1/31/14 Episode 93: Top Down Design

All podcast content by Mark Rosewater

Okay, I’m pulling out of my driveway! We all know what that means! It’s time for another Drive to Work.

Okay. So today’s topic is something that I’ve talked a lot about in my column. What we call top-down design. Okay, so let me start by answering the question I get a lot, which is “Why is top-down design called top-down design?”

Oh, real quickly, let me define what it means. So top-down design is when we start with the flavor, and it’s the jumping-off point for the set. So from a flavor bent, Theros and Innistrad are examples of recent top-down designs. Like, “Oh, we’re going to do a Gothic horror world.” “Oh, we’re going to do a world inspired by Greek mythology.”

A bottom-up design is when you start from a mechanical place. So for example, Zendikar started as the land set. Ravnica started as a set where we were going to represent the two colors, make people play two-color gold.

Now, no matter which direction we go, we intermingle flavor and mechanics. So if our job is done well, Zendikar feels very flavorful. Ravnica feels very flavorful. But what I’m talking about is, technically when you’re building the design, which direction did you come from?

So anyway, people always ask me, “Where in the world does the term top-down design come from?” Now, at first—sorry. Sorry… traffic. Car coming right at me. Okay. So where does top-down come from? Now, I kind of assumed that it was—like, you know how you use an expression long enough that you’re just like “Oh, this is just an expression people use.” And then I realized “No, actually, I think I came up with it.”

So I did a little research. Here’s where I believe it came from. So if you look at a Magic card, the top of the card is the title, then there’s the art box, then there is the card type line that has the subtype, has creature types, then you get to the rules text, then you get to the power/toughness, then the legal text and collector number.

Well, the top half of the card has the vast majority of the flavor. Because the title and the name and the creature type are all on the top half of the card. So top-down came from “Oh, well you start from the top of the card and go downward.” So a top-down design is, “Well, I know what the name of the card is, and kind of what it looks like, and oh, I’m trying to match the flavor of the card, making the rules text.”

Where bottom-up is, “I have the rules text, now I figure out what the flavor’s going to be.” Now, we have an awesome creative time. Oftentimes we make a mechanic and they come up with a perfect match. And so it feels very connected. But I’m just talking about where we start. In the end we want everything to feel meshed and interconnected. But anyway, top-down design is when you start from a place where flavor is the jumping-off point.

And I’ve talked about this a lot in this podcast especially, in that one of the keys to creative endeavors is you want to make your brain attack things from a different angle. I’ve talked about this a lot. That if you go at something the same direction, you’ll hit the same neurons and you tend to get the same answers. Because the brain will follow a familiar path. But if you come at it from a slightly different vantage point, then you’re thinking of it differently. And you find different answers.

So for example, when I do top-down design, I’ll just use Theros as a recent example, everything I was doing was saying, “Oh, does this match Greek mythology?” And so my lens that I was looking at things through was a lens I had never used before. I had never designed Greek mythology before. And so I was making all sorts of interesting decisions, because I was using a new thing to judge by. It’s one of the reasons top-down is nice.

And bottom-up also does that, where when I was doing Ravnica, I’d never done a gold set before that focused on two-color play. And so when I did that for the first time ever in Ravnica, you know, it guided how I wanted to put things together.

Okay. So today’s podcast is going to be about how we do top-down design. Sort of the lessons of top-down design. Okay. So thinking about it, I realized that communications theory does a very good job of explaining the key lessons. So for those that didn’t listen to my podcast on communications theory, or podcasts, plural, in communications theory there’s three things you have to strive for. One is comfort, one is surprise, and one is completion. So I’m going to walk through top-down design from those three lenses, because I believe it gives a good example of what you need to do to do good top down design.

Okay, so number one. Comfort. Okay, so let’s say I’m doing a top-down design. I’m going to use Innistrad and Theros, as those are the two recent sets I’ve done that were top-down design. Okay. So what happens is, the first thing you want to do—when I have both my first meeting of Innistrad, and my first design meeting of Theros, I said the following to my team: “Okay guys, we’re going to brainstorm. What is everything you would expect that this, for Gothic horror or for Greek mythology, what would you expect?”

So for example, Innistrad, we started writing on the board, like, well we expect vampires, and werewolves, and zombies, and ghosts, and we’d expect nighttime and candles and bump in the night and we’re just writing everything down that you could think of. Victims, and everything that you could think of that Gothic horror would be to you.

And the same for Greek mythology. It’s like centaurs and Medusa and a pegasus and minotaurs and gods. We wrote everything down that we expected. And the reason this is important is, part of doing top-down design, the reason to do top-down design is that it plays into the concept we call “resonance.”

And what resonance says is, okay. If you are trying to make somebody—I, as any kind of creative person but especially the game designer, I’m trying to connect with my audience. I want them to be invested in the game so that they are excited. Right? Remember, game design is about creating entertainment and experiences. Right? That you want people to experience something, and you want them to feel something. Experience an emotion.

And one of the tricks to doing that is to take things that your game player already understands. Because they have lived a whole lifetime. They have all these experiences. If you tap into pre-known experiences—and that could be real life experiences, it could be pop culture experiences, but something that they know and understand, what happens is that you are piggybacking if you will on emotional things they already have. Emotional feelings they already have.

And so, for example, if we’re doing Gothic horror, and you—there’s expectations that come from Gothic horror, because they’ve watched movies and they’ve seen TV shows and they’ve read books, and they have expectations. So part of doing top-down design is you want to meet that expectation.

That there’s a level of comfort that for example—Champions of Kamigawa—so historically speaking, there are three blocks that have done top-down design. Champions of Kamigawa, Innistrad, and Theros. Also, we did one set, Arabian Nights, that was top-down design.

Now the difference between Arabian Nights and the other three blocks is Arabian Nights was top-down, but Richard was trying to as best he could match Arabian Nights. He wasn’t trying to create a world inspired by Arabian Nights, he was trying to actually capture Arabian Nights, where what we do now is, Kamigawa and Theros and Innistrad, we were trying to create our own Magic world inspired by those top-down flavors. We weren’t trying to do exactly—for example, in Theros, there’s no Zeus. There is Heliod, which has elements of Zeus, but it’s not Zeus. It’s different. We were doing our versions of things.

I bring in Kamigawa because Kamigawa is a good example of mistakes that we’ve made. I think Innistrad and Theros are top-down done well, and that Kamigawa is top-down with some major mistakes made. And we learned a lot, one of the reasons I think Innistrad and Theros are as good as they are is we learned some important lessons from Kamigawa, and this is one of them.

So Kamigawa went in to look at Japanese mythology as inspiration. And so one of the things that they got into was Shinto—I do not, by the way, know tons about Japanese mythology, so I’m going to talk the best I can, but if I’m a little off, it’s because I do not—this was not my set. I did not lead that set. And so I’m not quite as—I spent a lot of time and energy on both Greek mythology and Gothic horror.

But so we did some stuff with Shinto, we did a lot with the kami. There’s a belief in Japanese, I think in Shinto, that like every object has a spirit associated with it, I believe. [NLH—to my understanding, roughly yes.] Anyway, what we did though was we captured something that while being somewhat true to Japanese mythology, was not very known by a lot of the Western audience. So we created a lot of stuff that wasn’t as resonant as it could be. I mean, I think there are a lot of resonant tropes for Japanese mythology.

But instead of sort of hitting the more obvious ones, we would try to stay true to it, but the problem was we weren’t as resonant as we needed to be. In fact, Kamigawa also taught us that it’s very, very important that when people first experience your top-down, that the thing they expect is there at common. You want them to run into the stuff they expect first.

 So if you’re doing Gothic horror, well, you want the monsters to show up pretty quickly. You don’t want like all these unknown things sitting at common. You want your vampires and werewolves and zombies and ghosts and stuff sitting right there. And same with Greek mythology. You want your minotaurs and your centaurs and your cyclops, and the things that you would expect. You want people to see them quickly, okay?

Now, the idea is, that doesn’t mean you can’t ever do more obscure things. Hundred-Handed One should be an excellent example. Which is, we have a Hundred-Handed One, if you know Greek mythology it’s a big part of Greek mythology. We didn’t want to not do it. We wanted to include it. And we wanted to have a reward for people that really knew the source material.

But, unlike Champions of Kamigawa, where we put some of that stuff at low rarities that people just didn’t know, Hundred-Handed One is rare. It’s not going to be the first card most likely that you’re going to open up and see. And the idea is, one of the lessons of top-down is, make sure that your as-fan, your low rarities are things that are comfortable that represent the thing you want.

Because one of the things that’s important is that when you’re trying—the point of top-down is to connect with your audience on an emotional level. Right? You want them to go “Oh yeah.” You know. “Oh, you’re doing so-and-so? Oh, is there this?” “Yes there is.” “Oh, is there that?” “Yes there…” you know, you want the audience to be hoping for things, and deliver on most of that.

For example, if I went out on the street and polled Magic players, and said, “Okay, when you think of Thing X, what do you think of?” “When you think of Greek mythology, what do you think of?” I want us to hit most of those top ten answers.

 

Now, I’m not saying all those answers have to be the major thrust of what’s going on, but if you ask people and they really, really think there’s going to be a gorgon, well, you kind of want to make sure we have a gorgon. Like, Greek mythology—Medusa’s pretty central to Greek mythology. I’m not saying you need a lot of them, and obviously we didn’t have a common one, but we did have an uncommon and a rare that were high profile, like “Hey, you want to see a gorgon, we got a gorgon for you.”

And a lot of what we also did in top-down design is you look at what Magic has done before, to see if the things that—like, for Greek mythology, one of the things that Ethan did, Ethan Fleischer, I had him do some research ahead of time for Theros, and he wrote down a booklet. And the booklet mostly broke into two categories. First category is, “What are things we would expect to do that Magic has done?” and “What are things we would expect to do that Magic hasn’t done?”

And we wanted to make sure that things Magic had done, that we had enough of those there, because there was a base already built up. One of the examples was, I knew I wanted a little bit of tribal. Not tons, because the way the set played out, as I explained during my lengthy, lengthy Therospodcast, that we found that in—when people knew Greek mythology, they didn’t think of races together in the same sense like they would in Innistrad, where monsters you do kind of click together. But, I knew I wanted some tribal, so I went and I said, “What is the tribal players would most expect and want?” And I decided it was minotaurs.

Minotaurs have gotten a little bit of tribal stuff in the past. Homelands, obviously, did some tribal. I know Didgeridoo was very popular. I knew that there was this desire for minotaurs and minotaur tribal. Okay, minotaurs are super resonant, even though ironically the actual Greek mythology has one minotaur. In our world, there’s many minotaurs, not one minotaur. And it’s become a race that Magic has used quite a bit, so that seemed like a very good place.

Okay. So first off, number one, comfort. Figure out what your audience is going to expect, deliver on expectations, make sure that expectations are at low enough rarity that that is what your audience is seeing.

Okay. Time to move into the second part. Surprise. Okay. So now that you’re doing something, it’s important that you do everything you can to capture the stuff that people expect. But then you want to make sure that you add some unknown element to it. So for example, I will use Theros because there’s a clean example here. So we wrote on the board, “What do you expect?” And as we wrote things on the board, it became crystal clear that there’s an expectation of gods. Gods are about as central to Greek mythology as you get.

But, while we wanted to deliver Greek gods, we wanted to surprise the people a little bit. And the reason is, when we do a top-down set, our goal is not like Richard was doing in Arabian Nights. Our goal is not to just capture 100% in which it’s just that thing. We want to do our spin on it, we want to make it a Magic thing.

And so with the gods, as I explained in my Theros podcast, we wanted to take the essence of Greek mythology, which was the gods, and the essence of Magic, which is the color wheel, and put them together. So what that meant was, we ended up making fifteen gods, five major that are monocolor, ten minor that are two-color. And each one of them had to embody either the color or the color combination that they were god of. And then what we did is, we chopped up all the existing Greek gods and took elements of them to make our new gods.

Now, why did we do this? Because there’s a level of comfort. You expect gods. We have gods. You expect attributes of gods, we have attributes of the gods. But we’ve mixed and matched them in a way to make something new and different. And it’s important for top-down, I feel, that you have some newness to it.

Now, even Richard, doing Arabian Nights, which was he was trying to capture, did make up some stuff. While Arabian Nights might have had djinns and efreets, Richard made up specific djinns and specific efreets.  He definitely gave some spin of his own to it. And in Magic, now, you know, Theros and Innistrad, even more so. We definitely took our take on things. Yeah, they’re gods, they’re our gods. They’re color wheel gods. They’re Magic gods.

And the reason that is important is that essentially what you want with a top-down set is you want to create comfort, meaning you want to draw the audience in. Oh, and here’s something important. I’ll use Innistrad to make this out. Which is, one of the fights you’ll have sometimes is trying to be realistic to your source vs. fighting expectations. For example, in Innistrad, we were doing Gothic horror. In Gothic horror, the zombies in Gothic horror, there’s not a lot of them, and probably the most famous is Frankenstein’s monster from Frankenstein.

Now, if you know the actual story of Frankenstein, Frankenstein is the story—he actually is an intellectual. The monster. That he is intelligent. That he has conversations. The idea of Frankenstein being like, “Urrrgh, Frankenstein,” that’s modern. That is not an old-school version of Frankenstein. That’s not what the book does.

But that is what people expect. And likewise, when you say “zombies,” people expect what I call Dawn of the Dead zombies, which is zombies that kind of slumber and are did and aren’t really bright, like “Braaaains, braaaains!” That’s the kind of thing you expect.

So, when we say we’re doing zombies, well if people are going to expect Frankenstein’s monster—like, the Universal-style Frankenstein’s monster, not Mary Shelley, and they’re expecting Dawn of the Dead zombies. Well, guess what. That’s what we’re delivering. Did that actually match our source material? No, no it didn’t. But that was the expectation, and you cannot fight expectations. If your audience expects something and you’re trying to deliver to it, you have to match what the audience expects.

In Theros, the other example is the kraken. Kraken’s from the Scandinavian mythology, not from—I mean, now there were sea serpents in Greek mythology, so that’s not a crazy stretch. But we knew thanks to Clash of the Titans, both versions of the movies, that, you know “Release the kraken!” People expected a kraken. And they’d be upset. So we made sure to deliver that. That even though it’s there.

And my other example of this is, Richard Garfield made a game that’s called, “What Were You Thinking?” Its design name was called Hive Mind. And the premise of the game is that you get a topic, and then you’re trying to write down the answer that everybody else is trying to write down.

So one of the categories one day was insects. And it turns out that one of the top answers was “spider.” “Now, now,” you might say, “What? Spiders aren’t insects.” Well, you see, it didn’t matter. The goal of the game is to write down what other people wrote down, not to be correct. It wasn’t actually “write down five insects,” it’s “write down the five things you think people will think of when you tell them to write insects.” And not only that. What do you think they think other people will write?

So the funny thing is, everybody in that room might have known that spiders weren’t insects, but enough of them felt that other people might not know that, that they wrote down spider, and the correct answer was to write down spider. So when doing top-down, you do have to match expectations.

Now, that said, once you match expectations, once you give something, players then want to be shaken up a little bit. Players do want us to say, “Okay, we’ve given you Greek mythology, but here’s some stuff that’s a little Magic-oriented.” You know, “Here’s our versions of the gods. Here’s our versions of some of the stories.” We definitely captured a lot of things from Greek mythology, but we put our own twist on them.

And the fun of that is, once you have the comfort level, once you—“Oh, you’re doing Greek mythology, you’re doing top-down Gothic horror. Oh, yay, vampires, yay Greek gods.” Whatever. At some point, you’re like “What’s new and different? What are you giving me that I haven’t seen before?”

And that’s where we get to sort of put our own spin. And it’s fun for the audience, because Magic’s still a game of discovery. You still want people to want to look and see.

Now part of it is seeing things we already know, and if we did them, and part of it is finding new things. And you want a balance there. Top-down isn’t about being slavish to your source material. But, and this is important, you do have to follow the feel. Meaning everything you add has to have the right feel to it. You can’t just do Greek mythology and go, “Hey, look! It’s…” you know, something out of Norse mythology that has no connection to Greek mythology. Can’t just go, “Here’s Thor! With his mighty hammer!” And like, “What? That’s not Greek mythology.” So you have to sort of make the thing fit.

Okay. So now that gets us to the third part, which is completion. Which is, it’s not just about making individual pieces. One of the things that I’ve talked about a lot, and this is—I mean the completion aspect of design talks about this quite a bit, which is you can’t just think about your designs in isolation, meaning if you just make every card and don’t think about anything else in the set between each card, even if each card in a vacuum makes total sense.

For example, let’s say I say, “Okay, I’m going to do a Greek mythology set.” And I make a card—I forget about everything else I’m doing and just make a Greek mythology set, those individual cards in a vacuum might be awesome cards. But the thing is, you are working together. Your game does not live in isolation. Your cards do not live in isolation. And so you have to be very conscious of how things interconnect.

Especially in Magic, a game in which people will take cards and put them in their deck. Meaning—and this is important to remember. We, as trading card game makers, are making a game in which we’re giving you components. Now, remember, when I talk about trading card game, that is a really, really important distinction. That most other games, when you take the game and open your box, when you open your box of Monopoly, it’s an experience that’s not unique. Every person opening a Monopoly set, or a basic Monopoly set, is getting the same things. And if I go to my friend’s house, and he has the same Monopoly set I do, we’re going to play the same game.

That’s not true for Magic. Magic is different. When I go to my friend’s house, if I’m playing with my friend’s Magic cards, they might be different cards, or I’m playing against his cards, they’re different cards. And that is one of the things that makes Magic a very different game. But remember, people are going to experience your game by taking the components and putting them together. And so our job as trading card game designers is to make sure there’s cohesion. Now, in top-down design, what that means is I have to be very conscious of how things fit together.

For example, let’s take Innistrad because (???) about Innistrad. I was trying to tell a story of the monsters impeding on the humans. Really what it was, was the story of the humans in peril where things were bad. I mean, remember, in story you want to start your story at the farthest end away from where you want to end up. Well, the ending of our story was the freeing of Avacyn, and the good guys come and save the day. Well, then I needed to put my heroes in a pretty bad place to make that story interesting. If the humans were in good shape, well, Avacyn coming back, what would that mean? No, things have to be bad. In order for things have to be bad, I wanted a relationship where the monsters were impeding on the humans.

Now, the monsters weren’t working together. It’s not like the vampires and werewolves like drew up a plan. No, the werewolves were doing werewolf things, the vampire were doing vampire things. But in order to create this sense that they’re in trouble, I decided that I needed to isolate the humans. Well, how do I do that? Well, one of the ways to do that was to create structures that I left the humans out of.

So for example, I did a bunch of cycles in which the monsters all got stuff, but the humans didn’t get things. And the idea being, to give a sense that the humans (???) and separate from the monsters. And make them feel isolated. I tried to isolate them in my design.

Now, there is no way for me to do that—that is a design that only works in conjunction. Now, the good thing I had going for me is I was doing top-down. That top-down was humans were the victims. The reasons we had humans, and the reasons we did human tribal for the first time, was it was very important to me that the victims got represented. Why? Because in horror stories, the victims are a key part of the horror stories. You don’t tell horror stories without victims. And the victims are human.

Why are they human? Because the whole point of a horror story is to get your audience to identify with the hero so that the horrible things that happen to them, you go, “Oh, that could happen to me. I’m scared. That would be scary if it happened to me.”

Because a big part of any sort of creative endeavor is you want to give your audience sort of what we call a POV. A point of view. That you want them—if you’re trying to get emotional responses, usually what you do is you connect them to the hero. Or in the game, the center of the game, such that they’re experiencing things they need to experience.

So in Magic, you’re the planeswalker. You’re having a duel with the other planeswalker.              And so I’m trying to make sure that when we build our sets, I wanted the humans to feel like victims. It’s very hard to do that on one card. I mean, I could do “Human, Victim,” but I mean it’s a little heavy-handed.

What I wanted to do is demonstrate that the humans are in trouble. Now, part of that is making the humans smaller. A monster and a human, the monster’s going to win. Now luckily, white, its nature is humans teaming up, it has smaller creatures, so it works well there. I mean, it’s why humans were in white in the first place.

But the key is, you want to make sure that when you’re doing top-down, that not only are you creating the elements you need, but that you understand how they structure, because the audience is going to want a sense of structure to what’s going on, and it will help them give that sense of completion.

In fact, one of my pet peeves is, I did not do a good job of explaining to Erik that the curses were a cycle that left out the humans, and so the green curse got left off, and then, “Well…” like, it’s hard to recognize a pattern when the pattern’s not complete. Having cycles of four, you really need all four, and that was me dropping the ball in that one tiny area. But (???), it’s that notion of completion that’s important.

Okay. Now I’ve got off the freeway, which means I’m not too far from work, so let me set to do some recap here. Okay. So when you’re doing a top-down design, you start, and make sure that what you’re doing is resonant, that there’s comfort, that you are capturing the things people expect you to be capturing.

Then, we get to surprise, which says, “Okay, make sure that not only do you capture what people expect, but you put some twists on it that are your own.” That when people come to see something, yes, they want familiarity, but they also want some sense of identity. You are doing something that’s giving it your own spin.

For example, like in movies, (???), but let’s say you’re going to take a classic story. Hollywood seems very into doing fairy tales right now. Hansel and Gretel. Well, hey, you want Hansel and Gretel, there better be a house made of candy, you want a witch. But you want some spin on it. And I believe the spin in the last one was like, “They’re witch hunters.” Right? That they lived through this experience and now they’re hunting down… I didn’t even see the movie, but from the poster I got this much.

That you want to have some comfort, but then you want to have some spin on it. You want to have some take on it. There’s been a bunch of Snow Whites. Well, how is your Snow White different from other Snow Whites? I didn’t see these either, but Mirror, Mirror I think was more modern take on it, where Snow White had a little more active role. And Snow White and the Huntsman—both of them, I mean she becomes more of an active fighter. She’s more involved. She’s not as passive. And that’s a big part I think of how—I mean it just was a modern take on the story, but it had its own take on it. It wasn’t just a story you knew.

But—and once again, I didn’t see the movie, but—it’s called Mirror, Mirror, so I’m sure there’s a magic mirror. I’m sure there’s an evil queen. I’m sure there’s dwarves. There’s probably seven dwarves. You definitely want to deliver that there.

Okay. So you have the comfort, you have the surprise. Finally, completion says, “Okay, not only am I making the components, but what does my audience expect of the components?” That I can’t just make individual components. Once I’ve made certain components, there’s an expectation that other components are going to be met. Right? That I have to make sure that I am not just building pieces, but building a whole. And I have to figure out in my design, what my overall feel is. Right? That part of creating the resonance that I want is creating resonance in the micro and resonance in the macro.

And once again, one of my ongoing themes, I have a bunch of themes, is in the macro, in the micro. That says, “If you want something to happen, if you want an audience to feel something, it has to appear both in the big picture and in the small picture.”

In Magic that means the set has to show it. The cards have to show it. That I want to show it through my set design, I want to show it down to my individual cards. That whatever theme I’m doing, I want to keep hitting that theme.

In Theros , for example, I was trying to build up a sense of adventure. Well, I wanted cards that built. I wanted mechanics that built. I wanted a style of play that built. That my theme was hit on every level, from the micro all the way up to the macro.

And remember, when you are building something, that your big device is made of a lot of small devices, and so you want that to carry through.

Okay. So top-down—comfort, surprise, completion. How—by the way, if you told me when I was studying this communication in school, that this thing I learned—and by the way, this completely applies to like TV and film and all sorts of stuff, and like “Oh, this would be very good for game design.” I’m like “Wow, kind of cool.”

But anyway, I am driving in. I see the Wizards building, I can pull in my space. Or not my space. A space. I always say “my space,” people feel like I have my own parking space, which sadly I do not.

Okay. So I hope you guys enjoyed hearing about top-down design. It’s something that I think we are getting better at, it’s something that has gone over really well, both Innistrad and Theros went over like gangbusters. So we will be doing more top-down design. I know the future seven-year plan, so there’s more top-down design coming. Not every year, A. because I want to mix it up, and B. because there’s not an infinite number of things we can do top-down design. It’s a lot smaller than you think.

But anyway, thanks for joining me today. Ooh, it’s pretty much about an average 30-minute ride. Not a lot of traffic today. Which is good for me. I don't know if it’s good for you. But anyway, thanks for joining me. And while I always love talking about doing top-down design, I also like making Magic. Talk to you guys next week. Bye.