What Can Video Games Learn from Collectible Card Games?

Hello. Um I'm Skaff Elias. I'm here to talk about what video games can learn from collectible card games. That's what CCG stands for. Um and first a little bit about who I am. Uh I started working on in game design on Magic the Gathering with Richard Garfield. We were uh classmates at uh University of Pennsylvania in grad school there. And um eventually over the years I became uh senior vice president at Wizards of the Coast Hasbro. Um and throughout the years working on game design. But sort of the unique perspective of the talk is how business and design are intermixed for this sort of certain class of games. Uh and first came to that realization in 1994. I actually became the business and brand manager for Magic. Uh and so that

that was a large portion of my job there. But I continued to work in design. I co-designed or developed almost all of Wizards of the Coast CCGs. Not all of them but the list is is very long. And part of a large part of what I did there was I designed the the Pro Tour, the organized play system, the tournament system, leagues that both Magic and then eventually Pokémon used. Um And there at Wizards after a while because I was going back and forth between design and business it turned out that uh really I was to borrow phrase if any of you know who the sports guy is, he's an ESPN columnist, the VP of common sense. That's basically what I was doing at Wizards. And I was I was very overworked. Um So in 2004 I left Wizards of the Coast, started a company called Three Donkeys with Richard Garfield. And it's mainly a consulting company and we've done a lot of computer game consulting and also design. We did a Xbox Live

arcade game called Schizoid with a company called Torpex. And and we've also published Spectromancer which is something like uh a collectible card game. Uh You can go look at it online. And also we've been teaching courses at University of Washington and I do a podcast with Richard where a lot of the things I'll talk about here are sort of gone over in more detail. That's at 3donkeys.com. So the interesting thing is by their nature distributed object games mix design and business in an inextricable way. A lot of concerns that are sort of traditionally in one field turn out to not be able to stay in that field. But first a little bit about collectible card games for those of you who don't know. Magic the Gathering was the first collectible card game. And um sometimes inside the video game industry people don't understand how large some of these properties were. Magic was doing

since 1994 so for a very long time between 100 and 150 million dollars wholesale and because of the the way the distribution margins work there the wholesale numbers that's actually about translates to 375 million dollars retail. And actually these numbers are sort of semi-publicly available and really they're probably about double what I'm what I'm putting here. Um Uh Yu-Gi-Oh Pokémon on the other hand our best year there did about almost 2 billion dollars at retail. Uh and it had several years pretty close to that. And it's interesting it hasn't quite gotten back to its peak but more than 10 years later it's it's doing you know more more than half that. Uh again. Uh and then Yu-Gi-Oh is sort of the top seller all time and it had several years where it was making at retail over over 3 billion dollars a year. Probably about 3 years in a row it came

it came close to those numbers. So they're they're actually very large properties and um and sort of the attitude that Richard and I take is you should look at all games all the time constantly because there's valuable lessons that you can learn uh that you can learn from them. Especially in a game that's as popular as these or major video games or even basketball. You know you really um something if that many people like something there's there's certainly some valuable lessons there. Uh so before we start um kind of a definition that I want to use for distributed object games just because mainly the lessons from the collectible card games apply to these sorts of things. And uh and uh it one of these games sort of should contain hundreds of semi-individually purchasable objects. Uh that together constitute a large portion of the game play. And semi-individually purchasable is is

actually kind of important. It means that I don't just pay one price. I get a game with hundreds of objects. And then I uh use those throughout the course of the game. That's certainly a multiple object game but not necessarily distributed where where maybe the players don't own all of the objects and they're purchasing them over time. And these types of games are interesting. They take a basic game play interaction, a standard set of rules, and they use these objects to change the game play in fundamental ways over time. And and and that's interesting because they become almost more of a hobby than a game or traditionally what you would think of as a hobby compared to a game where they're taking enormous amounts of time. People can be involved in them for a very long period of time, several years for even an average customer. And and actually they spend a lot more money than they might with a standard game because they're putting in putting money into it uh across a long time. These types of games are very difficult to construct. Um and enormously

difficult to construct. They're also enormously difficult to play. Magic the Gathering is except for basketball the best game I've ever played. And and if if I wanted to teach someone how to play and I knew they could only play it for one day I wouldn't teach it to them. It wouldn't be worth their time because of the amount of difficulty there is in getting up to speed with these things. However on the back end they're extremely rewarding for both the publisher but and the player as well. You almost can't have one without the other because you you get to um really have a almost a lifelong involvement with these types of things if they're done right or at the very least several years. So you get a very big payoff for getting into the game in the first place. Um first a few assumptions. One thing is that I'm only going to talk about game play concerns. What game play lessons can be learned from video games from

CCGs. Uh there's a lot of other things they could learn as well but just for the scope of the talk we'll talk we'll concentrate on those. And also um what type of video games can learn lessons from collectible card games. So I'm going to make just a few assumptions here. You can relax these assumptions and there's a whole other series interesting series of talks I could talk for a semester on this sort of thing but just to try and narrow the focus. Um The first one is a repeat purchase revenue model. Uh and as I mentioned before with a distributed object games um not just that the object get parceled out over time but that people are making purchases over time. Uh essentially you're creating an operating system and providing sort of new software for it constantly throughout the course of of your of your project. Uh it's very difficult for players and designers. So it needs to be built very robustly because if there is no long-term involvement it's it's a very dubious

value proposition. Again both for the publisher and and the player. Uh The fact is you don't have to purchase it all at once is a huge benefit for the player. And they get they get to try it first before they decide how involved they want to be. And and then hopefully ideally they're getting rewarded at virtually any level of money they decide to input into the game. Uh one thing about this is it requires actually a very large degree of sophistication in the audience. It's very difficult to fool people for 2 years and several thousand dollars of their money. Uh and uh this is something that's probably underrated even for a game like Pokémon where maybe only 20% of the people played, maybe only 10% played correctly. Uh those other players were extremely sophisticated about their purchases. I shouldn't say players. Those other customers were extremely sophisticated about their purchases. They knew more about Pokémon and what they were getting in their cards than you know 90% of the

average video game player knows what he's actually buying a video game. They knew the prices of the cards, they knew what they were getting, they knew the back history. They were allowed to associate with this enormous amount of story. And the reason they had this degree of sophistication is they were watching the television show or playing the handheld game. And so they had actually in some sense put in hundreds and hundreds of hours of research into the brand. Um and by respecting that sophistication we were able to provide them with a much better product. Um Another assumption, this is a very personal bias of mine. It's limited SKUs. Uh I've seen a a lot of times the sort of analysis paralysis on the part of the customer. And so so I really like it when there's very few things for them to buy. And it does imply since you're selling hundreds of objects that there's some sort of randomized selling going on.

Um Uh It doesn't have to be fully random. I mean, you can be creative in your product. For instance, when people are purchasing new cards, they could get a random set of them. And then the next time they purchase, they could get a random set that they don't already own. Uh you can change rarity levels. You can do all sorts of things uh here. It's not fully random, but um but the limited excuse apply some degree of randomness in that that actually has some implications later on. Uh and another assumption is that you're planning for success. You um you are expecting your product to be around for a long time, let's say at least a year of people act you actively supporting it, they're actively purchasing. Uh and also you're planning for a very very wide range of customers. Um In terms of the amount of skill, in terms of the amount of free time they have, uh their psychographics, um the amount of money they can spend. And uh and another uh thing when you have a game that's lasting this long is since your customers come in at different entry times, that's actually um that's

actually another very important consideration when you're trying to build these games. So, what are the implications of these assumptions? Uh For these games, the purchase decision of the consumer is not made once. Uh one-time deal heavily influenced by the marketing guys who have put a big spend behind your initial launch. They're constant decisions. Absolutely throughout the course of play, at any time someone's able to make a purchase. Uh Every single game is a sales pitch. Every single game that you have that someone plays, you could you should be thinking of in a way as a sales pitch to uh that customer because they can just turn right around and make another purchase decision. And also respecting all your customers pays off because you will have an enormous amount of them. And uh if you start cutting off sections of that market, uh pretty pretty quickly you'll you'll see the effect in terms of revenue. Um If every game is a sales pitch, uh this

is one thing to learn from a collectible card games. Um Uh the objects actually have to be pretty easily understood. Uh your your audience um your audience is pretty sophisticated. Uh they have um they have a put a lot of thought into the game. And uh and they will tend to, again, over time put a lot of thought into their purchases. And so uh clarity is very important. You really need for them to understand exactly the benefit of their purchase. There are other ways uh you can somehow obfuscate to them what the real value of these objects is, but uh let me just say the easiest way uh for them to uh the easiest way for you to have them see value in what's going on is actually give them value. And uh and clarity in what they're getting is a is a very important thing. Which is to say uh so for instance, if you had objects in your game that acted like black

boxes, and you were selling them uh Tamagotchi-like things, maybe with some hidden characteristics, uh the value of those purchases as it goes over time and your involvement is um it's very hard to make a case for people over time that they're that they're actually getting something uh of value. Um So, this is a is an important one. This is uh it's was really critical for us um uh in producing these games that lasted I guess Magic now is almost 20 years old, 18 years old. Um Uh for objects to have a clear benefit across the different entry times and the different purchase levels, as we sort of were talking about before, um uh the value of every purchase in terms of winning percentage ideally should only have logarithmically increasing value. And the reason for this is sort of obvious. If I came into the game 5 years

ago, and I keep buying cards, and now you came into the game today because you just turned 13 or 14 or whatever the low end of your age range is, um if these objects that I've purchased over time all have some sort of linear value, there's no way you can catch up as a new player, a new a new customer. Um That that's sort of the most critical case. Uh and then uh so you cuz you don't want to cut off your new customers coming in. And then additionally, even two people that have entered the game at the exact same time, uh just naturally one of them probably will spend more than the other. You want your players to get a benefit for spending extra money. That is after all uh how you make your money. Uh but you uh you don't want that benefit to be too high because if it is, it scares away other customers. Uh and sort of the solution to this is this logarithmically increasing value. There's a lot of ways to accomplish this. Um the easiest way is that you have some

sort of uh uh slight extra benefit for the combinatorial variety that you're able to get by purchasing a lot of objects. Each object isn't necessarily that you purchase more powerful than the next one. Maybe they're all of equal power, but uh but by owning certain subsets and increasing subsets, you you sort of the natural tendency will be to uh to have a curve like that. Uh and also um it it may be useful uh to reduce the amount of physical skill involved in your game. Uh when games have a lot of hand-eye coordination, uh then you are much less in control. It's just this extra variable in terms of the difference that your audience skill level that you're not in control of. And uh and people can have very very very wide ranges of uh of hand-eye coordination, especially uh if you're talking about uh a game

that appeals to uh uh a a wide age range. And these sorts of this sort of thing is is sort of important because you um you really it's really difficult for a person to see the benefit of uh let's say I'm making a first-person shooter, and this gun is 10% better than the other gun. Well, that may only matter if I'm uh playing against someone who's in the exact same skill range that I am in. And that might be a very small fragment of the audience. Most likely, if you gave me a gun that's 10% better than another gun, uh I'm just going to get killed anyway. So, the value of that purchase is very very minimal for me. Uh and similarly, if you're at the top of your game, uh those sorts of purchases may not matter again in 99% of the contests that you're in. Uh This isn't a hard and fast rule, but it's just it's it's sort of critical to realize that you um you definitely reduce the value of incremental purchases for your customers in terms of their win percentage uh the wider apart they are in skill.

Um Another really uh important lesson that we learned, and uh and we actually uh at at one point when we were doing Magic, we were uh concerned about reducing randomness in the system. And uh and the more we tended to do that, we we did actually notice effects of uh of, you know, sales essentially going down. Um For small audiences, uh a high degree of randomness in your game is is is uh is critical. Now, what's high? I mean, I'm not saying that you should be flipping coins. There needs to be skill there for one object to make a difference over another one. But a higher than you would sort of normally think because for a a very small audience, um it it it is uh you want people to be able to play against each other. You want your, you know, your bottom 10% tile to be able to play against your 80th percentile guy. And without randomness, that's hard to do. Um You uh

Some people think that think of randomness is the opposite of skill. And uh that's not true. You can perform a a slight thought experiment by considering, say, the game of chess. We could um take a game, let's call it Rando Chess, where uh you play your game of chess, and then at the end of it you roll a six-sided die, and if it's a six, the loser wins. Uh may not be the greatest game in the world, but uh but then you sort of think about it, and you say, "Well, is there less skill in that game than there is in chess?" Well, no, not really. Uh every little bit of skill that you had in chess, you still have in that game. Every opening you need to memorize, all the games you need to study, they all still apply. And even more to the point, the exact ranking of chess players in terms of their skill in the world would be the same if that were the game of chess. Uh what are you losing? You're losing accuracy in skill testing. That's the only thing you're losing. It might take more games of playing that for you to determine what the proper rank order is,

but um but it's still the same rank order. And still every single skill applies. You still want to win. Uh and so uh so that's sort of instructive. Obviously, that's a terrible game because the randomness that's been added to that game uh is very overt and very unsatisfying when the loser wins for the loser. Although, I'd still be plenty happy if I beat Kasparov anyway because he might be pretty mad. Uh and and and so um so that's the sort of thing you shouldn't think of randomness as the enemy of skill. It's just another completely separate characteristic of the game. And if that's blended into the game along the way, even in a game of high skill, um you get the effect that you have in poker. So, you know, uh you you you have a very wide set of people that are able to play with each other and get positive feedback along the way of determining who the better person is. Uh if I play 10 games of chess against a better person and I'm winning zero of them and maybe once I came a little bit

close, that's a very unsatisfying experience. I play those 10 games of poker against someone who's better than me and I win three or four of them. Uh whatever figure games is some set of hands uh or or 1-hour blocks in a 10-hour session. Uh that's actually pretty interesting for me. That keeps me involved. If you ask a a tournament level poker player if they would want to remove that randomness from poker and have it be more skill testing, they would they would have a heart attack, right? The whole reason they're able to make money long-term is because the game is fun for people who get feedback of exactly how close they are to that better person. Of course, they have to pay for that privilege of finding that out. Um the other value of randomness in your game is that if you have an astute audience, uh it uh it's very helpful in maintaining the discussion about the exact value strategic value of the cards. Uh when you don't have anything muddying the feedback for many many

games, uh people can coalesce fairly rapidly on what the exact best cards are what the exact best strategy is before you have a chance to update and change that and keep the game fresh for people and that's a disaster. You don't want that to happen because again, that's a disincentive to purchase uh for people that have already bought, you know, the elements of that strategy. Um another thing is if every game is a sales pitch, uh how can you go about adding real value to the objects for your players? Uh and this is an interesting one. You're not trying to through the course of this is sort of going to sound like, oh yeah, you're just trying to get as much money out of them as possible and that's true, but the best way to do that is to actually give them value for the money that uh that that they're that they're giving you. Um so it's not some sort of evil plot, it's uh it's I want to give them something that they want to buy.

Uh so adding real value to the objects is a is a interesting question. Um one thing you can do is you can cash out the objects. When we first started um with these collectible card games as many of you know, there was a sort of a thriving secondary market uh in the cards um because specifically because they were purchased randomly to begin with. So you didn't have all the cards you wanted. Um some people you would get a sort of a better overall value if you're buying them randomly, but if you really only wanted specific things, you could just buy that one card. Um the people at the company hated that for a long time because uh we felt that we should be capturing that money. We felt that, you know, uh somehow this was uh we were losing out, but the nice part about it is is it gave players the feeling that they could cash out if they wanted to of the purchases that they've made. How many do cash out? A very low percentage. Uh but uh that that didn't matter. There was the belief and it was correct that if they wanted to cash out,

they would get some of their money back. So the objects had a real value uh for the players. Um it wasn't it wasn't a completely entirely sunk cost. The other way is uh oh and so when we did Magic Online uh with uh what we did there and this I guess it launched almost 10 years ago uh pretty close. So um so for a product like that, when people were purchasing the digital objects, we were very worried that uh that they wouldn't find value in them. So there's a cash out system there where you can convert your digital cards into paper cards. You say, I get rid of my digital card, we'll mail you the paper card. There's some restrictions and constraints, but essentially we felt that was important when the game first started uh in order to um give people real value for the objects they were purchasing that they clearly understood. Uh I think that's less important in today's day and age. People are much more used to buying digital objects. Digital secondary markets are uh are common, uh but back then it was

it was a really a really critical thing. Uh tournaments uh let people play tournaments of your game if it's a competitive game. Um and uh and that gives people again real value for the objects that they've bought because they go they go to a tournament, they compete, they win. There's social social social currency in winning, there's real money in winning uh and um and so that again adds value to their to their to their purchases. Um fulfilling people's desire to collect, uh building community, these are all sorts of things that don't aren't quite tied into gameplay. All very very very important when you have a product that's going to be like a hobby, you need to think about these things. Again, returning value to their players for being loyal, for being in the system for a year, for making the purchases across uh across time. Uh there's a lot of other things uh that are sort of outside the scope of the talk that you can do, uh but the point is to figure out um how you can provide that value to your

players for being loyal customers and feed that. Um if every game is a sales pitch, then the scary thought is that your your customers are your salesmen. Uh and uh and that's that is pretty scary. Uh intermixing uh but but in order to facilitate them um uh being your salesmen, uh it's really nice when you can intermix large populations of customers because it gives people more opportunity to play and uh more opportunity to play against a a a very wide variety of people. Uh again, here's randomness absolutely helps. I mean uh you you look at tournament games played for money and they're sort of in the paper realm and you have a 50-year tradition of or more of uh of chess being played for money and it's nowhere near as successful as poker and there's a big reason for that. The the randomness and the allowing a much larger group of people to play,

much easier to find a poker game than a poker uh than a chess game, uh much more positive feedback for doing okay um is uh critical in in that sort of uh poker's tournament success. Um cuz it allows you to um to have fun when you're playing against people that might have a better deck. And uh and here's another important thing. If you're playing against someone in any of your any of your game interactions, uh you really do want again because you want your customers to find value in purchasing more objects, you want them to run into somebody who has bought more objects, who is beating them and they think they're beating them because they bought those extra objects. Uh if there's no randomness or a low amount of randomness in the system, they will not have fun throughout that process. If there's a enough randomness in the system, they will have fun because maybe they're losing maybe they're winning 45% of the time and uh and that's actually really uh an important sort of number. I mean you you really would would love that to happen

and you would love people to think that the extra purchase they make is going to be the one that puts them over the winning percentage. If they're winning 0% of the time, they're going to have an awful awful awful time and they're going to think they need to make too many purchases that it's too far beyond them. Um so this whole thing actually uh implies this is something we struggled a lot with uh in the tournament formats for Magic. Um and actually a little bit even Pokémon, uh it implies limiting the potential number of formats you have to play. Uh you can play StarCraft in a million different ways on a million different maps and if you have a billion customers, you're fine because all of those things will fill up. You'll have plenty of people in each of those, but except for the very top end games, that's not a luxury that most games have. You need to get in control of the number of formats that people play your game. And uh it's a it's a it's a really hard lesson because every one of those games offers something fun uh and offers

a little bit different way to play and you want there to be different ways to play, but you have to balance the consideration of making sure there's a robust audience in each one of those uh or else you start again to lose the value of players as customer uh players as salesmen and uh and you get you lose sort of the value in any sort of social currency built up in uh in being accomplished in any one of those formats. Uh this is a really important lesson, especially for video games. Uh it's sort of a side digression. It's not so much about the intermixing of business and game design, but it's uh but it's something that I think a lot a lot of people don't realize. Uh and that's reconfiguration in these distributed object games is is half the game. Well, that means that losses are critical uh for two reasons. One to drive sales because if I win every game I play, I'm not incentivized to buy anything extra, but even worse from the customer's point of view, they've put in all this effort into learning this very very complex system where a large portion of the fun

is the reconfiguration of their set of objects. So if they're not losing, they really actually don't get to access half of the game that that you've uh that you've designed for them. Um in a multiple strictly multiplayer game, uh losses are easy to come by. You look at your average population, everyone's playing, half of them are losing, you know, uh half of the games are are losses for people. Um so you might think it's trivial there in that case, but it's actually a little bit trickier than that. Um you you want it to not just be for instance, you could achieve that if your bottom half of the players lost every game and your top half of the players won every game. Again, as you can tell, a disaster situation. Very few people are actually accessing that reconfiguration part of gameplay. So, what you want is you want the right amount of win-loss ratio for people

spread across your whole population. Um and uh it's something that you can There's a lot of statistical ways that you can measure this. Um for instance, like in in Magic, we used a a variant of the Elo system that chess uses. It's essentially tells you what your win percentage chances against someone else. Uh like TrueSkill ratings, that sort of thing. Uh you really want to look at it across your population and and and figure out um if you're getting anything too aberrant in that. Uh and uh once again, skill can be high. It can be extremely high, as high as it is in chess. But that doesn't mean that accuracy in skill testing does not have to follow from that. You can have enough noise in the system through the randomness that you you have as much skill as you want in the game. You and then you dial up that noise so that uh so that people can have fun playing people of higher skill. And you're getting a much uh sort of uh smoother feedback on your actual true skill position than you would if uh you were

simply taking a measurement. Let's play the fun game of who's the tallest. And we line up and okay, you win and you lose. Um Uh and and the perhaps the more important implication for most computer games is that uh the single-player player version of your games uh actually has to entail a lot of losses. Uh and this is hard. I mean, we we submitted games over the years to major publishers and would get consistent feedback that uh the single-player game was too hard. The campaign was too hard. You were losing too much. And uh we tried to make this exact argument that I'm making now of Well, you've got all these digital objects. If they're if they're winning all the time, like the player should always win. Well, if they should always win, then there's no reason whatsoever to have a distributed object game. There you're just They're not accessing half the gameplay. Uh the fun in these games is in losing, reconfiguring, going back, playing, and winning. And uh and I don't know what loss percentage you're supposed to have,

but it's pretty high. I mean, even if it's you're losing 70% of the time, it's fine. It's a single-player campaign. The win Sure, they should be able to win at the end of the campaign. But along that way, you need to be They need to be generating losses at a much higher rate than you would normally sort of uh most people from a game design sense would consider, especially a long time ago. Um And again, if reconfiguration is half the game, uh repeat play is critical. You want to be able to play against that same person with that same deck after you make a small tweak. You want to be able to get the tactical part of the game under control so that you can access the strategic part of the game. Uh keeping playtimes low is uh I I cannot stress how important it for each each individual interaction. Can't stress how important that is for these games to achieve their full value. Uh if you're playing a 2-hour game and then

you go back and you reconfigure, I mean, to play another 2-hour game. There's a certain audience out there, I'm sure, for games like that. But in general, uh for most people, most uh players, that's just too much. That's just not a high enough percentage of the time where the thing that makes your distributed object game unique, where it sets it apart, they're just not accessing that part of the game. And uh and and repeat play is one of the biggest drivers um for players uh getting positive feedback. Because there's somebody that's really annoying that's beating them, beating them, beating them. They go back, pull another object out of their collection, change their thing, and beat that person. That's I mean, there's no better feeling. That now you have this recurring villain and uh and or best friend or whatever, and now you're beating him. And um And uh we we really struggled with this at Wizards. We we we we kept trying to um in R&D move move playtime down. Um And if I were making a new game today,

it would would have an even lower playtime than Magic. Uh and now uh design and business in these sorts of games uh are intermixed in a in a in an interesting way. Uh some examples. Um When you change the rarity of an object, you're not just changing it in a game design sense to the frequency that it might appear for a player. You're actually changing the price of the product. Uh you uh You you At Wizards, we used to uh look at, you know, uh well, if you're a competitive player, roughly what percentage of the objects will you need to buy? How much will you need to spend to be competitive? Um and if some of those objects change in rarity because you decided, "Oh, instead of uh one in 10 for this card to appear or object to appear, it's going to appear one in 12 times." Uh now all of a sudden, you've actually made sort of what would normally be considered a major marketing decision. It's not like you're I'm the head of

marketing for a project and I decide to call down to R&D and talk to some junior-level designer and say, "Should we put this out at $29.99 or $49.99?" Uh and uh they thought, "I don't know. $49.99." Great, we'll go with that. And that's actually the decision that's being made constantly in these types of games. Uh something no less uh no less powerful than that. Even when you change the power of an object. I mean, a rarity sometimes, maybe that's a little more major decision. But when you change the power of an object, "Oh, I've got a dragon of uh 42 power and I don't know, man. He's not quite good enough. I'll make him a 63 power." And just tweak it out. Again, it's this sort of seemingly very very very minor decision, which ripples through the whole system. Now all of a sudden, your average price for your game for a competitive player could be uh I don't know, $30 a month instead of $20 a month. And you've and and you've sort of completely unthinkingly made that decision uh about how many sort of top-level cards are needed and and and

what's required to access play. I I cannot stress enough how important it is to understand um what your audience is. Uh at Wizards of the Coast, we would sort of break down our audience into various categories. These are competitive people. These are kind of casually competitive people. These people over here don't care. They like the art. They love dragon pictures. Uh and and you sort of look at all those people and when you make these decisions about small power rarity changes, you are again changing the price of the product for those people. Just as much as if you actually changed the dollar sign. Uh and another thing is the release schedule. Uh something as simple as changing the release schedule has uh major implications for um for not just the gameplay, but for um but for uh Well, I'll I'll give you a specific example. Um So, uh these these games, these

combinatorial games, are very difficult to coalesce around sort of proper strategy. And uh the very top-level players uh enjoy that uh and like a fast rate of that. But the just one step below them, if the rate's too high, if it's too fast, uh they they can't they they can never feel like they're um uh anything solid in their sort of efforts in the game. So, in uh '95 or so, this card came out called Necropotence. Uh Some of you might be players. Necropotence is like the most hated card in the history of Magic uh among sort of top players. Uh we had so many complaints because, as it turned out, uh Necropotence was a really freaking good card. And a whole set of decks, all Necropotence decks, a little bit different different in interesting ways, but still the the professionals kept bitching about it. Uh the that card dominated play. Dominated

play for a whole summer. And uh it's going to ruin the game. You've got to ban the card. You've got to get rid of it. It's terrible. And an interesting thing happened, which was sales shot up, even though there weren't any new releases during that time period. And uh and it's all It's largely anecdotal, but at the retail level feedback, it was like, "Oh, yeah, it's great. People know what deck to play, finally. The average the low-level customer, he knows what he's doing now." Because he knows that if he builds a Necropotence deck, he may not win, but he's not coming with the worst deck. He's whatever. He's halfway there. And uh and this was an enormous And the message boards took off. Strategy articles took off. Um and it was all because sort of the strategy had slowed down enough that a large enough percentage of the population was able to catch up, access this, know what they were supposed to buy. I know I go make these cards. I make this deck. I know I'm I know I'm no longer an idiot. I may not win, but I'm not a fool anymore. And um

and I'm not saying you you want everything to stagnate. Of course, you don't want to lose the top end of players. But you have to respect everybody. And changing a release schedule as far as how rapidly you're changing the strategy has very strong implications on again, the the value of the incremental purchases for players at different levels of skill. Um and don't cut your best salesmen, which are these sort of rabid top-level players. Don't cut your best salesmen off at the knees by limiting their time. Uh They should be bitching about not having enough new things to think about at the same time they're posting articles saying this is clearly clearly clearly the best strategy in the game. Uh that's right about the point where it's perfect and and then move on. Um, yeah, changing the price of a product changes your potential customers. Uh, you when you change the price of a product either by because you made the dragon have an

extra whatever three hit points. Uh, you um, or if you make an actual change to the price, you change your potential customers are. Uh, and this is another thing where I I think sort of more traditional marketing people might actually be more concerned about these sorts of thing where they'll say, "Well, uh, I'm making game for 10-year-olds, so I'm going to put it out uh, at a at a certain price point versus uh, a game for mature audiences." And uh, and it makes big difference. Uh, and so uh, it's something to consider. You need to actually really know who your customers are, who the customers are on the fringes, and how sensitive they will be to price changes in your products that you're doing all the time by changing sort of what mix of objects that they need to buy, what rate they're expected to buy the objects at, uh, etc. Um, one thing that we did, it's a lot less important for digital games, I think, uh, is we were very big on the

organized play. Um, tournaments, leagues, that sort of thing is very powerful tool. Uh, sort of the interesting part about this is that is it marketing or is it design? Uh, when we started it, it was sort of traditional event marketing. And uh, and I think that's sort of how how many people would view it depending upon the company today. Uh, and there it was an intermixture. It's design was the only way we had to patch the game. Um, you know, games aren't just their rules. They're not just the rules that come in the rule box. It's how people play every game. Nobody plays by the rules or there's variations of rules. So by being able to pick that, you're actually effectively changing the the rules that people are are playing by. Uh, it's a very powerful tool uh, and it it's very similar like the way we thought of it is um, to patching the game uh, by patching the format. Uh, we we actually were sort of patching the the game as per rules that that we put out. Um, and it's a it's a really interesting

question because there's a lot involved here in the actual um, I would say more sort of event marketing mold. Like uh, one thing that we ran up against uh, especially with Pokémon is keeping these playtimes short. You've got kids. You've got 6-year-old kids. You've got 7-year-old kids. So they're going to come in to the game and they can't play. They can't sit there for 2 hours and play. I mean, that's at the low end. Most of them are 10 years old or whatever, but even a 10-year-old, he can't sit and play an hour and a half in a row. It's just not going to happen. It's a terrible event. It's terrible event marketing. Doesn't doesn't work for anybody. Doesn't work for the parents, the kids. A lot of crying. There's a lot of crying at Pokémon tournaments. And uh, and so um, so then you you have to start thinking about again the design implications of that. Lower your playtimes so that they can play. Well, what does that mean? Does that mean that I can't play best out of two out of three? Maybe I can only play one game. Maybe I need to change my mix of cards. Maybe I need to start discarding defensive cards so that

the game moves faster. Then of course then that this whole big cycle then that affects how much skill there is in the game. If you're playing just one game at a time or best two out of three. So you play two out of three, it's more skill testing. You can develop more consistent winners. A lot more social currency is being able to be built up among the audience for winning at tournaments because there's these consistent winners. You feel like you might be able to get a prize. You feel like improving skill at the game or or buying more objects will actually help you. Or do you dial down toward the randomness end so that the tournament will run more smoothly? Well, you know, the answer is somewhere in between. You you you you need to take all these things into consideration based upon your audience. And uh, and and and that that's uh, probably happens just as much with uh, online play of these games. Uh, are people expected to play a 1-hour game at lunch? Or is it a game for kids? Is it 10 or 15 minutes? Well, your design in the very little things you do in those objects that come up are going

to affect that playtime. If your tournaments are important, then you really need to be concerned much more about playtime than you were if they're not important. Um, and another thing is that organized play might be sort of you think of the organism uh, the rules as unimportant as I actually think they are most games are the skeleton. If my game is let's flip coins and see who wins or cut for high card, you're not going to have a pro tour. It can't happen. There's no way for you to get consistent winners on that. So you do need uh, a certain underlying structure rule structure to uh, in order to build up certain types of tournament structures, especially the competitive ones. So, couple extra concerns, long-term concerns. Um, how do you keep objects interesting uh, in the long term uh, over the course of several years? And uh, and why should you care? Well, you should care because hopefully your players are playing. And I don't mean to say that every game that has distributed objects is going to last for uh, what uh, how long did I say

Magic was? 18 years? But uh, but it might last for a year across several releases. That's sort of maybe a little that's that's not normal for a game. Um, so uh, yeah, it definitely does matter. Um, we found that we couldn't just allow uh, cards to live forever in the environment because when a new player came in, they felt hopelessly outmatched because you have thousands of cards that I don't have. Um, so we would rotate objects out. That was part of what we did with the tournament formats. Uh, and an interesting thing in the long term, you need to in the medium term you need to rotate objects out so no one is expected to have to make purchases at too high a level. Uh, and in the really long term, you want to start rotating objects back in because just as I was saying about the cash out value, um, you if if the cards have no value after 2 years, again, I should use objects instead of cards. If the objects have no value after 2 years, uh, that's bad. But uh, if they can always keep rotating back in, there's always a chance of

rotating back in, um, that gives people hope and uh, and actually does in fact increase the the long-term value of those cards. Avoid complexity creep. This is the Yeah. This is possibly the worst Magic card ever made. Um, the text you see here, I hope you can't read it. Uh, uh, uh, because because you'd be some sort of mutant if you could read all all those words. Uh, uh, is about half the text of the card because there was another set of rulings that was about this long of the actual text of the card. I created this card, by the way. Um, and uh, yeah, what a disaster. And uh, and it uh, it's the easiest way to sort of oh, I've got to put out more objects and more objects and more objects. Well, um, yeah, let's just add another line of text. Let's have it be resistant to fire. Let's uh, uh, have it I don't know, increase your accuracy by 10%. Let's uh, oh, let's also have it have you draw a card and take a card away from it. You just keep

adding this list of things that these things do. And then if that's the way that you're trying to keep your design fresh over across 2 years, uh, you're going to end up with stuff like this. So you you really do need to get it in control of it. Uh, I would set hard limits right in the beginning on how much how complex your your stuff's going to be. Uh, and avoid power creep. Here's another really bad card for Magic. I didn't create this one. Not that dumb. Uh, and uh, if you have power creep, two things will happen. One, you won't be able to rotate objects back in uh, because they'll be awful. Uh, and that's a minor concern. The major one is that you won't be in control of your game. You won't be in control of if the power keeps going up, you will not be in control of your playtime because now maybe all of a sudden the game is ending on turn four instead of turn 10. You won't be in control necessarily of the amount of randomness in your game. For Magic, it's a very specific example. You're drawing cards. Something that ends on turn 10 uh, is going to be in a way inherently less random than some than a game that ends on turn four

just cuz you're accessing more cards. Your distribution is more uh, even. Um, but in all these games, just a little bit extra, a little bit extra, a little bit extra. I keep making my gun a little bit more powerful, a little bit more powerful, a little bit more powerful so that people want to buy the next version, buy the next version, buy the next version. And then pretty soon, you'll end up with oh, everything's a one-shot kill. Clipped his ankle, he's dead. And uh, and so that's a that's a real problem. Uh, so you ideally, if you can handle it uh, over the long term, you would like the power of your objects in year 18 to be the same as the power of your objects in year one. And uh, that won't happen, but as long as you keep having that as a goal, you'll cycle up and down. And uh, in conclusion, um, there's probably uh, I would say there's two major lessons that I would like to get across. Um, one is uh, there's no substitute for studying games, all games. Uh, any successful game, I don't care whether

it's basketball or uh, poker, uh, the chess rating system, chess formats, any of these popular games, they have a valuable lesson for the game you're working on. Uh, it it just I I can't I cannot possibly understate uh, uh, uh, this this particular thing. Um, probably you have a problem that's been solved by someone. Maybe it was solved by someone making the golf tour back in 1925. I don't know. Maybe it was solved by someone last year who released a great RTS. Uh, the the answer is out there or at least something that's going to give you an idea for a problem that you have. The other one is that your customer does not care what your org chart is, uh at all. So, uh the fact that it's this person's job to um to uh set the price for a product, and it's this person's job to figure out what the power level is, uh you people need to get together and and really figure it out because uh because you're

both working on the product. All the customer cares is what that ends up to him in the long run. Uh and this is hard. This is one of the hardest lessons for us at Wizards because uh it's very painful when people are stepping on each other's toes, and uh you're going to you're going to have that both from the business people and from the marketing side, but uh but it's critical in these games to work it out because there can be such an amazing value proposition for the customer and the publisher if uh if you're able to sort of throw those divisions away and think of everything in terms of the end utility of the decision. That's it. Uh ques- ques- questions? Uh use use the microphone. Oh, everyone, please fill out the cards. And uh

Yeah. There's a couple of different CCGs I've been involved in, Magic being one of them, the other one being Legend of the Five Rings. One of the differences I noticed between the two is that in Legend of the Five Rings, the common most Sorry, could you speak more in the microphone? I can't quite hear you. Sorry. Yeah. Um in as two games I've been involved in, Legend of the Five Rings and Magic, um and one of the differences between the two is in uh Legend of the Five Rings, the commons were of more variable power than they are in Magic. Usually, the commons in Magic are almost kind of throwaway cards when you get to the higher level. It's the uncommons and the rares that people tend to play with. Do you have any thoughts about that, one being a better design? my main answer to that is to disagree. Uh Magic, uh there's plenty like if you look at it's varied over the years. Certainly, if you look at a if you look at a tournament decks in Magic, um I I think there's going to be, you know, uh as many common cards in those top-level decks as there are in uh in the other rarities, or at least it'll be

close. Certainly, when Magic first came out, uh lots of the most powerful cards were in were in common. Uh so, the intention at least for Magic was uh and I'm not saying that they did this, and I know at certain times people have had different viewpoints, but the intention was to push uh more complicated cards to rare, and to push cards that had more um specialized effects that may not occur as often to rare. Now, certainly, a lot of the powerful ones in the very beginning did end up there, but I I would say if you've noticed a consistent trend, it's unintentional. Thank you. Um Mag- Magic has different formats for the tournaments, and then Yu-Gi-Oh! has the uh banned restricted list. And could you comment a little on uh the differences there that you see in the plusses and minuses? Uh sorry, the question is uh Yu-Gi-Oh! has a banned restricted list? Correct, yeah. And and but only one format? Cur- well, it has banned and restricted, and then just limited and restricted. But the the cards from the very first set can still

be used, whereas Magic has the format set that rotate them out. Yeah, um well, uh we made a conscious effort I didn't work on Yu-Gi-Oh! We made a conscious effort to keep the um expenditure for the player down. Uh and so, what happened with Magic in the beginning is you you could use all the cards with a banned restricted list, and it became really really hard because our print runs were so limited in the beginning to get those older cards. They're much more expensive than, say, first-run Yu-Gi-Oh! cards. So, I don't know where they I don't know how much it cost to put together a Yu-Gi-Oh! deck, but we looked at the the price was just becoming more and more prohibitive. That's when we decided to start rotating out those older cards, and we had to do it wholesale because Yu-Gi-Oh! had a television show. They print whatever, 100 billion cards literally to start. Uh so, their prices never got to the level that they were at the original Magic cards. We had to cut those off. Now, maybe they're making a mistake cuz it's still too expensive to collect. I I don't know. I haven't really looked at the chart, but that's the first thing I would do would be sit down and look uh

you know, at that. It'd be great if the car- older cards were free enough that everyone could just use all their cards all time. Okay. Thank you. Uh over here. Um sometimes when you you made newer iterations of, like say, Pokémon or uh Magic, like there were sometimes when you made a new card that would just was better than a previous card. Yeah. How do you counter that later on? If you made it if if you don't know if you did it intentionally, if it was accident, but later on like sometimes like people will have cards that are just not really worth anything cuz they have the better version. So, how do you counter that later on? Well, uh that's gone back and forth over the years. Like we we had a rule at one time that no card was supposed to be strictly better than any other card in the whole set, and uh uh it was a good sort of a rule of thumb, but it it didn't really work out. So, eventually, we we just started making cards better or worse because of where we wanted the, you know, sort of the end goal to be. You will be forced into sort of obsolting certain cards.

Uh when you make there's tens of thousands of Magic cards, different Magic cards. So, uh you you'll be forced to do that. And the same thing if you're designing a computer game, same same exact sort of thing. Uh if if you have enough objects, some are just going to be strictly better than others, and and you can't help it. I mean, would it be better to try to plan ahead of time to make sure that it doesn't happen, or is it does it really matter, you know? Is it better to just like, you know, to make sure like try to put, you know, a lot of effort into trying to not do that, or just let it happen? Um I think you should plan for it to not happen, but then when it does, don't be too afraid and too resistant if you're actually going to end up in a better space. Yeah, certainly, the ideal is that that doesn't happen. I I don't think that's realistic. Uh so, I I do think sometimes you'll have to just say, "Yeah, you know, Lightning Bolt should cost whatever, R for two damage instead of R for three damage." That kind of thing. It just yeah, you'll be forced into that. Yes. Over here. Uh yes. Oh, if someone has an

idea for a trading card game, what advice would you give them to help flush it out? Um I wouldn't launch a new trading card game unless it had a major license attached to it today. Okay. Uh at at least I wouldn't put money into it. Uh if you want to launch it sort of uh on your own, and you know, 10 years ago I would have had different advice. Now, the way the market is, it's it's sort of difficult if you don't uh if you aren't able to bring a lot of initial customers in. But um if you want to let's say let's ask it a slight little question a slightly different way. If you want to make a really good trading card game, I would sit down. I would play test it. I would play test it with people that are not your friends. I would play test it with strangers, and I would do that for at least a year, and I would make changes on that with the thing with, you know, we really sort of as nice as you can get the cards to look. I mean, that's what we did back in grad school. Um you know, we had art. We photocopied it, cut it out, uh and these games are really really really hard to get right. Uh and there's a lot

of group think that goes on in your group of in in any group of friends. And so, um yeah, you know, put put it out in front of people that aren't your friends for about a year, and keep making changes, and you'll end up with a a much better product at the end than you would. Great. Thank you. Sure. Yep. You mentioned that in single-player campaigns, you needed to introduce very frequent failures in order to motivate reconfiguration of the deck. Yep. I'm wondering if you might think it would be possible to do uh something to motivate players to reconfigure their decks proactively, so that they still get that rich reconfiguring experience without the uh the feeling of failing all the time. Yeah, you can do a little bit of that uh because you uh you you know, you can say, "Now you're going to fight the fire guy." So, they sort of go ahead and think they should put all the anti-fire things in. Um and and that's okay. It's just really difficult to um

There's There's two things about that. Most of these systems, if they're going to have hundreds of objects, are are sort of too complex for people to be able to successfully do that if it's a hard job. Mhm. And I'm not saying that there aren't people that could do it, but if you really want to provide a lot of value from the complexity of the system, that that's almost impossible. Uh and then secondarily, when you get to the end of that campaign, you I guess, I don't know. I would question how smart I was if I didn't lose any of the things uh any if I didn't lose any battles, it's very difficult for me to get feedback on how well or poorly I did. Mhm. Um yeah, I I think it's more the na- the nature of the complexity of the system makes it uh yeah, I I definitely understand them. And in fact, the game that we submitted that was rejected had a lot of that in it. Um but but but I think we you still needed to generate the losses in order to um

yeah, in in order to uh really do the hard thinking about your deck. Thank you. Yeah. Um one of the advantages of going digital is that you can update uh cards at any time. So, you don't end up with a Necropotence or a Channel Fireball situation. But at the same time, you saw that sales surged as a result of players seeing which cards were needed. Right. So, how do you how do you So, um updating cards is a very hard thing, even online. uh just because it's possible to update does not mean that you want to update them. Uh because someone's purchased this card. They feel really good about this purchase because that was very powerful object that they bought. Um and then you decide later, oh Uh sorry, this is going to be the last question and then we'll go. So then uh so uh uh then you decide uh later that that was that thing was too powerful. Well, maybe I went out and bought that thing specifically because of its power. Well,

now you nerf it. Well, how does that make me feel? How does that make me feel the next time I think there's a really good object out there that I want to buy? Am I going to spend money on it? Uh how about if I wait and see if you nerf it or not? And that that's a decision you don't want to make. It's much better to constantly move forward if you can. Sometimes there'll be something so broken that it's just destroying the game and you have to deal with it. But uh aside from that, if you at all if there's any way you possibly can, don't take away what people have have bought. It's a it's a terrible thing. Okay. Thank you. Sure. All right. Sorry, no more time.