Game Economist Cast

E21: Subs, Doms, Surprise, and Suspense

Phillip Black

Send us a text

Eric starts 2024 with a new paper on League of Legends published by leading researchers. Is there an optimal amount of excitement in matches, and if so, where is it? Chris returns from Italy, and holy cannoli, did he have a "time" while  Phil laments subscription-based pricing as the antithesis of game monetization.  The crew agrees to spend money on Magic The Gathering this year.

Speaker 1:

It's pretty. I don't know how to describe it. It's hard to describe because I have to watch it, but yeah, I guess.

Speaker 1:

So this YouTuber got super big off these memes and they just made a bunch of meme videos about this egg dog thing, and then I guess there's it's a random dude in Seattle who's a millionaire and is like I don't know what to do with this money. So they just started throwing money at snatch tournaments because there's no money in eSports right, esports is a big money loser. So this person is just like I like these players. I'm just gonna sponsor a bunch of them. So it's okay, that's cool, I'll buy it. It doesn't work.

Speaker 2:

I wanna be one of the weird internet communities you two are a part of. I feel like you guys are always in like just very strange annals of the internet.

Speaker 3:

Well, I'm all over the home improvement Reddit, but I don't know other than that.

Speaker 1:

Yeah, I got your board game kinks.

Speaker 3:

That's true, that's true.

Speaker 2:

Let's start with utility. I don't understand what it even means.

Speaker 3:

Everybody has some kind of utils in their head that they're calibrated. There's hardly anything that hasn't been used for money.

Speaker 1:

In fact, there may be a fundamental problem in modeling when we wanna model.

Speaker 2:

Game Economist cast episode 21, back in the saddle. Do we all bitch about the inefficiencies of gift giving during the holidays? Did that cross anyone's mind?

Speaker 1:

The dead weight loss. I bitch about that every year.

Speaker 2:

You bemoan that Do you give gift cards?

Speaker 1:

No, I hate gift cards. Gift cards are like just give them cash. Gift cards are so awful. Why are you just gonna have any chance that this thing doesn't get used? Or you give them a $100 Best Buy gift card, or they wanna go to Circuit City or whatever the fuck.

Speaker 3:

Dude, that's the whole business. That's the whole business.

Speaker 1:

Why buy into that? Or those Visa cash cards where you pay a 20% fee to the credit card company or something?

Speaker 3:

It's beautiful, it's beautiful. I prefer gifting experiences.

Speaker 2:

Oh hey, let's go out to a show Ultimate Dead Weight Loss or something like that.

Speaker 3:

What? No, that's the perfect thing, because you wouldn't normally go do that by yourself. I've always been of the belief that the best gifts are things that you do want but you wouldn't buy for yourself. Like I can't. I just can't justify buying Scotch, like because it's too much. It's 80 bucks for a bottle of Scotch. I could go buy bourbon for $30 in the US, but when somebody gifts me an $80 bottle of Scotch, I'm super pumped. That, to me, is where the that's where the gift giving is not all dead weight loss Trying to save you from some sort of guilt. Yeah, probably, it probably bypasses, like if you were to take a behavioral approach, there's probably like a social cost of buying the Scotch where you're like oh God, I could probably put this in my like 401k or my Roth IRA or something.

Speaker 1:

Wouldn't you rather them just put the $80 Scotch in your Roth IRA?

Speaker 3:

Like the Scotch as an investment, or the 80 dollar.

Speaker 1:

Instead of giving it the 80 dollar Scotch, just put that $80 in your Roth.

Speaker 3:

IRA. No see, I want the Scotch. I just can't justify it. Who am I gonna? I'm like, oh yeah, honey, I just like nickel and dime to the HVAC people and got pissed off that they were charging me this for a part. But I bought an $80 bottle of Scotch, Like how could I look myself in the mirror, man, I can almost hear Richard Thaler, just like running from the hills, trying to justify Chris's position. I'm serious For me, I think like the person.

Speaker 1:

If the person doing the buying is very understand Scotch and they know what a good Scotch is and a bad Scotch is and what a good price is, right, they're better at shopping for Scotch for Chris than Chris would for himself, right? So in that situation I see gift giving as positive. That's fair.

Speaker 3:

The problem is I'm a nerd and I, like, know a decent amount about Scotch. I watch Scotch YouTube videos.

Speaker 1:

There must be some product category that you don't know much about, but you still need to, so it's almost like you want to be a gift similier.

Speaker 2:

That's a good point. You're trying to solve like an asymmetric information problem and you need to understand. You could understand someone's utility function better than they do. That's very dangerous.

Speaker 3:

So how could you? But you would like for that to really be like net positive. You would need everybody in your gift giving group to be a specific connoisseur of each thing, and then, or you just have to know more than them, that's all. So Secret Santa shouldn't be randomized. It should be based off of each person's expertise. That's actually really great.

Speaker 1:

That's why I think White Elephant is better is because you can show which item you're more valuable.

Speaker 2:

I like where Chris was going too. Like he want to designate similiaries for each like gift category. Like I guess I would be the games guy you could have. I don't know what other gift categories are there. Home decor.

Speaker 3:

Home decor Scotch whiskey.

Speaker 2:

in general, you can tell we haven't talked to each other in a while. We have two wonderful articles to talk about today. Speaking of unexpected gifts. Eric, what are you going to be talking about?

Speaker 1:

Talking about surprise and suspense in games and, just generally, if you view a game as like a random process that results in a winner loss, what types of those random processes look good or bad? As a part of a research paper, we took a ton of League of Legends games and tried to figure out what types of games people liked or disliked in terms of like swinginess. Are you winning early and then you got shut out late? Do you want a close, long game or a fast shutout? Do people prefer comebacks over stumps, stuff like that? Looking at the way the game itself plays out, so what's the thesis here?

Speaker 3:

Like I get what makes a sports game, a political election or a mystery novel exciting to observe Is the idea like this is better for profit or something, and that's why political elections tend to be more like exciting and explosive I think, maybe not political elections, but like certain events that you watch, like when you watch a sports game.

Speaker 1:

some sports games are more exciting to watch than others, and so how do we quantify that? What attributes quantitative attributes about the sports game make them more exciting? And, as a game designer, can you intentionally recreate those types of things? For example comebacks are really exciting, and so can you make a game with more comebacks, and would that actually increase engagement with your game?

Speaker 3:

Ooh, that sounds interesting, but this is just the intro. We're not talking about it yet.

Speaker 2:

So Eric's gonna be talking about his own paper that he got published. It's out, although I remember you were telling us you had pulled this like years ago Did this long?

Speaker 1:

to get out this.

Speaker 2:

It was to be like eight years ago.

Speaker 1:

Oh my God, it's not published in a journal yet, so still working, oh my God.

Speaker 3:

Yeah.

Speaker 1:

Yikes academia.

Speaker 3:

That's how it works.

Speaker 2:

I will be talking about game subscriptions. There was a piece out of the Wall Street Journal that Netflix is reconsidering using IEP and ads in their game subscription services. This is something I've been writing out for a really long time. Why don't subscriptions work in games? There is a morgue that is brimming with game subscription bodies, but I think the answer to why game subscriptions haven't worked Both subscriptions that gate multiple pieces of content, like Xbox Game Pass I also mean individual game subscriptions like World of Warcraft. Why haven't those survived? I think, strikes at the heart about why games are games and why they're different than other forms of media, and so we'll spend a moment and talk about that. Before we do, let's talk about what we've been playing.

Speaker 1:

You've got a selection of good things on sale stranger.

Speaker 2:

Chris, what have you been playing?

Speaker 3:

I returned to a classic, the Witcher 3. I played this originally on PC. It was like I bought a graphics card and built my PC, like I think my senior year of freshman year of high school or freshman year of college, and right around then the Witcher had come out and they were like bundling the Witcher with graphics cards. It was like a GTX, like 9.9, 20 or something, so it was a pretty old graphics card and like back then it was like this amazing, beautiful game. So I remember enjoying it a lot and since then I've read the books. Obviously, the TV show came out and married a Polish woman. So I was like, oh, I'm gonna play this Polish game that has a Polish like, all audio is Polish, so I can have the text in English and the audio in Polish and it's been good so far.

Speaker 3:

I remember why I never fell in love with the game. The controls are just so clunky. I wanted to have the story of the Witcher, but the gameplay of a Souls game. I just want the combat to be more engaging. It's just always felt clunky to me. It's even more clunky on mouse and keyboard, less clunky with a controller, in my opinion, but still a little clunky. Like dodging feels weird, like just the way your character moves has always frustrated me with those games. The Witcher, it's something that I wanna like a lot more than I do, so that's what I've been up to. Good old fashioned single player game. Got all the DLC playing on Xbox, enjoying it for the most part.

Speaker 1:

Do you think it holds up?

Speaker 3:

I don't, like I said, I don't think I liked it that much to begin with. I don't know if I'd say it holds up, but the story is still pretty, like, pretty incredible. I think also like it has a at least in my opinion, it has one of those timeless graphics, like I feel GTA 5 was really timeless in the color scheme that they used for the game. So, even though graphics have gotten a lot better since GTA 5 came out, it still looks good. When you're just watching the gameplay and you get like a screenshot, you're like, oh, that looks like a good game, because they came out of this awkward time period, like in the late 2000s and 2000s, early 2010s, where games were trying to go super hyper realistic but they weren't. They weren't, they didn't have the graphical power to really pull it off and they just, in my opinion, looked ugly.

Speaker 3:

This game, I feel, was really on the like on the right, in the sweet spot, where it looks great, but it's mostly because of the colors and the polygons that they chose to shapes. So, even though we've got a lot better stuff out now, I think it still looks super good. So story's always going to hold up. I think the story is great and that's why I'm playing. It is because I've now read all the books and watched the show and I'm like, oh, let's see what this story actually is, because I don't really remember it when I played it originally.

Speaker 2:

So I think it holds up.

Speaker 3:

I like it.

Speaker 2:

Eric, what have you been up to? What have you been playing?

Speaker 1:

Went back to Magic of the Gathering.

Speaker 2:

Oh surprise.

Speaker 1:

Playing Magic Arena.

Speaker 2:

Hit the pipe again.

Speaker 1:

Yeah, one of my friends was hosting a draft in person and I was like, oh, I haven't played Magic in two years. I got to learn the draft set and stuff. So I played a bunch of Arena, dropped $100 in the game. That game is expensive ever.

Speaker 3:

Biggest understatement of this podcast and gosh. Everyone knows what.

Speaker 1:

Magic is, but just a few things that kind of came to light to me. Like one is that Arena is huge, right, and people were saying I don't know if this is true. My friends were saying like yeah, arena might be bigger than Paper Magic for standard and draft formats, like it might be way more popular just because it's way more accessible and you can just play a lot more games. And apparently Paper Magic has shifted hard into Commander format, which is like a casual four player free-for-all format, which again is the opposite. I would have expected Paper players to be the more sweaty try hard players, but I guess Paper is more casual and Digital is more competitive because you can sweat all day.

Speaker 3:

Well, a lot of people are playing Commander Paper, right, that's like the most popular way to play Paper now and that's a very casual format.

Speaker 1:

Yeah, and I guess it makes sense that if you're going to meet with people in person, your pool of competitive matchmaking is a lot smaller but your pool of the social fidelity and like the social enjoyment of the game is much higher, and so I guess it makes sense in person would shift towards that more social, casual play with your friends with a wider skill gap experience. But yeah, that surprised me. And still, no trade Powergroup has been insane, just absolutely. I look at the one drops that are common now and I'm like this would have been like a four of rare bomb in like a you know people are winning by like there's four cost creatures that just end the game.

Speaker 2:

People are winning by like turn five, turn six. I think they've it's gone down quite a bit, which is sad. I mean there just isn't a lot of strategic flexibility in five turns.

Speaker 3:

Do you think it's just like eventually they'll hit a cap, like I feel, like for the last? So I've been listening to magic podcasts for probably like 10 or 12 years now and it feels every year if they're like, oh, this set is broken, they need to ban everything, they can't keep doing power creep like this. But it's been 10 years and they just keep doing it. So I'm curious if there's just like this natural reset that happens over time with the set, the set structure, and eventually it comes back down a little bit and then it creeps up, maybe over time it's like trending up.

Speaker 1:

But bear markets, yeah, I I.

Speaker 3:

I just I wonder if there is any. It's like, what is the limit? How long can this go on, or is it as bad as we think it is?

Speaker 1:

I think the game still has plenty of strategic complexity, even when it is way more turbo. But you do lose out in all that whole back catalog of cards, which is content that if everything is power creep, that content gets removed, whereas if power levels kept the same, that's relevant for deck building and stuff.

Speaker 2:

Yeah, I don't know. I think they really fucked themselves with the mana system. I mean, we've been talking about just how bad the mana system is in general, like every CCG has moved away from a mana system to, or at least a flexible mana system where you're playing mana to mana, just you get one mana, one additional mana per turn, and that is an effective cap on the number of actions you can take per turn. But the fact that you can generate mana free flow seems to be like just this huge design problem. And they've actually made it worse, in my opinion, because they've introduced so many dual lands that they've reduced the barriers between both of the colors, and so that just gives you so many more strategic options and so that's so many more combinations you have to balance against. There isn't like this class system that you can say, okay, the blue cards are usually not going to mix with the green cards. It's oh shit, now everyone's running three colors. The level of strategic balancing has exploded. I play drafts.

Speaker 1:

So three colors is like you're committed to a slow deck.

Speaker 2:

Man, I'm a drafter too, and it's man. 17 cards in my deck are going to be mana. That's so boring to draw, and that's literally this. Almost half my draws are like boring lands.

Speaker 1:

Yeah, lands are the original sin of magic Like what Whoa?

Speaker 3:

No, it's one of my favorite parts about magic.

Speaker 1:

That's why I'm using the original sin what you like lands, are you? Yeah, that's why I never got into hearts.

Speaker 3:

I didn't like heart stones, like all this, like single one, like mana Correct. I like the colors.

Speaker 2:

No gross.

Speaker 3:

I like the colors, but I'm the kind of guy who go ahead.

Speaker 1:

A third of your draw. You draw seven hands. You got cards. You got one land. Are you mulligan? You draw a card. You got five lands. You mulligan again?

Speaker 3:

Yeah, no, I'm not saying it doesn't have its problems, but I love mana. That is what I think of when I think of magic gathering. It's iconic.

Speaker 2:

I'm just going to leave that there.

Speaker 2:

My bad take can just go away, fail or even Speaking of, I'd say, near CCGs. I played a shit ton of Clashmini over the break. It's a game I've gone back to time again. It's a game that's in soft launch. It's a supercell title. They released a big update. It looked like a huge content drop. I think the game is probably going to get killed the next six months.

Speaker 2:

This was one of their initiatives out of their Shanghai studio Studios. It was the one of the first studios inside of Helsinki. It looks like they just haven't been able to solve auto chess in general, which is really sad to me. It's the dream that never materialized. Underlords from Valve came out and died very quickly. Remember that was based on a mod from Dota 2. It looked like the general story of basically every Valve game and most Blizzard games was going to be fulfilled yet again, which is, hey, they took a game, they made a mod based on it and the mod was really successful and they spun it out into their own title that did not have one under. They attracted some DAU, but they couldn't just keep it going. Tft, at least in the West, didn't seem to work out either, although it's interesting.

Speaker 2:

The Chinese version of TFT teamfight tactics from Riot thought their auto chess version which they put together very quickly, by the way was extremely successful. It's called Fight for the Golden Spatial. It's extremely successful in China. It's always in the top 10 in terms of revenue. It is actually operated by Tencent, so Riot doesn't have their hands on it. If you remember, riot also refused to adapt League of Legends very early on for the Chinese market, which resulted in Honor of Kings, which was the Tencent version of League of Legends and that has been the most successful game in China. They snubbed their nose and thought I don't know, maybe Eric has different opinions here, but there was an arrogance about hey, you can't take League of Legends and adapt it to a mobile format. Yet here we are with League of Mobile being released last year, but teamfight.

Speaker 1:

That's exactly what happened, but they learned the lesson for sure. They were like all right, you guys know mobile, you guys make the TFT mobile port.

Speaker 2:

Don't think we have a ton of compelling reasons why it seemed to work in China. I was talking to Javier Barnes who wrote a great Deconstructor Fun piece on ClashMini right now and I don't think he had a great explanation necessarily for why TFT in China worked, except that it's a different market and they're used to more sophisticated experiences. I'm not ready to buy into that. I think the West is plenty, plenty into sophisticated experiences. They're just talking about Magic, the Gathering. It's a pretty sophisticated experience. They're into 4x games. Those are pretty sophisticated experiences. I'm not quite ready to buy that thesis but nonetheless it looks like ClashMini isn't going to be able to solve a lot of the autochist problems, particularly on mobile. They had a format where you would build a deck and you remove most of the economy pieces. You build this deck, you play these units. The units could move in between rounds. It became this mind game about okay, this unit's going to charge, do I think there's going to be a unit on my opponent's board that I'll be able to charge, or are they going to move that unit away? That was a super interesting game. It looks like they didn't think that was going to be successful. They issued an update earlier this year where units would be grounded in particular places and they've kept that alive and you still can't move units. They've just essentially dumped a bunch of units. In this update, I think a lot that they had in the warehouse.

Speaker 2:

I was actually on an auto-chess game earlier this year. The thing that no one can still solve beyond just making the core gameplay compelling, is out-of-round progression. No one knows how to add out-of-round progression to auto-chess games. I think that was a lot of why underlords couldn't take off. How do you add power to a game like this? It's very hard to add power, to add vertical progression One of these games horizontal progression seems to be on the table. Like in a CZG, you get a new card, but it's so much more expensive to make an auto-chess unit than a new card. They've just struggled to figure out what the right content pipelines are for this game. What is the right progression ultimately? What is the right modernization method?

Speaker 2:

The thesis we had on this project I was on was what if you took Squad or PG's like Marvel Strike Force or Star Wars Galaxy of Heroes? What if you actually just added auto-chess mechanics onto that so you would have the normal gear system, you'd have Ascension. You'd have a character collector, but you would also have this like deep real-time not even real-time, but you'd have this deep combat. We didn't get very far, I don't know. We saw some extremely high CPIs. I still think there's something to that Like to see someone take a shot at it, see whether or not there's something towards that thesis. But Clashmini I just uninstalled it yesterday. It was too addicted to draft. Free drafting is not healthy for us, gentlemen. Someone needs to take that away from us. This is why I was into Valve's game. I don't know if you guys ever played Artifact Valve's.

Speaker 3:

I fucking love that game.

Speaker 2:

I thought it was like one of the best CCGs of all time.

Speaker 1:

Oh yeah, it felt too much. It was like a lot of options that weren't that meaningful. Scroll across all three boards Eat.

Speaker 2:

They fixed some of that in a second update, but it was actually set somewhere between I would argue it's somewhere between auto-chess and CCGs, so it actually was like a forerunner for, I think, a lot of these problems, like if we had lanes, which is interesting, and similar to that. Anyways, we don't have to go back into Artifact. It was a Richard Garfield game, crater Magic, and he, by the way he's cooking something new these days. He has a game in soft launch.

Speaker 3:

Yeah, that's how you heard. It's not Web 3, right, not Web 3. I'm remembering something else. No, he was on a wrestling, he was on a wrestling Web 3 game on Wax.

Speaker 2:

That was like a CCG and I tried to play it because I'm like, oh, Richard Garfield, which is what I think they wanted everyone else to do, but I had to pay too much. Just fucking disasters. Fucking Web 3 infrastructure disaster.

Speaker 3:

Is Wax a blockchain.

Speaker 1:

Yet another one.

Speaker 3:

Yeah.

Speaker 1:

Speaking of auto-chess progression, I highly recommend checking out God King Castle. It's a mobile game, it's a single-player auto-chess game, but it's got all this classic squad RPG progression.

Speaker 2:

Oh, so someone has figured out this. What is it called again?

Speaker 1:

Yeah, god, king Castle. It's very foreign, like everything's poorly translated. But yeah, it's single-player only though, which I feel like limits it a bit and you hit like progression walls. But at least on the modization front they figured something out.

Speaker 3:

Is it mobile? Yeah, yeah, yeah. I'm surprised they slay the spire type of thing doesn't work for these games.

Speaker 1:

I mean, how does slay the spire, monetize right.

Speaker 3:

Oh yeah, I guess monetize. I don't know how they monetize because I play it through Game Pass. They have mobile right.

Speaker 1:

Yeah, but they just sell a box price product. Yeah, it's like $20.

Speaker 3:

I think they can do that because they're like an indie superstar. Okay, fair enough, fair enough.

Speaker 2:

This looks like they actually did exactly what we were trying to do, yeah does auto-chess have?

Speaker 1:

this monetization have the same problems that rogue-like monetization has? Basically, it's the in-and-round progression versus out-of-round progression problem.

Speaker 3:

Wait, what did you say?

Speaker 1:

It's similar to rogue-likes right when.

Speaker 3:

Yeah, it is a rogue-like, it's just a CCG or it's like a trading card game. Actually, I played the shit out of that while I was on vacation, just while we were waiting for stuff at the airport and stuff. Yeah, I love that game. I gotta play this. Eric King, God Castle.

Speaker 2:

Should we talk about surprise and suspense and beliefs that entertain?

Speaker 1:

Yeah, let's do it.

Speaker 2:

Cool.

Speaker 1:

Yeah. So this came from a big problem in game design at League of Legends, which is game pacing. Right, game pacing is how do you, where do you stick the content of your game? How quickly do players level up? How quickly do they find new areas, new bosses, new items, new weapons, whatever, you know? If it's too fast, you know you, it's all front-loaded and they, you know they consume it all. If it's too slow, they don't stay until the slow trickle.

Speaker 1:

I think Phil's referred to this as like a cake-eating problem. He's like, how quickly do you dole out the cake? And so this is the problem in all games, but in competitive games in particular. In a game of League of Legends, everyone starts at level one and over the course of the game the players level up, accumulate power and then bash their heads into each other until somebody wins. And so in a competitive game, pacing is okay.

Speaker 1:

How impactful is the early game? If you get an early lead? Does it mean you're pretty much guaranteed to win? Or how likely do you make comebacks? If you stick all these Mario Kart, rubber band comeback mechanics in, like devalues the meaning of getting an early lead If some wacky, game-breaking shit can happen blue shells and bullet bills. And so this is the big question of do we need more comeback mechanics? Do we need early leads to be more meaningful? Should we speed up the game, slow down the game, that kind of stuff. And it was very feels-based for a while, and I think one thing I set out to do while I was there was how do we measure, quantify this and say, oh, these game states are actually more comeback heavy or less comeback heavy? And, moreover, do players want it to be more comeback heavy?

Speaker 1:

But inspired by this paper Surprise and Suspense, which essentially does this, for it creates this theoretical model that, for any sort of imagine a sports game or a political election or a mystery novel, any kind of spectator event where you're like you're not sure what the outcome is, but you want to watch. You want to watch the two sports teams go back and forth and have your heart fluctuate, as you know the chances of winning, and go up and down. Why do people like watching these? And so the way they instrumented, operationalized, is if you're watching the Lakers versus the Miami Heat or whatever you can imagine, it's like a. You see a probability of winning that starts at, let's say, 50% and over time, you can imagine as you're watching Lakers get a lead, the probability goes up, they lose their lead, the probability goes down and there's this path of probability of winning until it reaches at the end. It either goes to one or zero and they say, okay, and what types of paths are more exciting? Presumably a path that goes up and down a lot is more exciting than a path that kind of just goes straight. Presumably a path where there's a lot of large jumps maybe is more exciting than one where there's just a lot of small jumps and yeah, so that kind of stuff.

Speaker 1:

And yeah, funnily enough, this guy, josh Tastoff, reached out. He read the same paper and was like hey, I want to do this League of Legends data. And he knew the design director at Riot at the time, tom Cadwell, from college. Anyway, I pulled a bunch of data for them and, yeah, they were able to essentially find some kind of interesting results. The first is that kind of consistent with loss aversion. People like to believe that they're losing and then win. They would rather think they're losing and then make a comeback than to just have a straight line to victory, and I think that's intuitive. But on the flip side, when they lose. They would actually prefer also to be thinking they're losing and then lose, rather than to think that they're winning and then suddenly have the game swing. You choked, you threw away the game, you had a huge lead and you lost, which seems contrary to the other point of people like to make a comeback. But people like to make a comeback. They don't want to come back made on them.

Speaker 2:

Is that the most damaging of all? Is that the most utility loss Is when you so? You said, players have a preference to think they're losing and then come back and win, and I guess the worst position to be in is to think you're winning and then have someone come back on you if we were to flip it. That's poof, poor Atlanta Falcons. Right now there's some Atlanta Falcon fan is crying.

Speaker 3:

Eric, quick question. So it's to me. The big question here is what is the? What's the dependent variable when you're running this empirically? And is it a latent variable? How do you? I think I understand your character. You have a list of characteristics, some like vector X. It's a bunch of different things that might be like the state of game and in time period one and the state of the game and time period two. What's the like? How are you measuring enjoyment? How do you? How did you determine people like it more when this happens?

Speaker 1:

Yeah, so outcome we used here was did you play another game and how many games did you play in the next X hours? So the assumption being, if you enjoyed your previous game more, you're more likely to play follow up games, and this was the big thing we did that most other things were like theoretical or they use like spectator data, but this paper we had a tons of that actual behavioral data. Yeah.

Speaker 3:

Okay, so it's more like what types of conditions game conditions lead to higher conversion. Converting to another game.

Speaker 1:

Yeah, continue to play.

Speaker 3:

Okay, oh, okay, that makes sense. So the underlying model being about this feelings and utility, but ultimately, empirically, a relatively straightforward approach that we would think of when we're doing our everyday job.

Speaker 2:

Yeah, and all this is super consistent with research on match threes and near losses being some of the most exciting moments in match threes, when you just think that you're about to run out of moves and lose and you find that one move that lets you win and you're down to zero, like that's something people are actively designing towards and it totally makes sense to me as a design principle.

Speaker 2:

In a PvE environment Like this is how you should design things. It's like you think you're losing and you think you're going to win, like you overcome something. That is just to me seems like a great revelation when it comes to progression. I guess the thing I'm interested about with this paper is, in a PvP environment like all this stuff is immediately mirrored towards the player who loses. So I'd want to say your design advice would be like make sure that a player can come back no matter what, like there isn't a lot of path dependency in victory or in loss over the course of a match, but because it's so damning also to think that you're going to win and then lose the last second, can you take away any PvP lessons? Yeah?

Speaker 1:

So that aspect of it is a little zero sum. But there were some other parts. Like you said, if one team has a comeback, the other team had a choke, necessarily. But there's some other elements that are could I don't know what the term is positive Some here, for example, in the original papers, surprise suspense they talk about two different. They define surprise as a change in the probability of winning and suspense as an anticipated change, variance over the next period's time stuff. So let me give you an example. So in basketball is a very surprise heavy game, which is that there's a lot of points scoring and so the probability of winning is moving up and down by small increments very frequently. Soccer or football, both kinds is a very suspense heavy game where there's not a lot of scoring that happens, but when it does you do score. It's a big deal Right. So like the score in a game of football is like pretty flat, but then when you make a goal or a touchdown it swings dramatically and so there's always a small chance of a huge swing.

Speaker 3:

Okay, so it's. The probability of winning is much like in time period two or time period t plus one is much narrower for a more surprising game or yeah, let's call it basketball versus football, yeah. Yeah, okay, that's pretty sweet. Yeah, eric, can I ask a question, since Phil is here and Phil is the killer of behavioral economics and I haven't read through the paper, but I'm just scanning it. Scanning it and I see prospect theory in here.

Speaker 3:

And that might explain this idea of to think you're losing and then win is much better than to or, and then it's just as bad as, or is maybe worse than to think you're winning and then you lose Is how much does the prospect theory play into this? Prospect theory, obviously being like losses are worse than gains.

Speaker 1:

Yeah, they found results consistent with that. That and that the the effect of the the loss side was bigger. Choking a lead was worse than making a comeback Like the. In terms of the effect on your next gameplay, interesting.

Speaker 3:

So we think about this from like the outcomes point of view. You'd be like if you were winning and then you lost. Your probability of moving on to the next game is like 10% worse than otherwise, versus if you were losing and then you won. It's 5%. So it's just like more.

Speaker 1:

Exactly, yeah, you got it.

Speaker 3:

Now, phil, is there a neoclassical explanation?

Speaker 2:

Well, one thing I want to clarify is that loss aversion is not something that behavioral economics can steal from neoclassical economics. It's a neoclassical result as long as you assume a diminishing marginal return utility curve, the n plus one step on that curve going to be higher than the n minus one step to downward, because remember that it's concave the curve bends inward, so going backward is always going to result in more utility loss than going forward again because of diminishing returns. So that's something that I think is commonly misunderstood as a behavioral result, when in fact it's a neoclassical result. So I would always put that out there. I think we could. I don't know if we want to call it behavioral economics, but I do think there's something to expectation formation and how people have results that are below or above their expectation formation, and so it would make sense to me like we could develop, I think, a lot of really nice models here about expectations.

Speaker 2:

Like to me all of this is about expectations. It literally is If you expect something's going to happen and it's bad, then you are safeguarding yourself. That makes sense, like you're building a defense, and I guess if something, if you unexpectedly win, it's all about shooting above and below expectation, like those are the things that have the most powerful emotions. That seems to make sense to me. I don't know if that's a behavioral, though, or not, but that makes intuitive sense to me. But yeah, like, yeah, the worst result is like thinking you're going to win, then losing it the last second. Yeah, it sucks. That is a pre-reversed for your expectations. So I think everything is above and below expectations. Right, that's where, like, the true utility changes come from. That makes sense to me.

Speaker 1:

But yeah, the results work. I think that the interesting part from the game perspective is we can actually change what this like, what the shape of these curves looks like, right, whereas, like you can't really change basketball that much. But in a game you can add in a comeback mechanic, right, and you can say, oh, actually we think that, or we want the early game to be less impactful and the later game to be more impactful. I think empirically you see this in like Jeopardy, for example. Right, in the first half of Jeopardy, the point, the questions are worth half as much. In the second, are they harder questions? No, right, it's just, it's more exciting that way. But in a game of basketball if you said, oh, in the second half the games are worth twice as much, I feel like people would get upset. But yeah, so we actually use these on league to make certain decisions.

Speaker 1:

For example, there was this feeling that the early game was too fast. So, broadly MOBAs, there's kind of two phases of the game. The first game called laning phase. You're going to your three lanes and you stay there against one opponent and you fight them. At one point each one of you has a tower that you're defending and if your tower goes down, it opens up the map and then it goes into the teamfight phase, where everyone kind of groups up and runs around and tries to kill each other.

Speaker 1:

And so we were like, hey, actually the early game is too impactful. People with early leads accumulate a lead and it's hard to make a comeback. They made the towers bulkier and harder to kill and we could measure this. We could say, okay, the size of early leads went down by this amount and this frequently the teams are able to make a comeback and so, yeah, that was. One nice outcome of this model is that we can actually measure these time series in the game and actually, rather than someone just saying the game feels faster now, because I know we can actually measure that the game is faster.

Speaker 3:

So how did you use that? How successful were you in taking that and applying it Like it's working with designers to try I don't know make the game more exciting and to create these moments of suspense?

Speaker 1:

Yeah. So the nice thing is, in terms of organizational buy-in, they already wanted to do these changes, but during this time there was this real data nerd guy in the leadership team who was telling everyone you got to measure your results, you got to measure your outcomes. And they were like I don't know how to measure the game, feeling more early game focused, and I was like, hey, I can measure that. So thankfully there's a lot of buy-in.

Speaker 1:

But once we got that, like all of their other future major systems changes. So, for example, they wanted to make a game ending mechanic, because there's this problem sometimes where the defender's advantage the defender will just pull up and turtle and it's hard for the attacker to finish out the game, even though they're winning, and it turns into a slow grind. So they added this kind of coin flip mechanic, like this super dragon, that if you kill the dragon, your team gets super powerful and you can just steamroll the other team. But it gives the defender a chance to come out and try to steal it and steamroll the attack. And, yeah, we were able to measure that, like, actually there's more comebacks, games end faster and more conclusively, but also it creates more aggressive action. And, yeah, so the nice thing is we now had a way to measure the major systems changes they were making.

Speaker 3:

That's. I think that's half the battle when you're doing any sort of empirical work, like within your organization, is making sure that the department that you're doing the research for it has buy-in, because a lot of times you might get something across your desk and they're like oh, I want you to just tell me something about this, can you research this? And that's straightforward. And you say and it's typically sometimes it'll be like looking in historically at something and you'll say, oh yeah, this is what happened. It's a lot harder to say, oh, I think we could, I think we could measure this and come up with a prognosis and say, hey, if you do X, y and Z, it's going to improve the game. Especially to tell that to a designer, that could be super challenging, at least from yeah, especially when you come in at the data.

Speaker 1:

You don't know anything about game design. Don't tell me what to do.

Speaker 2:

This conclusion around having some sort of mechanic to end one side of games early and to extend closeout games is such an obvious result and I mean that in a brilliant way that it's incredible we don't have more mechanics that we do this. There's the concede mechanic in a lot of MOBAs now where, if enough people agree, you can concede out early, but nothing's worse than just sitting through losing, and it's such an obvious result of the paper that it's incredible that more games don't implement something like this. The ways to concede early or to extend games that are close, like overtime, also sounds silly that there aren't more overtime mechanics in a lot of these games. There should be far more than there are. If things are getting intense, like you want to stretch out, you want to stretch out that utility, you want more overtime. It's crazy how little of that there is. There's none of that in magic.

Speaker 1:

Yeah, to be fair, in a single player game you don't have the four fitting problems that you have in a multiplayer game, or a team game, no for sure. And so I think team games suffer the most from this problem, where maybe two people want a four fit, the other three don't, and then you're stuck losing and two people don't want to play, but the other three do.

Speaker 2:

No, that makes sense.

Speaker 1:

But yeah, I think one other interesting extension of this. So you think about a slot machine, like you put a coin in a thing and then some number of coins comes back out. There's no reason for any graphics or rolling wheels or like the three sevens lining up, Like it could just be a black box and you put a coin in and then it just tells you what number you got out. But that process of the wheel spinning and the gradual information getting revealed is exciting. Or if you think about a lottery scratcher like you could, it could just be you scratch one square that tells you how much you got, but instead people scratch all of them and then they have to like line up the numbers and count it to three sixes or something. And yeah, I think it all points to this idea that the process of gradually revealing information is exciting in and of itself and on top of that we can design that process to be more.

Speaker 2:

So you have this like crazy claim at the end of the abstract that says we find that optimizing the information revelation process could substantially increase player engagement. The average effect is 40 to 99% as large as the effect of making a losing player feel like a winner. Where does that come from?

Speaker 1:

Yeah, I don't actually believe that claim they're cool. They basically looked at the time. Yeah, so it was an interesting idea. Which is? They looked at the pads on League and said, okay, if we made our own path, what could be better? And the thing they found was in the first five minutes, very little happens, and therefore, if you made more stuff happen in the first five minutes, the game would be more exciting. But, first of all, you can't make it. The point of the first three minutes is people are people are connecting, some people have slower internets, it's a waiting room. It's not really when the game action is supposed to be happening. And then, on top of that, I think all the results are extremely local. You're measuring all the in sample data is all the current state of the game, and so I don't know if you can really extract that.

Speaker 2:

So you guys, do you do derive an optimal game path? Yes, interesting. See, now I'm gonna take into this paper.

Speaker 3:

Okay so, eric. So this is a very interesting paper to us and I think it's also very interesting to people in this industry. I'm curious. I think we all know at least a few people who are trying to get papers published in this on these topics within academia in particular, like business analytics, or like econ was like a pretty popular place where you would get published, like information science, these types of journals, like marketing, research. How has that experience been for you guys? Where do you see it publishing? Where would you? This is more of a I know what I've gone through and I'm curious to hear your thoughts on in terms of the paper.

Speaker 1:

Yeah. So honestly, I'm not too plugged into the academic side of it. I know that they struggled to, so at first it was like, hey, we got a mountain of data, we can definitely do something with this. And then we ran all this stuff and then it was like, okay, but what are the actually the interesting conclusions? Because it's about games, but it's specifically competitive games, so it's not like games industry as a whole. So they took just a pile of behavioral economic stuff and threw it at it. They throw five different behavioral economics models at it. So I think that was their strategy let's throw all the behavioral stuff at it, see what sticks and then try to publish based on that.

Speaker 3:

Who is it Is it like? Because a lot of the especially theoretical papers from like the 80s and 90s especially neoclassical I don't know so much about behavioral, but maybe behavioral. One of the things that that sucked was like you come up with this model that people were like that's intuitive. I think I've experienced that and your model makes sense, but they have really no good data to study the problem. This is an opportunity to say hey, here are all these interesting theoretical models. We have data that actually lets us show you that the model works, which could be like that's generally interesting. So that's like top five publishing material if it's correctly situated. I'm curious if that is the approach.

Speaker 1:

I think so. The paper that inspired the surprise and suspense one was Pure Theory and so since then there's been a bunch of papers trying to prove it out. Most of those use sports and most of those use viewership data and much smaller sample size. So initially our shtick was like, yeah, we're going to validate this model using a much larger and hopefully better data set than just viewership. On a few hundred We've got a few million games. But I don't think that, yeah, that that particular theory, the surprise and suspense one, didn't get that much traction. So I think that's what started throwing other stuff at it. I do think that our results probably show in the surprise and suspense one that people like suspense more people like the football style like low chance of a crazy event happening, much more than the basketball style of like constant scoring back and forth, and that is consistent with some of the other papers.

Speaker 3:

But so is the takeaway for game designers to add more randomness to their competitive games.

Speaker 1:

Add a chance of a large swing happening. You want to. You don't want the ceiling on a particular play to be capped.

Speaker 2:

The most you can, the craziest thing that can happen, I think it aligns with what I experienced on Sunday when I saw multitude of pick sixes in the NFL. Like those are 14 point swings. That's one of the biggest swings, I would argue, in all of sports. It's why they're so exciting. One team is marching down the field Do you think they're going to get a touchdown? The other team intercepts it and returns it, and then they get seven points. It's 14 points. So I think it's huge. I don't think there's anything else like it in sports. Okay.

Speaker 2:

So, there was a piece that came out in the Wall Street Journal where Netflix is considering IEP and ads. A couple of interesting things about Netflix's gaming business. They spun this up in 2021. So far they've done a bunch of mobile games. They are hiring for a AAA HD shooter out of Los Angeles, which is fascinating. It's not.

Speaker 2:

I'm not quite sure how they're actually going to deliver this to Netflix customers, because you got to remember they don't own a lot of the devices that people actually end up using Netflix on. They're playing on Chromecasts, they're playing on Xboxes, they're playing on Roku's, but the thing is like, how are you going to get controllers into these people hands? I think there's a tricky problem to solve there. I'm sure they can solve that one. They haven't done anything in AAA quite yet. It's been mostly mobile. The way you get mobile games is that you have to go onto the Netflix app and then they advertise games to you and then, when you click on one of the icons inside of the Netflix app, it brings you to a separate webpage on the app store where then you have to download the game, which totally violates a bunch of app store rules about having app stores within app stores, but they've been grandfathered in because they're a really big company and if you look at the mobile download charts, netflix is almost always number one, at least in the United States, when it comes to downloads. They just have such a strong subscriber base so they've been able to earn a lot of sway with app stores.

Speaker 2:

But at the end of the day, their games program has so far I don't want to call it a disaster, but they basically have bought Riff Raff games and they've turned off IEP in them. They've got a lot of free to play games. They've launched them original games based on their IP. They have this choices segment of the games program where they have a lot of story based games based on like reality shows like Too Hot to Handle, but so far it hasn't really gone anywhere. They've amounted to about 1% of their active subscriber base playing games on a daily basis, which is just a disaster. They've spent a billion dollars on games it was reported in the last year which is an incredible sum considering they only spent 17 billion about on content. So they're spending 5% to 7% of revenue on games for a return that's less than 1% of your DAU. That is not a winning proposition. So there has to be some way out of this, and so the thing that I've been writing about really since I said writing about writing on the internet is why subscriptions have been such a disaster in games, and so it's not just Netflix. Nearly every game subscription has been an absolute shit show.

Speaker 2:

So we think back to online. If you guys remember that, that was an early attempt to do subscriptions and game streaming. The tech worked really well, but it looked like it was too early for its time. We had EA Play, which was a more full game subscription service from EA. It was originally called Origin. We had another service called Gamefly, which is originally doing Netflix style hey, we're going to send you a disk, you send it back and they also pivoted and tried to do full game subscriptions. Then we had another service that was called Game Tap. That was another disaster. We also had Stadia, which is trying to do the subscription streaming thing.

Speaker 2:

We have Uplay Plus, which there was an article that came out recently that they're rebranding. They don't seem to have gone anywhere. And we also had Xbox Game Pass, which is like the big hope that subscriptions are going to break out, but they started to stall in terms of their MAU growth. They're certainly not profitable right now. Remember they were running like these $1 discounts to get people signed up and then we had just like this litany of individual game subscriptions where you play an MMO Probably almost always an MMO they charge you like 10, 20 bucks a month and you get to continue to play in the game. Those are dead. Those are basically dead, except for World of Warcraft. And even inside of World of Warcraft you could remember, they've made two changes. One is that you can play any character I think it's up until level 25 at no fee. So they've moved to more of a shareware model. And not only that, they've expanded their MTX business, so there's a lot of cosmetic stuff you can now buy in World of Warcraft.

Speaker 2:

So a lot of this like a game subscription stuff has gone the wayside, which is, let me not forget, dunking on Apple Arcade, which has also been a disaster. We have no numbers on Apple Arcade. They haven't reported any quarterly earnings report, which should tell you that it's been a disaster. But the reason that I think people keep pushing subscriptions is rather obvious. They want steady revenue and I see a lot of this like thesis about like people don't actively engage in the subscriptions, they just forget about them. That certainly may be the case, but that is not a sustainable. That is not a sustainable bajillion dollar business, even if you're a gym, because the thing that happens is that credit cards expire. So you might get like a nice uplift from people not renewing their stuff, but credit cards are not going to get renewed. There's behavioral things credit card companies are doing now. That is not a business model. To run a subscription basis business on the fact that people will forget to cancel, that is not how you win.

Speaker 3:

Duolingo charged me like five times for the subscription this last month and I didn't realize it. And then my credit card was like, hey, these people keep trying to charge you for a subscription we're pretty sure you don't want. And I was like, oh, thank you, I don't want that and I canceled it. So, yeah, totally.

Speaker 1:

I mean, even if the price is like five X what you expect, you still got to get people to sign up for the subscription.

Speaker 2:

There's just. It's just not how you want to bet. You don't want to bet your business on like trickery. That's just not a long run sustainable model. And so the thing that I think has defined gaming is that we've created these insane engagement tales. So when you think about watching a Star Wars movie, so you go in, you buy a DVD back in the day and you pay 20 bucks and you might spend three hours watching Star Wars, and so you take the $20. You divided by the amount of hours, you can get a cost per hour of entertainment.

Speaker 2:

The thing that tends to happen is that you won't rewatch that same piece of content, or at least you might rewatch it. What? Maybe two, three times. And if we were to talk about the average amount of times a movie has been watched that someone has purchased, it may be 1.1, maybe 1.2. I'm sure it's. Even if we were to give it to two or to three, we're talking about maybe six hours total that someone has paid a fixed price for that piece of entertainment.

Speaker 2:

Now we think about games. The engagement tales. Wild, it's wild. It can be thousands of hours. If you look at my Dota 2 account, which is still embarrassing, I think I have a thousand hours on Dota 2. And so these engagement tales have the capacity to be so much larger than other forms of content, and I would argue that's because the type of content that you're consuming in games is so different than other forms of meat. It's not the same piece of content over and over again. What you're consuming in a lot of these games is PVP or even in a PVE environment. You're consuming a lot of the social interaction that happens, and so if you play a game like Clash Royale or you play Magic Gathering, you're consuming having to solve the puzzle of what is my opponent going to play, what is the best deck to collect right now? Or from in a PVE environment, what is the next build? What is the best build to have? So you're solving all these really interesting puzzles and these puzzles, ultimately, can provide thousands of hours of gameplay. And on the supply side, it might only take, let's say, two weeks to come up with a Magic Gathering card probably even less than that, but those two hours that maybe three people will spend crafting that card can provide thousands of hours of gameplay. And even if you don't utilize that card, you still have to solve the puzzle that card presents if other players put it in their deck. And so you have this kind of almost universality of the content affecting almost all player bases, regardless of whether or not you yourself utilize it.

Speaker 2:

And we think about the supply side of movies. It takes hundreds of thousands of hours and hundreds of people to produce even like one hour of Game of Thrones content, and so you have this like incredible amount of demand and, I think, this very low, harsh supply side economic piece, and on the other side, you have content on the demand side that just it just hasn't changed in any sort of way, and so I think that's why you've gotten these incredible engagement tales, particularly when you start to see the emergence of online games. Even if you had local co-op games, there was a ton of instances in which people were playing those for hundreds of thousands of hours, even like the original Pong when it came out it was one of the first games, the first video games to come out. People could play that for hundreds of thousands of hours because, remember, it was PvP you were playing Pong for someone else. You had the little joystick that you would have in your house.

Speaker 2:

I think arcades were an answer to the monetization question about. Okay, how do we come up with ways to monetize this engagement tale? Fixed price Wow, that's actually going to mean that we're only going to be able to correlate to a very weak degree the amount of time someone's going to spend on this product with a fixed price. We need to figure out some way that we can continue to correlate those two things we need to do monetization, and monetization that count in mass game has defined games. Arcades are the first answer to this. You're going to have to insert a quarter and we're going to charge you on a time basis. If you want to play for another 30 seconds and you need to insert another quarter, you need to pay. Insert a quarter. Insert a quarter. The problem is that they didn't have any long run retention Every time you inserted that quarter. It may be the case that you were continuing. That was the real banger is that I play time crisis and I get shot three times and then I wouldn't be able to progress to the next level. I need to insert another quarter. The problem is, if I come back to that same arcade, I would have to restart from the beginning again. That's a pretty harsh and expensive roguike. Now we have safe states Seems like a pretty clever solution to this. Not only that, we had the emergence of MTX and app purchases. Again, we found a way to be able to correlate engagement and monetization. We've been able to grow average revenue per paying user in a lot of these games and we've been able to grow the number of hours per played per game.

Speaker 2:

The subscriptions just portray all of these economics because you're fixing the cost of these. You're fixing the cost to customers while the engagement tails are still really long. You're just not capturing a lot of that tail. Ultimately, in a competitive environment, you're not going to be able to have enough capital to fund into creating more content for your subscription. Ultimately, I think the wheels just fall off. That's why none of these subscriptions have won. That's why I think Xbox Arcade, xbox Game Pass, is also going to fail. The economics simply do not work out when how games are consumed. It does not make sense.

Speaker 2:

I would really love for at least some of these subscription services to sit down and come up with an answer to this challenge. Why are you going to be different? Why is your service going to change the economics of this industry? Ironically enough, when you run through this model, the best thing subscriptions should do is actually just create single player games, which is a service that all these subscription services understand, which is okay. I'm going to go out and make Piki Blinders. You're going to meet 10 hours of content, it's going to scale to Y number of people and it costs Z dollars.

Speaker 2:

Single player game is the exact same economic model as other linear types of content. That's the safe model. That's how you diffuse. That's what a subscription is really good at. It's about building your subscriber base to 10, 20, $30 million and then being able to fuse those costs amongst your subscriber base and then just keep feeding the furnace with more and more linear pieces of content. That is what subscriptions are good at and, ultimately, that's what I think Netflix should go at. If you do choose to pivot and go live service, you're going to need IEP and ads. All right, I'm done.

Speaker 3:

It's more expensive, phil, to Produce new single-player games than to keep a live service game going. You think?

Speaker 2:

so that one more time.

Speaker 3:

I don't know. That's more of a question, it's more so. My understanding is that you're saying, because of the way, that, because of the subscriber model, it makes more sense to let's take one subscription model that has wow, that has a live service game that they constantly have to update with new content but players are paying every single month, versus a studio that puts out One new game a month that has like 20 hours, 40 hours of gameplay. Are you saying like the live service is more expensive than the then producing a new game every Couple of every couple of months, or something like that? Because your thesis is Produce single-player games? If you're gonna do a subscription to Nick with Xbox game pass where you get to play to Nick for free, it's 30, 40 hours of this fun experience and it's over and you move on to the next thing. Compared to them, trying to keep. Obviously fortnight is not owned by Xbox, but try to keep something.

Speaker 3:

Yes, what night wise, which is, yes, arguably more expensive. Okay, what's what I'm conflicted about as a consumer Is that I love Subscription, like I know that for a fact that I played more games than I would if I didn't have Xbox game pass, and I know that I spent more money than I would if I didn't have Xbox game pass. Because I'm like a one, one game kind of a guy. I like to buy my $60 game and play it for 100 hours, 200 hours, 500 hours. That's why I like Monster Hunter, that's why I like Dark Souls. I'm into kind of these games that I can just grind over and over and over again. This addictive experience.

Speaker 3:

For someone like me and this is probably more a question of what's the counterfactual that you would, what's the experiment that you would run to Really get at the bottom of this but for a consumer like me, I know that I spend what like 20 months, $20 a month, on game pass, so roughly probably just around $200 a year on game pass. I don't think I've ever really spent. I don't think I've ever, before game pass, bought three More than three triple-a games in a single year, and maybe that was the max. But for me, for a con, from a consumer's point of view, I really feel like this is they're getting more money out of me than they would otherwise. But I guess that's assuming that the alternative is a fixed box price, which we're not talking about.

Speaker 1:

That we're talking about probably microtransactions being the Alternative they're saying Netflix content is you watch five hours of TV and then you're done. Right, the seasons over, you're done, and then you move on to the next show. Why does that fit a subscription model better Then the game engagement part curve which continues on that long. So I is it because the variety.

Speaker 2:

So I think you have large, understandable fixed costs that have that that don't have. There there are no cost tails, and so the way I think about it is that you can diffuse that really Large fixed number over a wide subscriber base. That is why subscriptions work right. That's why you want to build subscriber base really large, which is why all these subscriptions have started to eat up prices, because they already built their base and Because costs are so diffused.

Speaker 2:

At this point you think about a billion dollar piece of content. If use it across 20 million subscribers, it ends up being like very little per subscriber. So I think that's like very understood and you can just continue to shovel coal into it. I would argue in a competitive equilibrium, because the correlation is always going to be weaker in a subscription service. You're never going to be able to properly fund the amount of content that you should fund into a live service. You're never going to be able to fund as much money as you should into a live service game. Because you're not able to correlate those two things.

Speaker 2:

Because, like, how do you we think about? There's this game that came out on Netflix that was a Star Trek squad RPG and they had removed MTX, and so there's really interesting question about okay, if you remove MTX or in-app purchases from a game, like, how do you rebalance it, which I think is like a really fun system design question, especially because we haven't seen a lot of like unpaid squad RPGs. The problems is like, how do you funnel more money into that squad RPG If you don't have in-app purchase and so you have to do is you have to think about, okay, how do you do attribution within a subscription service? And I think that's really hard and not understood and it's hard to figure out how much coal I should shovel into that and I so I would argue I have a, so in a competitive equal.

Speaker 1:

Can't you use the same thing they use now, which is watch time or whatever their engagement? Heuristic is like playtime. It's the same question of should we make another season of you know this show right?

Speaker 2:

at the end of the day, though, the amount of money that you can fund towards that project is always going to be the same, right? So if I increase the amount of, if I have 20 million subscribers on Netflix and I add, let's say, 200 million hours of engagement, 200 million hours per subscriber, I don't know some like crazy fucking metric revenue didn't move. So they're never gonna be able to like, shovel, shovel, more coal into these products. They can never grow an individual product. To me, it's okay, you've got to to.

Speaker 3:

You've got two games that are identical in every way. The only difference is how you're monetizing them. And let's take fortnight, for example. Fortnight monetize is really well. They, they sell micro transactions. They sell pieces of content that are zero marginal cost, relatively low fixed cost, and the biggest cost that company faces is Probably, if I had to guess servers, like that's probably the thing. They probably pay AWS like a bajillion dollars a month.

Speaker 3:

I don't actually know what their infrastructure looks like. Of course, to me you take like you take epic and it's like I would Probably well, I almost do hey a subscription on top of the actual micro transact like. To me it's like maybe Xbox just needs a better way of allowing me to spend additional money Inside their subscription service, because I love their subscription service and I probably spend more money Using it than I would otherwise. I'm struggling to. So my understanding is your thesis is like this is not the. This is not an optimal Monetization strategy for the vast majority of games. Maybe it's a good one for single-player games, very much like the movie. They deliver content, you, you finish that content in 10, 20, three hours, depending on what it is, and then they have to go turn it out. To me that's like actually not great and one of the bigger so the oh who's the sharp Tom's not sharp tech Chatecary guy.

Speaker 3:

Yeah, ben Thompson like his biggest critique is they just have this insane cost basis that they're working on, where they have to put out like 10 million, 100 million dollar movies and they have a fixed Number of humans on the earth who can consume that content. And that's, from my understanding, it's been one of his bigger criticisms. So it almost sounds like from what you said earlier, like that is the model you would pursue, but to me it's. I would go for something that's like a zero marginal cost, like once it's out, there Are the only cost of keeping it up or maintaining it, maybe giving it an update and, of course, the servers to support online play. I feel like I'm missing something. I feel like there's, there's some spice, there's like a, there's something that I'm I'm missing and I know you you have it.

Speaker 2:

Eric's challenge is right. Okay, how's the live service different than a season? You're only playing about one game at a time, maybe two games at a time. There isn't room to consume multiple pieces of content and you can consume multiple movies and move all TV shows because they have a fixed. They have a fixed end, and so that's why you need to keep shoveling coal is that there's an end point, and whereas a live service game You're only playing like one or two at one time, there's all this LTV that's just sitting there. That's just sitting there that you're not able to correlate with price. So I'm gonna play League of Legends for a thousand bajillion hours and you're telling me that it's only about ten dollars. I'm gonna get unlimited access to almost all the content and potential subscription world. You're leaving just so much value on the table. And only that.

Speaker 2:

There's all this weird economics that games have discovered. You know we were talking about Melbourne goods before the cast started. We were talking about like negative externalities, you know, and a lot of the cosmetics. If everyone walks around with a legendary cosmetic, it decreases my value at it. So there's a lot of like weird kind of club exclusionary good stuff that you also get with MTX. So I think there's live service dynamics. There's the correlation issue. It seems much easier if I'm Netflix to just sit down and say, look, I understand how the subscription economics work. I shovel in content that gets me Y hours. I diffuse that fixed cost amongst my large subscriber base. They consume it and they move on to the next thing. Games can fit in that model. They have fit into that model. They still people.

Speaker 2:

We've been talking a lot about live service. I talk a lot about a live service. Live service is growing but ultimately, like, at the end of the day, like rising tides have lifted all boats, like the single player games are still doing gangbusters. Baldur's Gate Holy shit, we were talking about Baldur's Gate. That game has been enormously successful. That's a fixed cost. And now, chris, you were talking about Dark Souls.

Speaker 2:

It isn't just the eight hour action player games that I'm talking about. Those are like the movies to me. Those when I'm talking about like a what would be the equivalent of a TV series that has a fixed cost. It's more like Baldur's Gate. It's more like kind of these PvE single player experiences which don't get like a ton of live service or content updates.

Speaker 2:

That still looks like hey, we can diffuse those costs amongst our customers and still get like a pretty good return on investment. That or the movie model, those are both things that feel applicable to a subscription Shovel. It move on next thing. So I don't have a silver bullet for you, eric, like I think you're right in some sense. It's just that you're leaving so much money on the table and other firms are not leaving the money on the table, and so they're going to be able to reinvest a lot of the additional revenue that they get from M4, iep and, ultimately, like those products are going to be better services than ones that you can maintain their subscriptions. That would be the best economic rebuttal I could give.

Speaker 1:

Yeah, I think that makes a lot of sense that they example of taking a mobile game with MTX and just stripping the MTX.

Speaker 1:

There's no way the design of the game is optimized in a subscription format Right. Games are optimized either to sell boxes or to sell MTX and ads pretty much, whereas if you and so adapt, you can't just shove that stuff into a subscription, assume the form of the content matches the form of the modernization, whereas I think when TV did it, when Netflix, like cable TV, was already a bundle right, no one would pay for a reality show DVD but or very few but a lot of people would watch it was in bundled into cable TV and so they had didn't have to adapt the content to the modernization medium anymore, or very much compared to games which have to make a huge difference. Maybe the best format is I'm thinking like a short single player story driven game, right, not a super long one, not like a JRPG that's 100 hours deep, but like a nice five hour, 10 hour experience and then move on to the next one. It sounds like indie, single player indie games to me.

Speaker 2:

I think it could be. I'd love to see him, like, take another shot. I'd like to see him take a triple A. I don't want to say triple A, but Tumor, tunec, whatever it is, shit out a couple of those. Figure out how the controller thing's going to work.

Speaker 2:

I would love that, I think it's where they've been going is the completely wrong direction, which is mobile. I think that's again. We were just talking about the competitive equilibrium. I think a great example of this is like match three. So they had a match three game with stranger things, but obviously, like they can, they'll never be able to attribute a lot of revenue to a stranger things match three and so they can never reinvest enough money, which means they're never going to be enough content.

Speaker 2:

And one of the big turn points we know in match threes when you run out of levels and so I actually think that's probably a better example of what we were just discussing is like King is able to invest in having they're up to level 15,000 or some crazy shit.

Speaker 2:

You're just never going to have the attribution under a revenue caps description model that lets you reinvest that amount of capital in creating that much content, and ultimately that it's going to mean you're not going to be able to attract a lot of players and maintain them relative to their opportunity cost. The only other thing that I think we haven't solved to Eric's point is like how you redo systems, design and without a lack of spend, because ultimately, like we've been optimizing towards payers, of course we're optimizing towards payers, and so, if you don't have MTX, can you optimize in such a different way that the experience is that more enjoyable? And the thesis that I have been like burning around in my head, that I haven't been able to write down in the right way, is that actually think? I'm not quite sure how to describe that yet, but I think there's something to that idea that we've not been able to articulate quite well yet.

Speaker 3:

You mean, like I spend money today, I'll show up tomorrow because I spent money.

Speaker 2:

I don't want to say a commitment device, because that puts me in behavioral like, not even a commitment device, like I want to get my money's worth or some shit like that you could see it as an investment Right, like I've invested this money and I'll lose my investment. It's almost like an expression Like I want to be able to express, express things like I. I got to think.

Speaker 1:

And if you don't have the friction of the paywall, then there's no individualism in your choice, I think it's almost, it's about payment allows me to almost choose my progression speed in some sense.

Speaker 2:

Again, I go back to match three because it's such a simple box for us to analyze. But how fast I progress and match three is a function of how much time, it's a function of skill, which is actually a very low coefficient, and it's also a function of payment, and the more I pay, the more boosters I can buy, and the more boosters I can buy, the faster I can move. There's literally no option for me to choose my own. You're removing one of the coefficients on me, choosing my progression speed. What if there was the?

Speaker 1:

dial you could set, you could ink. All the mtx was free.

Speaker 2:

So I think that would be super interesting, and we were talking about the couple casts ago is like, why can't you just choose your own difficulty in a match three? And that's, I think, exactly the type of thinking that Netflix should start to think about if they are going to try these mobile stuff is like, how do you rethink about a lot of these mobile experiences when you don't? How can you optimize them without payment being there? I think that's exactly what they should try, eric, I think, and single player games, to their credit, do that. I think that's the next thing I would take a stab at. I think you're completely right.

Speaker 3:

I mean that's, but the it's, that's. That's a huge undertaking. Let's optimize every single one of these games for this subscription model. To me, that's that's. The thesis that I'm getting out of this is that, like the subscription model is good for some games, it's not good for other games. The problem is, the nature of it is that you throw a whole bunch of different stuff into the same bundle without, at least right now, any regard for what are we actually putting in this thing and is it the right structure? It's like you're going to an all you can eat buffet for sushi and then, all of a sudden, like, you throw Kobe beef in there. It's oh, we can't sell Kobe beef the way we sell sushi. Like this sushi is this completely different product. So that's what I'm getting.

Speaker 1:

And modifying those games takes a lot of time and effort. That's another thing, that's like the ideal situation is you just stick the game in.

Speaker 2:

I think there's a much easier way to do this, eric, which is to just remove prices from IEP. You could just literally copy paste the game. Why not? Why not users just let them click IEP whenever they want, without it actually triggering a purchase in a subscription based game, I think.

Speaker 1:

Candy Crush with free purchases is actually worse than so. Why though?

Speaker 2:

Because isn't that exactly what you were talking about with difficulty? Doesn't that? Isn't that consistent with what you were arguing previously?

Speaker 1:

I think if you let players opt into their own difficulty, they often choose one that's less fun. Have you ever played a game where, like you had cheat codes and like you enter the cheat codes and then you're like I agree, you're like wait, actually, because we forced players to ration content we're almost we're better at, because people don't they don't want to naturally overcome things.

Speaker 2:

We have to force them to. Not I agree with you, but again that I think would go back to like we've almost inadvertently done that with modization and free to play, right yeah.

Speaker 1:

Hmm, but there's at least like anyway. Yeah, at least there's like a default difficulty mode. Right, the free mode is like the default difficulty and that's the one you're supposed to be on, as opposed to with when I piece free it's you could pick any difficulty. Do you think it takes longer to play every level in Candy Crush or to watch every episode of the Simpsons Candy Crush?

Speaker 2:

Candy Crush, easily, easily Candy Crush. Oh, absolutely, it's not close in my opinion.

Speaker 3:

How many minutes of there are there of the Simpsons? I bet there's less. I could be 35 seasons.

Speaker 2:

Let's say 22, 22 episodes a season. That gets you 770 episodes. Assume 30, assume 22 minute episodes, go 770 times 30.

Speaker 3:

The timecom has already done this for us.

Speaker 2:

That's 2014.

Speaker 1:

That's 60.

Speaker 2:

So there's 385 hours of Simpsons content. Candy Crush easily trumps that knife. Close, not even close.

Speaker 3:

Yeah, not only that, you can remember they have the live system.

Speaker 1:

Candy Crush saga holds 16,000 levels.

Speaker 2:

You get the live system, though, and that's also a time gate, so remember, you only get five tokens and they like recharge it like 20 minutes, oh we can assume that away, which just give them unlimited lives. They hand that shit out like candy Candy Crush. It's pun yeah.

Speaker 1:

Even if each minute level only takes two minutes, it's still over 500 hours. Jesus Christ Okay.

Speaker 2:

And I bet they're outpacing Simpsons too. They must be, because Simpsons has been around for 20 years. Yeah, all right, cool beans, we want to call it here. Gentlemen, you have a holiday too. You waste your time with me. Ha ha, gotcha.

Speaker 3:

Oh, it's the perfect opportunity. Yeah, I'm going to get back to my Reddit comments from our HVAC advice. We should teach this to our children Economics is major, major, major. Everyone has to major in economics. Number one for personal survival Economics is major.

People on this episode