Interview: Richard Garfield, Designer of Magic: The Gathering and King of Tokyo

Editor's note: This interview was first published on Diagonal Move in March 2021. —WEM

Richard Garfield, designer of Magic: The Gathering, King of Tokyo, KeyForge, and many more joins Neil Bunker of Diagonal Move to discuss his remarkable career in game design.

DM: Thank you for joining us today, Richard. What inspired you to become a games designer?

RG: I find games fascinating and full of possibilities — like an exciting and largely unexplored land. It helped me understand the world and other people in a way that nothing else did. When I first got into games, I was amazed how little there was known about them relative to, say, books or movies or music.

The key moment for me was learning Dungeons & Dragons. That game broke all the rules for game design that I knew and thrust both the game master and players into the role of game designer to some extent. I figured if something so incredible existed that I had never heard of before, surely games were filled with many treasures to be discovered or created.

DM: You are probably most well-known for Magic: The Gathering, which has been a phenomenal success. Can you tell us the story of how Magic came to be, and at what point did you realize just how popular Magic was?

RG: Magic came about because I couldn't find a publisher for RoboRally. When I showed it to Peter Adkison, the head of Wizards of the Coast, he said he would publish it but needed something cheaper and fast-playing first to get started.

Soon after that, I had one of my few "flashes of inspiration" moments — most of my design is slow, intuitive, and experimental. I realized not all players had to have the same deck. I was swept up in all the possibilities for game play that had — and wary of the many problems it posed. It is sobering to think back to that time and remember — that amid the excitement — I told Peter that it might not be possible to make such a game. After all, Scrabble where you choose your pool of letters or Poker where you choose your deck are not necessarily good games, let alone better games. They are likely at best interesting puzzles. It was a matter of several weeks before I had a prototype that looked like today's Magic; it was built upon the framework of one of my many designs that I had enjoyed playing with, but didn't think was finished yet.

Looking back, it is easy to see that for years I had been fascinated by games where many elements of the game allowed the player to "break the rules". This interest first got kindled with Cosmic Encounter — and the spirit carried through many of my designs and was fully a part of Magic. My ideal was a game that was simple, but endless complexity was introduced through different cards. Anyone who sees the early magic rules knows I fell short of the "simple" goal, though probably not as far as it looks. 99% of Magic could be learned easily — and players could learn that fast and play a long time based on it. The remaining 1% was a nasty mess though.

There was no particular point that I realized how popular Magic was. I was perpetually surprised during the first few years, and honestly its impact on game design still surprises me from time to time. I knew Magic was a special game — the playtesters' passion was a testament to that — but I also knew many of my favorite games were not "smash hits", so I didn't think that meant Magic was destined for big things.

DM: Magic was followed by other well-known card game systems: Netrunner and KeyForge to name just two. How did you approach those designs? Was there pressure to repackage Magic, or were you free to experiment and take the designs into new directions?

RG: Usually I have been free to experiment with my designs, and that is what really keeps me interested. My first and second post-Magic trading card games (TCGs) were Vampire: The Eternal Struggle and Netrunner. With those, I was trying to figure out what mechanics worked well in this new kind of game. I learned many things about TCG design back then; for example, prior to V:TES I used standards that I had developed in board and card games, so I thought nothing about having a trading card game that ran for two hours with four or five people. After V:TES, I realized that so much of a TCG's value is in replay, possibly with a new deck or different tweaks to the old deck, that making the game short enough to allow for replay was a really good thing. Fans of V:TES either liked it as a TCG despite its length, or often liked it as a boardgame experience more than a TCG experience.

With Netrunner, I tried a lot of new things and I ended up with a game that I was really pleased with — but also learned some hard lessons. While Netrunner was in some objective sense simpler than Magic, the fact that everyone by that time knew Magic meant it was in fact much more difficult to learn than a game that was more like Magic would have been. I realized then that novelty in design comes with a cost, and as a designer it was my responsibility to make sure that novelty carried with it a payoff that was worth it for the player. For this reason BattleTech — my third TCG — was intentionally designed to be close enough to Magic to allow players to learn it easily, while being different in enough ways to make it interesting to them.

Fast forward twenty years and we get to the design of KeyForge — which was a game concept I wanted to explore for a long time, but couldn't because printing technology wasn't up to the challenge (or at least it would have been prohibitively expensive). With KeyForge I was trying to get the variety and uniqueness back into the game form which is diminished, if not destroyed, by players playing constructed decks with access to all the cards they need. For years I have been dissatisfied with that point in trading card games where one finds themselves removing cards they like from their deck because they just don't pull their weight. My preference is to play decks that are not honed to a razor's edge, but to play decks with more variety. In the TCG culture this is simply playing bad decks — and a player who does so is viewed as casual at best, and probably a bad player. But these decks, they can be very challenging to play and there is a great deal of skill to playing them well. I don't want to play casually — I want to play seriously with interesting decks. That is what KeyForge is about.

Board Game: KeyForge: Call of the Archons

DM: In addition to creating your own "worlds", you've also designed within existing IPs — Star Wars and BattleTech, for example. Can you describe how the design and publication process differs for these compared with your other games?

RG: Yes, I have done a number of licensed games, and the experience varies widely with how appropriate the license is for the game and how flexible the licensor is so that the best compromises between good play and best reflection of the world can be made. Working with a supportive licensor can be marvelous; it was that way with Star Wars, for example. Working with the other kind is soul killing.

I quite enjoy the exercise of figuring out the best way to frame a game within an existing world. There is a special pleasure to be found with a world elegantly reflected in an appropriate game. However, I will always lean toward making my own world since I know that I can do whatever I think is best for the game in that case.

DM: From a design point of view, how does iterating within a Living/Collectible System differ from designing expansions? Are there specific challenges that need to be overcome?

RG: Designing massively modular game expansions and expansions for a board game each carry their own challenges. In some ways, the massively modular games are easier to expand because that is what they are designed to do. Expanding a board game often involves challenges associated with adding complexity without a good enough value to the player, or the expansions undermine appeals the unexpanded game had. There are many times I have played board games and liked the base game — but then played expansions of it and for all the added variety the aggregate experience was worse, sometimes much worse.

A particular example from my own work is King of Tokyo. The success of the base game lead us to think about an expansion — but the challenge soon became clear. The easy expansion of "just adding cards" is not satisfying because cards are only a part of the game experience; some players play an entire game without getting any cards. Just adding cards impacts only some players, and the more cards added the less they each mean to the overall game.

So then let's explore another common request: monsters get unique powers. On the surface, this is an easy and obvious thing to add — but it turns out to be quite difficult to add without making the game worse. To see this you must understand that the basic game is a dice game with three principle strategies: attack, get VP, or get cards. A more casual player might pick a strategy and run with it, but a player who plays well will be adapting their decisions to the circumstances and the dice rolls they get. Being a dice game, either approach can win — but the "serious" player will win more often, a characteristic I really like in games.

Now if powers are added in a straightforward way and a monster gets, say, an advantage in attacking, suddenly the "pick a strategy and run with it" approach becomes stronger, and the player doesn't even have agency in that strategy since it is defined by their monster. The simple solution will be satisfying for a certain audience of very casual players, but many players will have the feeling that the expansion isn't as fun — even if they can't always put their finger on why.

Expanding a massively modular game is far easier in this regard – there are usually many different mechanics to explore, and even when there are limited mechanics, there are essentially infinite environments of mechanical combinations.

The challenges facing expansion of these games, however, in their own way can be quite difficult. As an example, let's talk about game balance. The stakes are generally much higher in balance, and the massively modular nature of the games usually make that balance much harder to gauge. To see why the stakes are higher, you have to understand the promise these games make to the player is endless variety and personal customization. A card that is too good must be in every player's deck, which makes both those promises less maintained. A card that is too bad shouldn't be in any player's deck, which does the opposite — which isn't quite as bad but still undermines the game's promise. Some degree of that is okay, but the more the expansion strays, the worse the overall experience becomes. And the more cards there are in the environment, the harder it is to manage that without making the game changes very conservative.

There are many reasons this is often not as big a problem in board games. Some of that is cultural; boardgame players typically have an easier time getting their group to not play an expansion they don't like — or even just play part of the expansion, or they modify it to their taste. The massively modular games tend to have massively modular playgroups — which makes that much more difficult.

Another reason is that often the imbalance in a game impacts all players equally, so going back to King of Tokyo, a card can absolutely be too powerful, but the system is much more forgiving since all players have access to it. A card that costs 0 and makes you win the game would ruin the play experience since the players' strategy would almost certainly be simply to collect energy and use it to sweep the board until they draw the game-winning card. However, a card that costs, say, 5 and was twice as good as another card of that cost? It would make collecting energy more appealing certainly, but is unlikely to break the game in the same way. That is a really large range of "error" one can get away with, especially if as a designer you aim for about 25% difference in power being acceptable rather than 100%. But in a game where players choose their own cards? This would be a "must have" card and make the play experience noticeably worse since every player would feel they have to have it.

DM: You have also designed some hugely popular board games. Most well known is probably King of Tokyo. Can you tell us the story behind this game, and why you think it has been an enduring success?

RG: King of Tokyo came out of a thought exercise around Yahtzee. A friend of mine was doing some serious analysis of Yahtzee at the time, so I was reflecting on how strong a design it was in that fashion that I really like: excellent play gives you better chances, but casual play can win. I wondered that if I were to try to design a game with the same principles, but interactive: what might it look like?

Interactivity in games can be tricky; done carelessly, it can involve a lot of "take that" political decisions which I am not fond of. I don't mind directly affecting another player, but I don't want to be in the position of choosing which player to affect, at least not often. The usual way to solve this issue is to make the interaction indirect, which, of course, can make an excellent game — but often one that feels "passive aggressive" rather than directly interactive.

My solution here was to make a "king of the hill" structure to the game. Being on the hill was rewarded, but carried with it risk in that you were the target. This made players in some sense in control of how much damage they were subject to and had a feeling of "low politics, direct interaction" that I often like in games.

Later came the flavor of "the hill" being Tokyo and monsters. I often design my mechanics first with some fairly generic theme, then completely redesign once I settle on what the theme should be. Once the theme has been picked, if you fail this redesign, it will feel much less integrated into the game play. For the record, I do design in the other way as well — where I have the theme first and build a game to that theme.

My own guess as to the enduring popularity of the game is a combination of the direct, yet low politics interaction — which is really pretty rare in games — and the cartoony and playful theme which IELLO managed to create around the concept.

DM: Staying with your board game designs: King of Tokyo sees players fight giant monsters, in RoboRally they control robots while Bunny Kingdom is an area control game about rabbits. These are thematically and mechanically quite different games. What do you feel is the thread that connects them?

RG: Mechanically I am driven to explore different areas of design, so I am likely to move to something new once I have gotten what I want out of a particular space. Sometimes it doesn't appear that way; in particular I had a number of drafting games (Treasure Hunter, Carnival of Monsters) come out shortly after Bunny Kingdom, so it may have looked like that was "my thing" — but actually they were all part of the same exploration at about the same time, and they were made in part because I was having a lot of difficulty getting a publisher to see them as interesting game space. Before 7 Wonders came out, they weren't really enthusiastic about it, and after 7 Wonders there was quite a stretch of time where they seemed to think there was no point in doing another because of 7 Wonders. Then there was a shift, and suddenly it was a class of game that could stand on its own.

A mild thread of mechanical connection, which is really more of a design style, is that all of these games can be played casually with a chance of winning or with great thought for increased chance. I tend to prefer games the casual and serious player can play together.

Thematically there is a strong connection between these games and most of my games which aren't made for existing properties: a sense of humor and playfulness. I like that more than "dark serious" game flavors because I think serious players can get past it if the game mechanics are worth it and the players are more playful with it when learning the game — which allows them to take the swings in the game a little less seriously when learning it. There is kind of a toxic "rush to judgement" with some players these days, and I believe this helps mitigate that just a bit — and if they stick with the games a bit longer because they don't take them seriously, they might actually get good enough to see how to play well and have fun with the mechanics.

DM: How has the huge success that you've enjoyed changed your approach to game design during the course of your career?

RG: I would guess that it is mostly the amount of time I can spend designing, playing, and studying games. The nature of my interest hasn't changed; I don't design more or less publishable games these days except insofar as my practice has probably made me better. Most of my designs are just for my own interest and that of my friends, and that has always been the case. Sometimes that leads to something I think other people will like — and then I look for a publisher. I have been in the fortunate position of never having to design to make ends meet, which might have lead me to working on games that didn't interest me or that I thought wasn't servicing the players enough to warrant.

Certainly, looking for a publisher is much easier than it was before Magic, and I do take pleasure in the fact that if I have a game that I think players will like, I can get a publisher to look at it and consider it seriously. That doesn't always lead to a product — or sometimes it takes a long time as it did with my series of drafting games — but that process of presentation and consideration always leads to improvement in the design, or at least the presentation.

DM: Is there a game you would like to revisit and do differently if given the chance and why?

RG: Hah. Every game I have made I want to redesign at least in minor ways. I am known to be reluctant to play any game of my own design once it is published — and I think the reason is that I get frustrated when I can't fix something.

For a major case of that perhaps I would go to SpyNet — which weirdly I would actually change very little about except for the messaging. I am disappointed that it barely got noticed after publication, yet find it one of my favorite two-player games. I think the decision to promote it primarily as a team game made people not give the two-player game a fair shake. Also, I think the special cards in the game gave a sense of "wackiness" to the play and players didn't take it seriously because of that — despite the fact that once you know what is in the game, there is a lot of interesting play dealing with that. I had some luck with friends incorporating a small card reference, I believe because it showed the players that they were supposed to anticipate the possibilities rather than just be surprised by them — which, of course, is common in first plays of any game.

DM: Do you have any upcoming projects you would like to tell us about?

RG: Roguebook is a digital game I worked on with the folk who made Faeria. It is a deck-builder in the same family as Slay the Spire. One of the key things we aimed to do was make constructing a big deck — a "tower deck" as we call them — a viable strategy. Most deck-building games are as much, or more, about removing cards than adding them. This is an interesting and skill-testing characteristic, but it is not logically required of the genre for an interesting game.

Personally, I like adding cards more than subtracting them, and I am particularly pained by removing cards that are fun to play but aren't quite worth playing with. The resulting decks are more challenging to play because the decks are less reliable and generally more flexible. One way we went about this was by adding a bonus that is unlocked for getting your deck to particular levels, so adding cards gives a bunch of cool powers to your deck over time. Of course, you can still play a lean mean deck if that is what you want.

Half Truth is a trivia game that I made with Ken Jennings, and I am really pleased with how it turned out. My inspiration was Ken's book Braniac, and I resolved after reading it to try to make a trivia game that could be played by a broad audience that wouldn't feel like the trivia nerds would always win. When I first shared the design with Ken, in fact, he played two games and lost the second one. (Not to me — I wrote the questions!)

The way it works is each question is multiple choice, and half the answers are correct. All players secretly make 1, 2, or 3 guesses. If a player misses any guesses, then they don't profit from the question at all. The players get only a small advantage for getting the second and third answer. Each question is a little minefield, and you can definitely get by with always just trying to get one answer correct. A lot of the fun comes from the really random and silly questions that are sprinkled throughout. One of my favorites was written by Koni, my wife:

Quote:

Is a poisonous mushroom:

• Fool’s Webcap
• Blinding Angel
• Destroying Angel
• Deathspore Thallid
• Deadly Galerina
• Night’s Whisper

Here I guessed Destroying Angel with some confidence, but I wasn't sure about the others. When the answers were revealed, it turned out all the bad answers were Magic cards! I saw the Thallid but missed the other two — but this opens up a really interesting characteristic of the game: You can have questions that you have no idea about but can still get a good guess in — occasionally even getting all three — if you can recognize the fakes.

DM: Finally, do you have any advice for aspiring game designers?

RG: Play lots of games, even games you don't care for. Learn what players like about them. Getting those qualities into other games where possible, ones you do like, will make your games better. Also, you will get more pleasure from games in general. Often a game I didn't like, once I really took the time to understand it, I not only understood the appeal but I acquired the taste.

Get a playtest network that has both casual and serious players. Listen to both. One common development standard that I regard as a mistake is just listening to the best players and looking to them for data. This is natural as a single group plays the game through many iterations over months or years. The problem is that game balance that is ideal for beginners and casual players is not the same as that for experts. One that seriously considers the former will often be much more exciting the first few times it is played and that is critical these days since there are so many other games to play. Development that relies too much on the latter can look very same-ish to the beginner — as if it doesn't really matter which strategy is chosen and the expert will always win by their 2% advantage.

If you use Kickstarter or some other method of self-publishing, get some playtesters outside your bubble, playtesters in particular that you don't teach the game to. One very important thing a publisher provides is an experienced sanity check on the game play and rules. I have many games that profited from their insight that I would have published on my own without hesitation. Sometimes that insight leads to improvement in mechanics; at other times it leads to improvement in the way the game is presented — both are important.