, does it?

Speaker 1:

even matter. You would introduce this because it could reduce CCU. You could basically design a difficulty curve, you could feed it to the player and there's still an expectation that this could be human or not human. I would say at the end of the day oh man, I feel like it's very hard for that information. I guess that information could leak. That would cheapen it. I feel like just that certain fact would cheapen it. It's like people do lip syncing If you're a musician and you go out and you don't actually use your voice, but people can't tell the difference. But if a tree falls and no one's around to hear it, did it really make a sound? It's the question that we're asking and I would say this information can leak.

Speaker 3:

So there's a probability of this destroying everything. But yeah, like on a spreadsheet, yeah, this could. This, of course, makes sense. It's all so many problems. Does the knowledge that some musicians lip sync devalue the?

Speaker 1:

live performances to you because you know there's a chance. Yes, we just had controversy where the lead singer of the foo fighters threw shade at taylor swift for implying that she like doesn't actually sing during our concerts and like that now devalues the experience. Just that accusation for all musicians, because we know this is sometimes a thing that people do. I think that makes everyone more soft. So it's kind of Louis Vuitton bags. We want to talk about internal competition.

Speaker 3:

Yeah, sure.

Speaker 1:

It's up, eric. Internal competition gone wrong. What went wrong? What internal competition?

Speaker 3:

This is an observation about game companies and people. But so games are little virtual monopolies, right, like you're the only one who can sell content in your game and there's zero marginal costs right. But because you have the monopoly, other people can't compete with you and you can take all the profit you want there. And that's where a lot of the power of game monetization comes from is that you are this little virtual monopoly. However, there's a pattern I've seen at larger companies especially they're more I would call decentralized, where there's more internal competition, where you've got all these different product managers who own a specific feature. Let's say, one person owns the shop, one person owns the rotating store, one person owns the battle pass, etc. And each of these product managers is being evaluated based on their product's performance. Or there's some pressure to boost some short-term sales, and the easiest way to do that often is with price discounting, and so the product managers will start discounting the product, be like, oh look, I cut the battle price in half and the sales went up and my chart looks great. Give me a promotion, thanks. But then there's ecosystem-wide effects where, when they reduce their price, it's competing against all the other products that the game is selling and the net. The net revenue is actually net and functionally what's happening is these internal products.

Speaker 3:

In theory, the game should be a monopoly, but they're functioning like a competitive market right. They're competing against each other on price and once they're competing on price equals marginal cost instead of price equals price equals marginal cost instead of setting a marginal revenue mixed up. But you're basically forfeiting market power once you have these internal products cost instead of setting a marginal revenue. You're basically forfeiting market power Once you have these internal products competing against each other, trying to undercut each other, trying to fight for their slice of your game's market share, and you're throwing away your monopoly advantage and your monetization potential there. So, yeah, anyway, this is something I've seen a couple times. I've heard stories that Star Wars Galaxies, for example, did this, where they had a product lead who was extremely aggressive with this stuff, got a ton of revenue, a bunch of promotions, left for a higher role and the person who got left holding the bag was the next guy. Oh, they torpedoed the game economy.

Speaker 1:

Brilliant, brilliant political play.

Speaker 3:

But this seems to be a common thing and normally I think internal competition is good for fostering innovation and productivity and all that. But this is an instance where that internal competition, like actively, is harming the product as a whole and which begs the question how do you deal with this, like how do you prevent this from happening? I think, on one hand, the big culprit here is price competition, that these products, rather than trying to say how do I make battle pass a better experience, what feature dev can I do here? They think how do I reduce the price to get more sales, because that's always easier and cheaper to do in the short term. But another thing is I think it helps to have some kind of central authority who's seeing over the whole thing right, looking at what is the rate of gold inflation in the economy and how much of this rewards budget can I allocate between the different groups, some kind of like a Federal Reserve or that.

Speaker 1:

Eric that's it. You just I think you nailed it. This is exactly what we see in the federal government, right? We have peer-reviewed research that the level of independence the Federal Reserve has is correlated to relatively stable inflation. So the more independent a central bank is, the more they seem to have inflation under control because they are not responsive to the whims of politicians. Why wouldn't you just have a Federal Reserve inside of a company that doesn't report into the PM org, One that's actually dis-insidious, Like? Why not create an internal Federal Reserve to avoid this problem?

Speaker 3:

I think you could. I think most games just don't do it because it's extra org bloat and you gotta hire someone and that person is mean to other people.

Speaker 1:

For instance, the difference in most organizations is that analytics does not report into the PMs. Now I've seen cases in which everyone reports into the PMs, which I think is dangerous for exactly the reasons Eric pointed out. But if analysts reported into central tech or into a different reporting line, which I've also found to be the case, they have the ability to call out PMs, which has happened to me frequently. It's like when I've not reported into orgs, I think the PMs also will sandbag numbers. I think there's a counter effect here. What's the target this quarter? Let's set it really low so we can overcome it.

Speaker 1:

That's another perverse effect that can happen, and sometimes analysts form the watchdog group. They're like, hey, these numbers are not good or you just ran a sale. I think that's. One idea is independence. That's a principal agent problem. But what about just having a holdout group every quarter? I've always dreamt of this. But why not just hold out 15% of your player base every quarter from all the updates that happen in the quarter and just evaluate product managers every quarter based on the difference between the control group and the variant which has all of the things the PM has done?

Speaker 3:

First, it's not always possible with things like matchmaking or clans.

Speaker 1:

Okay, okay, okay, okay. I'll give that to you.

Speaker 3:

The bigger issue is the long-term effects. There might be long-term effects that you're not seeing within your holdout window, and what do you hold them out forever?

Speaker 1:

Do you think that's better than what we have now, though?

Speaker 3:

I don't know. I'd have to make a model and parameterize it, and what parameter values lead to a bad equilibrium?

Speaker 2:

so for with respect to like evaluation. Like a popular topic in economics is personal in personnel economics is how do you evaluate the value of a specific employee? It's much easier to evaluate the, the marginal return for an employee who directly impacts, like kpis. But what if you're a pm who works on something that doesn't necessarily impact revenues quite as obviously, how do you evaluate the impact that person has? And is it fair to compare them? And do those people make less money than the? What exactly is your value?

Speaker 1:

proposition, I think is actually a pretty good question to ask. Now, to be fair, I do think there are legitimate answers to that question, but I think most of the time that question would do more to cut down the field of PMs than increase them. You should have a direct connection to the bottom line. But yeah, I'd probably concede on a lot of those points. I guess it's all about better or worse, right, I think it's another tool in the information bag for, ultimately, the principle to evaluate the agent and how you choose to weight that information, I think is up to each principle. But I do think that is a valuable piece of information that I'm confused on why more people don't try to seek it out. Great, just don't do the monetization features for the monetization PMs. That could be another solution here.

Speaker 2:

Eric, your thesis is that competition within firms is actually bad for the firm's bottom line or their product.

Speaker 3:

Yeah, in this specific domain, where the firm has a monopoly and the internal competition is on price rather than quality.

Speaker 2:

So this doesn't apply to a law firm where all the lawyers are competing to try and become partner, because the law firm doesn't necessarily have a monopoly over their product.

Speaker 3:

Yeah, I think so. The marginal effects of one group on another is not huge. There's a huge demand for lawyers.

Speaker 1:

I think a different analogy is like imagine if all the lawyers were selling peanut butter and all the peanut butter were in the office and like you could grab as many as you want and you can price it yourself, and you're just trying to sell as much peanut butter as you possibly can. So I'm going to choose probably it might be a lower price than I might otherwise have Because I want to front load all the peanut butter sales, because if I don't, other people are going to sell the peanut butter.

Speaker 2:

Plus, they don't really have the ability. Lawyers are not PMs, but bad example Peanut butter, peanut butter. I did have some peanut butter today. I don't know why I did that, either my wife started baking sourdough bread recently because my friend gave us sourdough starter the whole new game.

Speaker 1:

Messy. It's a game, it's exhausting.

Speaker 3:

There's so much crap you gotta do for that.

Speaker 1:

To make sourdough. Is there really sourdough?

Speaker 2:

We're like actually making sourdough.

Speaker 1:

Okay, I don't know man, you gotta clarify this shit. Okay, like I don't know man, you gotta clarify this shit. We're in the age of goat simulator and like power wash simulator and train shit.

Speaker 2:

I'm sure there's a sourdough game. I'm sure I don't have to Google. I'm sure there is. I'm gonna Google it.

Speaker 3:

Sourdough bread making game. Have you guys played?

Speaker 2:

it. What's the kitchen game where you're running around in the kitchen trying to mix Overcooked?

Speaker 1:

Have you played that that's insane.

Speaker 2:

I can't play it anymore. It's too stressful.

Speaker 1:

Not healthy man, Not healthy at all.

Speaker 3:

I can't find a sourdough bread making game.

Speaker 1:

I think this might be a niche Hypercasual, hypercasual.

Speaker 2:

One of the hypercasual shit together. Any of you listeners that are CEOs or CCOs, creative officers, go create a sourdough game.

Speaker 1:

That's all we got Episode 28?, 29?, 29, 29 in the case wouldn't it be crazy when we hit 50, 50 episodes? That's gonna be wild we should teach this to our children.

Speaker 2:

Economics is major everyone has to major in economics. Number one for personal survival, economics is major.

People on this episode