Decisions, Decisions, Part II

Posted in Making Magic on August 10, 2009

By Mark Rosewater

Working in R&D since '95, Mark became Magic head designer in '03. His hobbies: spending time with family, writing about Magic in all mediums, and creating short bios.

Two weeks ago, I wrote a column examining the use of decisions in game design. As this week's column is the second part of the two-part article, before beginning today's column I'd heartily recommend you take a look at the first part if you haven't already. Last time I stated that I would examine decision creation in five parts of the game: the card level, the mechanics level, the set level, the block level, and the meta-rules level. I did the first two last time, so odds are I'll discuss the other three today.

But before I get to that, I am going to take advantage of my medium. Due the nature of Internet column writing, I had the opportunity to see feedback on Part I before I wrote Part II. As such, there are a number of issues that came up concerning part one that I'd like to quickly address before moving on to part two. I feel this interactiveness is a key of what makes the Internet such a cool and unique medium for writing.

Anyway, here, in no particular order other than my own internal preferences, are the issues that I feel need to be addressed:

Expunge is a wonderful card. We joke in the Pit that the main topic of my thread is seldom the main topic of my column. So was the thread two weeks ago all about the role of options versus choices? No, the vast majority of my thread was spent defending the card Expunge. (I kind of said it was a bad design.) How dare I claim such an awesome card as a failure?

Let me clarify what I said, as I think the topic drifted a bit. I didn't say Expunge was a bad card. On the contrary, the card saw plenty of play, and as the thread shows, there are many players who liked it. I was claiming that it was a bad cycling design. My critique was a technical one, looking at how well the card took advantage of having cycling. Several posters objected that they would rather play Expunge than Terror, and that the cycling did occasionally matter. My response is that of course it matters. It would be hard to put cycling on any card and not have it come up some of the time, if for no other reason than cycling helps grease the deck to avoid mana issues.

This ties right back into my point from two weeks ago. Players love options because humans love options. Why wouldn't you want extra functionality on anything you use? One of the reasons I'm so gaga about my iPhone is that it seems to gain functionality every week. But, and this is the whole crux of this two-parter, what players want and what makes better game play are often not the same thing. That's so important I am going to repeat it: what players want and what the game needs to succeed as a game are not synonymous.

Yes, Expunge is essentially better than Terror. And yes, players will be attracted to the thing that is better most of the time. Does that mean we should stop printing Terror—that we should only make black kill spells that do stuff in addition to killing things? No, it doesn't. Your job as players is to want as much candy as you can get your hands on. My job as the game designer is to make sure that I don't let you have so much access to candy that you all get stomachaches.

I use the above child/parent analogy with some trepidation as there is a lot of baggage that comes with it. Let me be upfront that I do not intend to offend anyone with it. The reason I chose to use this analogy is that it is a good example of an interconnected relationship where one party is responsible for the welfare of the other. My kids might not like getting shots, but I still have to make sure they get them, as I know their lives overall will be better for having had them. Likewise, part of the role of being a good game designer is not giving the players what they want at the expense of giving them what they need.

I like the original version. Some players were curious what was wrong with the original version of my card in the section on card level design, the first "Guy with Spear" that just had activated first strike. There was nothing wrong with it. In fact, the simplest card is usually what we try to print. The question at hand, though, was how we layer on choices when we feel a need for an extra decision. In most cases, the original version is what I would print.

I will point out that the third version, with some hindsight, was a little less optimal than I would have liked for my example. Of course, you will shrink your creature if the only other option is death, so it mostly fails at the point I was trying to make: when you add on decisions, make sure the new decision adds reflection rather than just an obvious preemptive decision. Mea culpa. I used a suboptimal example.

Options need to and should exist. When I said choices make for better game play than options, I was not saying that options should be stripped from the game. As demonstrated by the conversation around Expunge, players love having options. The goal of this two-parter is to explain that options need to be limited, especially when there are choices that fit in the same space. Choices need to be the default, but in the right places for the right reasons, options can and should be used. (Remember, I'm using "option" to refer to additive decisions and "choice" to refer to modal decisions. I'm well aware that by the technical dictionary definition they are synonyms.)

My explanation of googolplex was mathematically ambiguous. You got me. I'm a word guy, not a numbers guy. I do know it's a very, very big number (and the name of the movie theater in The Simpsons.)

With that out of the way, on with part II.

Level III – The Set Level

When you boil it down to its essence, set design requires three basic things:


  1. Newness – Market research has shown time and again that the number-one thing players look for in an expansion is new things, especially new mechanics. Magic keeps reinventing itself, and players are always eager to see what new veins of design space we've found.
  2. Familiarity – Each expansion has to feel like Magic. As much as players want to explore the unknown, they want to do it within the confines of the known. While the new thing will grab the focus, the "same old stuff" actually needs to make up the vast majority of the set.
  3. Integration – The new and old things have to work together. The set needs to feel like a cohesive whole.

The key to accomplishing these three goals lies in making use of decisions in the set's infrastructure. As an example, let's take a look at Ravnica: City of Guilds. The newness of the set was in the addition of the guilds. True, each guild had its own keyword, which clearly helped getting the newness in front of the player, but the majority of the make-up of each guild was abilities normally found in those two colors.

As an example, let's examine the Selesnya guild (green-white). The guild had a commune feel, as the guild was very much about expanding the group and using that resource as the means for victory. From a design standpoint this meant we had to play into green and white's emphasis on the community. Before we could do that though, we had to take a step back.

When looking at the guilds and the overlap between them, we had two different possibilities. We could either build them so that each guild offered something the other guilds did not care about, or we could build each guild out of pieces that overlapped between the guilds.

The two guilds that Selesnya overlapped with in Ravnica were Golgari and Boros (black-green and red-white respectively). Each of these guilds shared one of the colors of the Selesnya guild. Now, when designing green cards, for instance, we could have made the conscious decision to make half the green cards mechanically relevant for Selesnya and half the green cards relevant for Golgari. When drafting the set, if you took a Golgari-only card, you would still have the option of taking a Selesnya card, but in practice, picking more Golgari cards would be just better. This is the "options" method, where themes do not fight with one another and allow each person to easily find the cards they need for their deck.

We could instead design green cards so that many of them fit both Selesnya and Golgari. This would mean that when you drafted a green card that was very efficient at making tokens, you wouldn't know yet how exactly you were going to use it. Would it become the backbone of your army building deck complete with white cards buffing the group or would it be the means of producing lots of tokens to fuel your black sacrifice outlets? This is the "choice" method. The key is that now you will have to make choices about your deck, instead of getting to coast past many easily selected options.

Obviously, we took the second path. Why? The short answer is better game play, but as my column really isn't about the short answer, let's look at the long answer: What does creating choice in this area do for game play?

Increased opportunity – If we divvied up the green cards into two halves, we simply lower the number of cards available for any one guild. By making monocolored cards that have dual function, we make more choices available to the players—will you use any given card in a Golgari deck or in a Selesnya deck? This is important not just to enable Sealed Deck play but also to make Draft more dynamic and have more replay value.

More diversity – If half the mono-green cards only made sense in the Selesnya deck, then only the Selesnya drafters that would ever get them. By making cards that have dual functionality, there are now multiple drafters that are interested in the card. This means that the decks have less chance of ending up identical each time.

Interactivity – When a card has a dual function, occasionally decks find a way to make both aspects matter. Perhaps you're playing Selesnya but you are playing in a format that allows access to other cards. There are green and white cards, for example, that can allow you to use cards much as you might normally use them in a Golgari deck. For example, the card Natural Order from Visions, allows you to turn a green creature into any green creature card from your deck. In a format with Natural Order, token making might be a means to get your biggest green creature out of your library and into play in addition to building your army.

Increased integration – And now we get to the biggest one of them all. How do we interconnect our newness with our familiarity? By making each part connect with the other parts, forcing the players to make interesting choices.

The lesson here is that good set designs want you to create choices between your set components. When the players have to evaluate one against the other, it helps create a holistic sense to the set.

Level IV – The Block Level

If set design is about integration, block design is about evolution. The key to good block design is figuring out how the design changes through the course of the block. Once again, there are two distinct paths available. One evolution is additive (what I'll refer to as the option evolution) while the other replaces (what I'll call the choice evolution). Let me walk you through each of the two types of changes:

The Option Evolution – In this block evolution, each set adds more for the deck in question. The fundamental structure of the deck never changes, but rather cards get swapped out as more efficient cards or more cards of the subset required become available. As an example, assume a deck is built around mechanic X. The deck using the first set has all the mechanic X cards worthy of Constructed play and supplemental cards of the color(s) mechanic X appears in. The second set adds more cards with mechanic X. The deck changes as it now gets more cards with mechanic X and external cards get weeded out. The third set comes out and there are even more cards with mechanic X. The deck now gets to add the additional cards with mechanic X, perhaps allowing the deck to drop the number of colors or allowing the overall strength of mechanic X to rise. The play experience of the deck doesn't change all that much from first set through the third, but individual card additions do make slight changes to how the deck plays. The mechanic X deck has gained options, but no choices of substance.

The Change Evolution – In this block evolution, each set changes the deck in question. As the block evolves, the essence of how the deck works keeps getting reexamined. As an example, assume a deck is built around mechanic X. In the first set, the deck is made of all the constructed worthy cards with mechanic X plus support cards in the relevant colors. In the second set, something is shifted in how mechanic X works, and possibly the support for mechanic X also changes, making new things relevant. Thus, with the second set, the deck starts morphing into a different kind of deck. Perhaps it changes colors or shifts its focus. Maybe a new win condition gets added or old cards that got cut now have new importance. When the third set comes out, the deck gets fundamentally altered again. The players have to take a step back and reexamine what they know about the deck. With all three sets available, the deck is a different animal than what you saw at the beginning of the block. Players have to make choices about how to best use mechanic X.

Both of these evolutions are viable and we've used each many times, but as I've grown as a block designer, I've found my preference for the change evolution. Here's why:

The roller coaster theory. I once had a writing professor who said that a good story is like a roller coaster. There's build-up and release. There are twists and turns. There's suspense and adrenalin. You get on the roller coaster having a general sense of what to expect, but an unexpected turn at the right moment sends chills down your spine. Basically, what he was saying is that an author needs to be able to take his audience on a ride and that much of the joy of riding roller coasters—or enjoying stories—comes from the surprise of the twists and turns within the confines of a pre-understood experience. I feel the same way about good block designs. I want the players to have a sense of where we're going, but I want the ability to surprise them as we travel along.

Change is good. I've talked many times before that at its heart Magic is a game about change. If that is so, why should blocks be any different? I like the fact that the environment continues to shift as the block evolves.

We must keep surprising. In one of my more colorful metaphors, I've described our player base as the Borg (from Star Trek). Our players adapt to whatever we do, so we can't use the same trick twice. This has forced us to get much more innovative in how we create blocks. Once upon a time we could get away with "here's some" and "here's some more" and "here's even more with some tiny twist," but times have changed, and I feel the players expect more from us. That means we have to deliver. We have to find ways to surprise you, and the change evolution seems more prone to accomplishing this task.

Technology keeps improving. In sixteen years, we've learned an awful lot about Magic design. Another reason why I lean towards change evolutions is that we have tools to do them in ways we didn't many years ago. Our willingness to be more flexible in how we shape a block, even willing to shape the sets themselves in size and number, has opened up all sorts of new opportunities.

The take-away from this level is that choices allow us to create a richer block environment. Yes, we'll still make some use of options, but when push comes to shove, I believe choices will create much richer blocks.

Level V – The Meta-Rules Level

And now we finally get to the topic that so many people have been asked me to write about: the Magic 2010 rules change, more specifically the removal of damage from the stack. Before I open that kettle of fish, let me say a few things about options and choices as they apply to game systems in general. I've explained the advice given to me by one of my communications professors: no scene is worth a line; no movie is worth a scene. This advice meant that no matter how good a line was it shouldn't be put in a scene if it didn't service it, and no matter how good a scene is it shouldn't go in a movie if it didn't help the overall movie.

My game design corollary is this: no rule is worth a game. If a game can exist without a rule and function smoothly, then the rule is not needed. Rules should exist only because they service the game. Essentially, a game should exist with as few rules as it needs to do the things it does. This is why we pulled mana burn from the game. You remove it and the game functions perfectly fine.

The tricky thing about this strategy is that often rules that seem irrelevant aren't. Two perfect examples from Magic are the cards Yawgmoth's Bargain and Tinker.


Yawgmoth's Bargain

The design of these two cards came from me messing with older cards trying to create "fixed" versions. The two older cards, if you are unaware, were these:


Transmute Artifact

In Yawgmoth's Bargain, I was trying to make a cleaner (and in theory less powerful) version of Necropotence. I took off the pesky "Put that card into your hand at the beginning of your next end step" text as I thought it wasn't necessary. Likewise I removed the whole part about needing to pay the difference in converted mana cost from Transmute Artifact to create Tinker because I thought it was unnecessary text. In each case I learned the hard way that what seemed like unneeded excess text was not.

Back to ridding the game of unnecessary rules. One way to rid a game of rules is to look for things that are not essential. Another way is to look for rules that are limiting rather than expanding the game. Essentially, I'm looking for options that eclipse choices. Before I jump into the recent rules changes, let me give an example from another game. Suppose Tic Tac Toe had a rule that said that once per game each player could play twice in a row. It's an extra rule. It grants both players an additional ability. Isn't that good? No, because it leads the game to a locked state. The first player puts his X on the board in either the center or a corner, says go, and then is guaranteed a win on his next turn no matter what action the second player takes. You have more options, but the end result is that you actually have no interesting decisions.

Now let's finally talk about "damage on the stack." As far as I'm concerned it's an option that reduces decision-making. Here's an example. Turn one, your opponent plays an Elite Vanguard. You play a Sakura Tribe-Elder on your first turn. Then on their second turn, they attack with the Elite Vanguard. What do you do?

Pre-M10 rules changes, you blocked, put damage on the stack, and then sacrificed your Sakura Tribe-Elder for a land. Sure, you had other things you could do, but all of them were eclipsed strategically by this play. Under M10 rules, the same thing happens. What do you do? It's not so clear. You could block and kill the Elite Vanguard or you could block and then before damage sacrifice your Sakura Tribe-Elder. This allows you to get a land but at the cost of not destroying the Elite Vanguard. It's no longer an automatic decision.

I understand why players were unhappy with us removing "damage on the stack," but it comes down to the same reasoning why so many players argued fervently for Expunge. Options are desirable. Of course, if given the decision between two cards, players will always prefer the one with more options (provided, of course, that the costing doesn't make it unusable). The same goes with rules. Players like being able to do more things, but that doesn't make it produce better game play.

But wait, didn't we say that "damage on the stack" was removed because it was unintuitive to new players? We did and it is. That was what made this decision so appealing to us. Not only did we remove an unnecessary option from the game, but in doing so, we also rid Magic of something that was confusing newer players. We were able to remove complexity and add to strategy. Win-win.

To be fair, removing "damage on the stack" did reduce strategy in one way: it lessened the ability for more experienced players to win because they were aware of a rule the lesser experienced player didn't know about. To be pretty blunt, I have no problem lessening that type of strategy. With a game as complex as Magic the player more experienced with the rules always has an edge. I don't see any need to make it a greater gap than it already is, and I am more than happy to lessen that gap whenever doing so creates better and more intuitive game play.

Why did we remove "damage on the stack"? To both make the game more intuitive and increase in-game choices. Yes, it affects how the game is played. Yes, it makes some cards better and other cards worse. Yes, it requires players to shift their thinking. My blunt response to that is: welcome to Magic. All of the above is what design (and development) does every day in every set in every block since the game's inception. And as always, the shift has affected how we design. For example, I designed a cycle of cards for "Lights" that never would have been printable power-wise under pre-M10 rules. As part of the game shifts, so too does the rest of the game.

Choice – The Right Choice

Our job in R&D is to create the best game possible for all of you. To do so, we have to make hard decisions that aren't always popular. My hope with this two-part article is that I showed you a very different vantage point when looking at the game. I want you to see that we have the best interests of Magic at heart. We just approach the game differently because our focus isn't the same as that of the players. It is my firm belief, though, that these decisions lead us to create a better game. And if things like the sales of Magic 2010 or the attendance of events like Grand Prix–Boston are any indication, I think we're doing pretty well. (If you disagree or just think I'm being too cocky, please let your voice be heard in the thread.)

Join me next week when I visit a place that few cards ever return from.

Until then, may you have fun even if you're not always aware of why.

The survey originally included in this article has been removed.

Latest Making Magic Articles


Darkest Hour, Part 1 by, Mark Rosewater

Welcome to the first preview week of Hour of Devastation. I'll quickly introduce the design team, and then I'll tell you how the set design came together, and along the way I have a sweet...

Learn More


Metamorphosis 2.0 by, Mark Rosewater

Making Magic is an iterative process. We create something, we playtest it, we get feedback on it, and then we use that feedback to either make more new things or to adapt the things we al...

Learn More



Making Magic Archive

Consult the archives for more articles!

See All

We use cookies on this site to enhance your user experience. By clicking any link on this page or by clicking Yes, you are giving your consent for us to set cookies. (Learn more about cookies)

No, I want to find out more