#741: Bill Rose
In this podcast, I talk to Bill Rose, the vice president of Tabletop Magic. Bill and I have worked together for 25 years.
Posted in Making Magic on May 25, 2020
With the release of Ikoria, I've gotten a lot of questions (a number of which I answered over the last two weeks—Part 1 and Part 2), but I saved the most common question I've received for today as it has an interesting, but somewhat lengthy, answer—what's going on with the complexity of Ikoria? I've spent numerous columns talking about things like New World Order and the importance of simplicity. How exactly did this set ever get made at its current complexity level? It's a good question, one which I'm dedicating today's column to answering.
I guess I'll start by giving the abbreviated answer. I'll then follow up with the lengthier answer for those who want a more in-depth explanation. Ikoria is an experiment. Magic is a game that's constantly reinventing itself. Who plays? Who do they play with? How do they play? Where do they play? When do they play? Why do they play? The answers to these questions keep shifting, causing the game to adapt in response. Part of our job in creating Magic is to be aware of all these different factors and adjust the game accordingly. A big part of that is a willingness to experiment and stretch the boundaries of what's considered acceptable.
Only by testing things can we get the data we need to guide us into the future. One of those questions is "what are the boundaries of complexity for a premier set?" (R&D has decided to stop using the term "Standard-legal set" as it implies a little too strongly that our new sets are just about Standard. Instead, we adapt with how our audience has chosen to play the game—you are all playing a lot of different formats, so we've started calling them premier sets.) We've spent many years wrestling with what the floor of complexity is for premier sets (aka the simplest the game can be while still being fun), but we wanted to get a handle on what the ceiling was (the most complex the game can be while still being understandable). For many reasons (which I'll get to below), Ikoria proved to be the perfect testing ground for this question.
When you boil everything down to its core, Magic design is about making a fun game for all its players. The biggest challenge is that the players want and need different things, and every premier set must deliver something for everyone. The focus on complexity as I mentioned above is about finding the right middle ground where the game is simple enough for new players to learn the gameplay and complex enough for ample exploration by the experienced players. Make the floor too high, and the game becomes too challenging to learn. Make the ceiling too low, and the game becomes too boring for the experienced players.
We've spent the last decade with a lot of focus on the complexity floor of the game. While these efforts go far beyond New World Order, I believe it's come to represent this quest to the public. So, when you all see something that pushes in the opposite direction, I understand that it feels as if we're ignoring something we've said is important. Before I move on to talking about the complexity ceiling, I want to talk a little about how the environment has changed in ways that impact how we look at the complexity floor. There have been two big changes that I want to focus on:
"On-ramping" is a term which talks about how you get a person to go from being someone who doesn't play your game to being someone who does play your game. Longtime readers know I consider "barrier to entry" (aka the amount of effort it takes to learn a new game) to be Magic's biggest weakness. It's not even so much that it's a hard game to learn as much as it's an intimidating game to learn. Wizards has spent years and years trying to better understand what makes for the best on-ramping.
Here's what we've learned. You need a teacher. You need someone dedicated to walking you through the steps of learning, holding your hand at each step. Giving you text to read, or pictures to follow, and expecting you, the player, to teach yourself has proven to be a miserable failure. Why? Because there are many pitfalls where a player goes down the wrong path and there's nothing to help them self-correct. If you read a sentence and assume something other than what we intended that sentence to mean (which people do all the time), you're doomed. A teacher will recognize when you've made a false assumption and correct you.
It turns out there are two types of teachers that are the most effective—friends and computers. Friends have a relationship with you. They know you and are able to pick the best path for you to learn. They can anticipate your reactions to help steer you down the correct path. The rapport you share usually means they can make the experience fun. In the current environment, the most common way to play casually (how most beginners start) is through multiplayer play, mostly Commander. Now, you don't have just one friend teaching you, but a bunch of friends. Multiplayer also compensates for skill imbalances, so your friends can keep you in the game as you're learning and gently point out mistakes as you make them.
Computers are good teachers because they don't judge. A lot of the fear of learning something new is failing in front of other people. A computer doesn't care if you make the same mistake six times in a row, and they won't tell anybody. It also has the ability to control the environment, so it can carefully let you only access what you're ready to learn. Finally, it has the ability to monitor everything, so it catches misunderstandings quickly. Computers are a core part of digital games, and digital Magic is on the rise (and that was before a global pandemic).
What this means is that the two most common ways to on-ramp are multiplayer play and digital play.
Another big leap in the way we make Magic is we've gotten better at monitoring why people do things. For example, we have an entire group focused on understanding player motivations. We also have a data research group focused on collecting as much data about our players as it can, and with a lot of the advances of the last decade, both in technology and a rise in digital play, we have more data than ever. Lastly, we have a group dedicated to analyzing that data to understand what exactly it all means. The offshoot of all this is we have a much better grasp on how and why people learn.
There are many takeaways, but here's the most important one to this topic. Players learn best when the act of learning fulfills an emotional need that they have. I often talk about how the most important thing at the end of the first play session is a desire to play again. What makes someone want to play again? It turns out that seeing something tangible, that is, that you can see how the game affected you, including how it made you feel, greatly increases your desire to play again. What this means to design is that we want to make sure that something about the gameplay resonates with the player. It's more important that a player enjoys a card than completely understands it. (Man, it took us a long time to learn this.)
I bring these two changes up because it's made us re-evaluate where exactly the floor has to be, and more important to today's conversation, where the ceiling could be.
I divide complexity into three categories: comprehension complexity, board complexity, and strategic complexity. Comprehension complexity is focused on how hard it is to understand what a card does. For instance, can you read a card and then use it correctly? Board complexity is focused on how hard it is to understand how cards in various zones interact with one another. For instance, do you understand how you can use the abilities of card A and card B together to accomplish the following effect? Strategic complexity is focused on understanding how to most efficiently use a card to maximize its effectiveness in the game. For instance, are you aware you can use a card you normally use to target yourself to instead target your opponent to help you win the game (such as making an opponent draw cards so they deck themselves)?
A lot of my previous exploration of how to raise the complexity ceiling have been about how to add complexity for experienced players without adding it for less experienced players. This led to my creation of a term I called "lenticular design" where you hide complexity by putting it in places invisible to the less experienced player. (You can read all about lenticular design here.) I'm still a big fan of lenticular design, but I'm always eager to find additional solutions to our design problems.
Today, I want to talk about a new concept that tries to raise the complexity ceiling in a completely different way. I've dubbed this "evocative design." The idea behind evocative design is that instead of hiding the extra complexity so that the less experienced player can't see it, it puts it front and center. It says, "This is complicated. It's going to take a while to understand this." It then wraps it in a very resonant flavor. The flavor does two things. One, it gives a big picture for what the complexity is trying to accomplish. While the less experienced player might not understand exactly how the complex thing works, they get a general sense of what is happening. Two, it makes it exciting. It encourages the less experienced player to want to learn about the new thing. Remember the lesson about emotional learning. It's more important that the player be excited about the new thing than understand it.
Let's use mutate as an example of how evocative design works. You can use this new mechanic to mutate existing creatures. The less experienced player might not immediately get exactly how mutate works, but they get the larger gist—I can use a creature with mutate to mutate another creature and make it change. The idea of being able to take two creatures and smoosh them together is flavorful and pretty cool. Maybe they don't know exactly how it works right off the bat, but it's exciting enough that they're willing to learn.
That's how mutate ended up in vision design. We were following the concept of a monster plane to its logical mechanical conclusion and ended up with this evocative design. I should stress that we weren't trying to make an evocative design (the concept didn't even exist yet). We were just trying to make what we felt the monster set wanted. What we ended up with was a super flavorful, very cool, complex mechanic. At first, my normal designer instincts kicked in. This was just too complex for a premier Magic set.
But as I mulled over all the information I shared about above, I realized that maybe having an exciting, fun, flavorful mechanic was worth the cost of complexity. Maybe the idea that less experienced players will just reject anything that they can't instantly understand was an outdated idea. It didn't line up with what our data had been telling us. Perhaps there was a new way to raise the complexity ceiling. My job is to find new areas of design to push in. This was a new area, so I kept it in the set.
I do want to get into one related issue before moving on. When looking at complexity, we make a firm division between what I'll call practical rules issues and theoretical rules issues. The former are things that should actually come up in expected gameplay in relevant formats. The latter are combinations of cards that we need to have rules answers to but aren't something that should normally happen all that often. For example, it's important for most players to understand how mutate interacts with cards in Ikoria as you might have that interaction happen in Limited. It's less important for most players to understand how mutate interacts with morph. I'm not saying the information shouldn't be made available, but when we worry about complexity, we're much more focused on practical rules issues than theoretical ones. Magic has over 20,000 cards. Most new things are bound to act weirdly with something from Magic's past. That's not a reason not to make new, interesting cards and mechanics.
Mutate wasn't the only flavorful thing to come out of us trying to bring a monster set to life. Along the way, we came across both keyword counters and companion. Keyword counters were interesting in that they pushed at complexity but in a very different vector than mutate. They weren't hard to understand, that is, they had a low comprehension complexity. A flying counter grants flying. No one in any of our numerous playtests ever played any of the keyword counters incorrectly. Their strategic complexity wasn't super high. Yeah, they occasionally made people think about what impact they could have on future turns (something less experienced players don't tend to do), but for the most part, they just used them when they wanted the ability. They added board complexity, but interestingly, not for the reason that most cards are board complex. They had a logistical complexity. That is, you had to track the information about them. We offset that issue by printing punch-out cards to create tokens to use as memory aids. At no point in the set, by the way, did the keyword counters exist without the expectation of punch-out cards being in the product.
Companions made me realize that there was a complexity that I never talked about—what I'll call deck complexity. That is, they make building your deck, something that happens outside of the game, harder. Normally, you don't have all that many restrictions of what cards can go into your deck. Companions add an extra layer. It's not a super-complex layer, but deck building is already challenging for most players, so adding anything to it is adding complexity.
This brings us to our next question. If mutate was in the set and we already knew we had complexity issues, why would we also put keyword counters and companions in the same set? There are a few answers to this:
1. Neither keyword counters nor companions were all that complex.
We did a lot of playtests with both keyword counters and companions. Each was pushing in a new area, but our playtesters weren't getting confused by them. (Mutate, in contrast, did cause a lot of questions.)
2. They weren't causing problems with one another.
Usually, the issue with multiple factors is if the interaction between them compounds complexity. If thing A is complicated and thing B is complicated, often the combination of thing A and thing B are even more complicated. That really wasn't the issue here. Companions forced you to think harder during deck construction. Keyword counters required you to figure out how to track them (and we helped significantly with this by giving you a tool found in booster packs). Only mutate required you thinking harder during gameplay.
3. They thematically played well with the set.
Players are more invested in figuring things out when they organically make sense to the set. Keyword counters and companion were both good fits for a monster plane.
Mutate, keyword counters, and companions all made it into set design. Each was fine-tuned and improved upon. Even then though, we did come back and ask ourselves if the combination of all of them was too much. About once a month, we have a game day, where all of Studio X (the larger group that makes tabletop Magic) stops what they normally do to focus on playing an upcoming product. This was a chance for us to interact with players that were both new to the set and less experienced at Magic. How would they fare with the set? They liked it. The flavor was evocative, and they were excited to play it. Did they understand everything they encountered? No, but it was cool enough that they were eager to play again so they could learn.
I'd talked with Aaron (Forsythe, my boss) while the set was in vision design that I felt that Ikoria was an experiment we should run. If we were going to get a better understanding of what the complexity ceiling was for premier sets, we had to be willing to try pushing things. I felt Ikoria was the perfect mix of factors to try. The other thing that got Aaron on board was that Ikoria was the last non-core set before rotation, meaning it was the one that would exist in Standard for the shortest amount of time. If we were going to experiment, it made the most sense to do it at a time where, if it went badly, it would have the smallest impact.
After the playdate, Aaron met with a bunch of people working on the set, including Dave Humpherys, Ikoria's lead set designer, and myself, and we made the call to leave the set as it was. It was playing great, and the feedback from the playtesters had been very positive. If we were going to try this experiment, this was the set to do it.
And that is why Ikoria is as complex as it is. We're experimenting. I'll let you know how it went during this year's State of Design.
Join me next week when I talk about what I enjoy about Magic.
Until then, may you run an experiment or two of your own.
In this podcast, I talk to Bill Rose, the vice president of Tabletop Magic. Bill and I have worked together for 25 years.
In this podcast, I talk to Brian Weissman, an old-school pro player. He and I talk about the early days of the game as well as how he created the first Magic deck to ever get a name.