The Design of Splitsville

In the weeks before a new Magic set comes out, magicthegathering.com runs columns about how it was designed; these columns also typically "spoil" one or two of the upcoming cards. I thought it would be fun to do something similar for Splitsville, so I did.

Some of the below essays refer to Saturday, April 25th (2009), which was Splitsville's "launch" date -- the day I first had a group of people over to play with the set.

  1. Why Splitsville
  2. Splitsville: The Numbers
  3. The Flavor of Splitsville
  4. Splitsville: Flip Fantasia
  5. Right Sizing Splitsville

Why Splitsville?

I don't know about you, but I like options.

I like going into an ice cream store and reading all of the flavors, even if I always order chocolate. I like going into stores period, and marveling over all of the wondrous choices that modern capitalism affords. Fixe prixe? Non, a la carte!

I especially like games with lots of options. My current favorite game is Race for the Galaxy, in which every card can be either directly played or used as a payment. I'm fond of the Vs. system collectible card game, in which every card can be either be played or used as a "land" in the Magic sense. I even designed a strategy game called Barons! (still due out in 2009, as far as I know) that combined the two to allow every card to be played in three different ways (directly, as a land, or as a payment).

I understand that not everyone is like this. Some people feel like their head is going to explode when presented with a combinatorial number of possibilities. Mine does too, but somehow I like the feeling ...

Anyway, my love of options is one of the reasons I decided to design a Magic set full of flip and split cards. We'll ignore the bigger question of why I decided to design a Magic set at all; that's best left between me and my therapist.

The bigger reason is that I felt that in Magic currently, split and flip cards were both badly and under used.

Take split cards. There have been only 23 split cards printed, in just 4 (5 if you count Time Spiral reprints) sets. And all in the "Cliche Word #1 // Cliche Word #2" pattern, which I frankly find just stupid.

Don't get me wrong. I've read http://www.wizards.com/Magic/Magazine/Article.aspx?x=mtgcom/daily/af175 in which Magic developer Aaron Forsythe correctly points out that split cards are a very broad mechanic which if left unchecked could encompass any number of other Magic mechanics (kicker, buyback, threshold, cycling, etc.).

What I object to is using names to narrow the split card mechanic. Especially names like "Supply // Demand" or "Research // Development" that make little sense in a fantasy game. (I will admit to being curious about how these names get translated to other languages.)

Think about it: outside of Unglued, no other Magic mechanic is name related. If Rosewater had proposed having all cycling cards start with the letter "C", or have every buyback spell be titled with a palindrome, they would have locked him up. (I think. I hope. To be honest, for all I know, Rosewater did propose those things and they did lock him up, and he's been head designing from a loony bin for all these years.)

In Splitsville, I attempt to demonstrate how split cards can explore unique areas of design space while adhering to some rules that both provide structure and prevent them from taking over all of Magic.

Here are some of those rules:

  1. No permanents: each part of a split card must be a instant or sorcery spell. There are two reasons for this. First, in a split card with two permanent spells, it is annoying to have to keep track of which one was cast. Second and more importantly, this sort of thing can be much better done through flip cards. (See Splitsville: Flip Fantasia for more about Splitsville's take on flip cards).

  2. Mechanically or thematically linked. By this, I mean similar in the same way as the cards in a cycle are similar, by being variations on the same theme or design. I think of mechanically linked cards as ones that you can immediately see the similarity from just the text box (White Knight / Black Knight, for example), while thematically linked cards get a little help from the name and flavor. The attached Heal Other // Heal Self is a good example of a thematically linked card; the two parts have no mechanics in common but are brought together through the concept of healing.

  3. No repeating: the two parts of a split card aren't allowed to repeat sentences or major effects. This gets around 90% of split cards muscling in on the territory of other mechanics. We do allow repeats of minor effects, for example Draining Touch // Fiery Touch both include "do 2 damage to target player" as one of their effects. What a minor effect is is obviously context dependent -- if this was the only text on the cards, it would be their major effect, but as is it isn't, so it is allowed.

So that's why I did Splitsville: to show that split and flip cards, properly designed, could do a lot more than they currently do in Magic. And, oh yes, be a lot of fun.

You will ultimately be the judge of both. I look forward to your reactions.

Splitsville: The Numbers

Splitsville was designed to support a 8 person draft. Let's see what that means.

In a draft, each person opens 33 commons, 9 uncommons, and 3 rares. Multiplied by eight, that makes 264 commons, 72 uncommons, and 24 rares.

Now, I didn't want to design 264 different commons, and even if I did, I wanted the commons to appear more frequently than the rares. I settled on ratios of 3:2:1 -- each common would appear three times, and each uncommon twice, as compared to each rare.

To put it another way, I targeted the set for 88 commons, 36 uncommons and 24 rares (148 cards total). Conflux, a similarly sized set (145 cards total), has 60 commons, 40 uncommons, 35 rares, and 10 mythic rares. With Conflux, you will see a given common 4.4 times on average in a draft; with Splitsville, the commons frequency will feel much more like a large set. (Shards of Alara had 101 commons, so each would appear ~ 2.6 times on average).

The downside is that every draft of Splitsville will have almost exactly the same rares, while with Conflux you would have mostly different rares. I figure if we end up drafting Splitsville so much that the rares begin to bore us, I can always design more.

(Though, perhaps because I am a limited player at heart, I find it much easier to design commons than rares.)

I actually ended up overshooting my targets by about 25 cards; this will also add to the variability.

How does this effect you? Well, when you arrive at 2 pm on Saturday April 25th, I'm hoping that you will help me sleave the over 500 cards we will be playing with :). If that sounds huge, keep in mind that it is just ~65 cards per person (45 card draft set + ~20 lands per player), or less than a constructed deck plus sideboard.

The Flavor of Splitsville

How did the denizens of the plane of Splitsville develop their unique duality based magic?

Some people say the answer is lost in the mists of time. Those people are fools, because the historical record clearly indicates that it was because of a decree made by King Bob the 17th (aka "The King Who Liked To Make Decrees").

Specifically, King Bob moved Splitsville off of the tax-per-spell (TPS) system so common throughout the multiverse onto a tax-per-card (TPC) system. The predictable result was a dramatic increase in the number of spells per card.

Man, I suck at flavor.

Which is all to say: a 98% split & flip card set doesn't end up having lots of room for flavor text, so you've all been mostly spared my puns and bad jokes. I know some of you were thinking about skipping the Splitsville pre-release for precisely that reason, so sleep easy.

I do want to apologize for two things.

First, the art. In a more perfect world, I would have found two pieces of art for each flip card (one for each head), and glued them together along a diagonal. There's still a chance that I might do that for a few cards, but don't hold your breath.

Second, the card names. Right now, they are a hodge podge of (a) words stolen from Magic's past (Aphetto, for example) and (b) what might be best described as "designer names" -- "designer" here not in the sense of "designer babies" but rather the sort of names that a real Magic designer would use as a temporary play test name, then hand off to the creative team to fill in. (Plush Toy, Mishra's Junkyard, Big Tree, etc.)

A better set designer than I would have invented a whole world to go along with the cards and used that to influence the card names and flavor text. Frankly, I lost my interest in that sort of thing when I stopped DM-ing at age 16.

So that's what Splitsville is, for better or for worse: heavy on the design, light on the flavor.

King Bob would have wanted it that way.

Splitsville: Flip Fantasia

Flipsville might have been a better name for my set, given that flip cards outnumber split cards by about two to one, except that Flipsville is a completely idiotic name for a Magic set while Splitsville is merely stupid.

Anyway, like split cards, flip cards are another totally under and ill used Magic mechanic. Only 16 have been printed (17 if you count Unhinged's Curse of the Fire Penguin), all in the Kamigawa block, with the majority flipping from a creature into a legendary creature, and a few from Saviours of Kamigawa flipping from a legendary creature into a legendary enchantment.

I've got nothing against legends, and I actually love the flavor of these flip cards -- what could be more mythic than an ordinary creature doing something unusual and transforming into a legend? It's just that there are so many other forms of transformation -- young into old, caterpillar into butterfly, novice into master, good into evil, etc. -- that also have great flavor and that flip cards could also easily represent.

And once we are opening up the doors on what flip cards can do, it is immediately evident that there are two whole new possibilities beyond what I call the "one-way flippers":

1) The "no-flip flippers". These are cards that can be played in one of two ways, but do not flip once played. These are mostly like two way split cards, except that the headedness keeps track of which version was played, making them a much better solution for permanents.

Flavorwise, the no-flip flippers are good for representing "mirrors", Magic R&D's term for two card cycles like Black & White Knight. They are also good for representing binary comes-into-play outcomes that would otherwise require counters to keep track of. The design space here overlaps some with kicker, but there are also some clear differences. Want a comes into play effect for extra mana? Use kicker. Want two different sets of abilities, or power/toughness differences that aren't easily expressed with +1/+1 counters, maybe even at the same casting cost? Use a no-flip flip card.

2) The "two-way flippers". As the name suggests, these are flip cards that can flip both ways. These are good for representing both temporary effects and entities with dual natures. X is an example of a two-way flip card representing a temporary effect (in this case, a version of rampage); these cards typically only have one casting cost. For entities with dual nature, such as Tick/Tock, both heads will often have a casting cost.

I should note that one-way flippers might also have a casting cost on each head; for example, Aphetto Optimist.

And now for a shameful confession: I had designed over a hundred flip cards of the above types when I made the discovery that no-flip flippers, two-way flippers, and one-way flippers with two casting costs all don't work according to Magic's current rules.

One of the problems is rule 508.4:

508.4. Flipping a permanent is a one-way process. Once a permanent is flipped, it's impossible for it to become unflipped. However, if a flipped permanent leaves play, it retains no memory of its status. See rule 510, "Status."
In Splitsville, we ignore rule 508.4. We also need to change rule 508.2:
508.2. In every zone other than the in-play zone, and also in the in-play zone before the permanent flips, a flip card has only the normal characteristics of the card. Once a permanent in the in-play zone is flipped, its normal name, text box, type line, power, and toughness don't apply and the alternative versions of those characteristics apply instead.
Basically, we replace rule 508.2 with the split card rules: if you ask a question about a flip card that is not on the stack or in play, then what you do is ask both halves separately and then take the "OR" of the results. (I.e., "yes" if either is "yes", "no" otherwise). For a flip card in play or on the stack, instead look at just the half that is either active in play or being cast.

Finally, we have to make a small change to rule 508.1c:

508.1c A flip card's color, mana cost, expansion symbol, illustration credit, and legal text don't change if the permanent is flipped. Also, any changes to it by external effects will still apply.
The problems here are color and mana cost; we want them to change when the permanent is flipped. To be specific, we always want to change the mana cost; we keep the old color if and only if the side we are flipping to doesn't have a mana cost.

All in all, my hope is that these rule changes merely make flip cards work the way you think they do. (Or at least the way I thought they did...) Still, there are two subtleties that may not be immediately apparent:

  1. If you Flicker a flip card out of and then back into play, it always comes back as the "unflipped" version (i.e., the version that puts the card number on the bottom).
  2. If one half of a flip card doesn't have a mana cost, then it doesn't have a mana cost. (Just like Evermind). Spells without mana costs can't be played, but lands without mana costs can be. The converted mana cost of a flip card flipped to a side without a mana cost is 0.
Anyway, I hope that you don't flip your lid as you flip through the attached cards. And, not to be flippant, I also hope that come April 25th, you will enjoy being flipped off by your opponents!

Right Sizing Splitsville

Splitsville has a "converted mana cost matters" subtheme. It is unlike Scourge's, however, in that rather than encouraging big casting cost permanents, we reward creatures and spells with the specific converted mana costs of one, two, three, four, five and six.

I put this mechanic into Splitsville for a bunch of reasons. First, I wanted to encourage some non-color based draft strategies, by which I mean something like drafting Arcane in Kamigawa block or Elementals in Lorwyn -- they might lean on some colors more often than others, but aren't defined by their colors in the same way that a red/black beatdown deck is. (To be honest, it isn't that I'm particularly fond of these sort of strategies as a drafter; rather, I thought it would be an interesting design challenge to put some in Splitsville.)

In particular, I wanted to show off the way split & flip cards could pack a rather large number of such strategies into a relatively few cards. I think this is big advantage because it reduces the number of relatively uninteresting cards for someone drafting a more traditional color based deck, and because it creates scarce resources for those who are drafting cmc = X decks. A given 8 person draft can almost certainly support only one cmc = 5 deck, but more to the point, it can probably also only support three or four cmc = X decks at all, because they are all going to be drafting highly the same set of cards.

Second, I wanted a subtheme that would play well with Splitsville's dominant mechanic of flip and split cards. A large number of flip cards change their converted mana cost when they flip, so while you might normally want to flip your 3 cmc dude into something bigger if you had the chance, you might choose not to if you had a Frehen Shaman out. Many of the flip cards that flip into something powerful have no converted mana cost on the awesome side, which lessens their strength just a little, as there are no cards in Splitsville which reward cmc = 0.

Finally, I'm a contrary person and I relished the thought of encouraging folks to make decks that in many ways went against the grain of the fundamentals of Magic's mana system. For a cmc = 3 deck, "curving out" means dropping a 3 cmc card on turns three, four and five :).

The specific choice of mana costs one through six came as follows:

  1. I wanted to avoid cmc = 0, just so I wouldn't have to worry about edge cases relating to lands or tokens. As mentioned above, I also wanted to slightly penalize flip cards without a converted mana cost.
  2. I wanted an even number, so that I could put a cycle of a particular type of "cmc = X" cards onto a whole number of flip cards.
  3. I wanted a number close to five, so that each cycle could have a representative from each color.
To make considerations #2 and #3 jell, I used artifact as the 6th color for all of the cycles.

Assigning all of the converted mana costs and colors to the cycles turned into a giant sudoku puzzle. Literally. Consider: I wanted to do 6 different cycles, namely

  1. Other creatures you control with cmc = X get +Y/+Z,
  2. Other creatures you control with cmc = X get some static ability,
  3. Other creatures you control with cmc = X get some tap ability,
  4. For every creature you control with cmc = X, get something,
  5. Whenever you play a spell with cmc = X, do something, and
  6. Some big rare spell that rewards creatures with cmc = X.
So, for each combination of cycle a-f and one of the six colors, I needed to assign a number between 1 and 6 (the converted mana cost) in such a way that for every color, each cycle got each number once and that for cycle, each cycle got each number once. If you put the cycles as a columns and the colors as the rows, that's the same thing as filling out a 6 by 6 sudoku! (Well, Latin square, actually, because there are no sub-boxes. But still.)

(And then, just because I'm an OCD mathematician who is obsessed with symmetry, I had to figure out a way to pair up the choices so that no color pairings appeared more often than any other, and the same for converted mana cost pairings. This is the reason why the rare, big spell [f-cycle] cards aren't flip or split -- 6 choose 2 = 15 doesn't divide 6*6 = 36, but it does divide 36 - 6. Yes, you should pity me.)

As with most everything else in Splitsville, this is all just a big experiment. It will be up to you next Saturday to try it on for size and then size it up. Until then, you'll just have to pick on someone your own size.