Playing Sudoku with the Set Skeleton

Developing the skeleton of a set is like solving a puzzle in many ways; figuring out what it takes to get the game to play the way you want is a daunting task, with many objectives you need to satisfy, and some aspects of it are like solving a Sudoku puzzle.

Here are two of the constraints and guidelines that the common cards of most modern Magic sets abide by:

• Core sets have 2 vanilla common creatures in each color. 
• Almost every card should conceivably play a role in some kind of strategy.

On their own, they seem inconsequential, but taken together they are actually very constraining. Most colors asides from Green don't have a variety of sizes to choose from. For example, Red and Black can usually only get a 2/1 or 3/2 for vanilla slots of casting cost 3 or less.

The creature curve of each color should be as varied as possible, so we don't want a curve like a 2/1 vanilla in the 2-drop slot followed by a 3/2 vanilla in the 3-drop slot in the same color — that's boring. A 4-drop 4/3 followed by a 5-drop 5/3 would also feel like it could use more variety. What this boils down to is that each color is likely to get one small vanilla and one large vanilla so that the vanilla creatures of that color are spread out across the curve.

Another constraint is that we don't want a creature in one color to be too similar to a creature in the same casting cost in another color. For example, if we give Red a vanilla 2/1 as its "small vanilla," then we don't want to also give Black a 2/1. 

So in most core sets, you would ask, "OK, in this set does Red get the 2/1, or does Black get the 2/1? Whichever doesn't get the 2/1 gets the 3/2."

There could be many exception cases, and in fact I started this article by writing them out, but that looked like it would make the article hard to read so I omitted them. Of course some sets might have a vanilla 1/1 matter for some strategy by having it count as a Snow Permanent or be a Hybrid-colored creature to enable color triggers. Maybe with power creep red could get a RR 3/1. Some funky set could have a RRR 3/3 or 3/4, who knows. If it's not a core set, there's less pressure to show the basic character of these colors and defensive sizes like 2/3 would be more acceptable for Red and Black.

But here I just wanted to point out that for most practical purposes, many parts of the skeleton boil down to some kind of "either/or" choices. If these parts are spotted early, they can be filled out first. Otherwise, you may start filling out other parts first and come across a situation where you want to add a vanilla creature to fulfill the quota but you can't do so without cutting a creature elsewhere that was starting to get interwoven into the set.

So this is our current common creature curve for Red and Black:

Red Curve (1cc to 3cc)

Black Curve (1cc to 3cc)

The first thing I notice is that Muck Rat doesn't play a role in an archetype. I'd like to cut it and put in either a 2/1 or 3/2. Otherwise, the higher end of the curve for Black isn't so rich with vanilla sizes that I could design multiple vanillas that are different from each other. (I don't want a 4-drop 4/2 or 5/1 followed by a 5-drop 4/3). The 3/1 Greataxe Mercenary isn't indispensable by any means, so we can still give either Red or Black a 2/1 and the other color a 3/2. 

Another thing to notice (which I talked about in a previous post) is that all of our 2-drops provide power 2 on the board. Replacing one of the 2-drops with a vanilla 2/1 isn't enough; some of the other 2-drops also need to be changed to something that doesn't have power 2.

So it's Sudoku time. If you were to cut one of the 2-drops, what would you cut?

I tried cutting the Eerie Nosferatu and replacing it with a 2/1 vanilla. The 1/1 vanilla Muck Rats are also gone and require a replacement.

Another constrained spot to design for is 1-drops that have a useful role. There are fewer abilities that are balanced and make sense on a 1-drop. Here, I tried to adapt the Eerie Nosferatu by changing the ability a bit.

Or, maybe there can be a one-drop that you can pay mana and sacrifice to get a Duress effect.

Now I need to change Scourge of the Seas to a guy that isn't power 2.

Since one of my goals with this set's development is to have a contrast between aggro vs control, it might actually be ok to ignore the rule that creatures in the same casting cost should not all be aggro attackers. We could ignore this rule only for 2-drops, since they are so important for aggro decks. However, I decided for this example that I would like to do that only in a few colors like Green and White, and Red already has an aggro boost just by having Kird Ape as well as another playable 1-drop. I decided not to do it with Black.

I found it's a pity that I have to cut Scourge of the Seas since it looks like a card that could inspire aggro versions of Black-Blue or Black-Red decks. But I replaced it with this:

In a Red-Black deck, this card would encourage you to burn the opponent's blockers so you can attack with a +2/+1 bonus. Now that I think about it more, while this card may mold players' play styles the way I want it to, it's not necessarily the most likable card as you can't turn it on at will. It might feel fun to get the rider bonus whenever you kill something with removal before combat, but it might be just as frustrating when you can't. If it had a bigger size to begin with, maybe players could be more happy about the infrequent bonus more. Also, it doesn't help UB aggro decks as much as it helps RB aggro decks.

Another card I consider is this:

I think it's likable and grokkable as a basic vampire bat. There is another flying, lifelinking guy at 4cc, but I think it's so good that it's detracting attention from the Bond creatures like Vaportrail Imp so I would consider replacing it for this smaller sized one.

Bloodthrone Vampire might also be good here, if Dragon Fodder is staying in Red.

Pirates are a subtheme in our set, so after this swapping out the Scourge of the Seas, we might need to find another card that could be concepted as a Pirate card so that there will be at least one common black Pirate.

Now, which 2-drop should be replaced in Red?

I kind of like the play of turn 1 Kird Ape into turn 2 Runaway Chariot, an archetype defining play for hyper-aggro RG decks in our testing so far. Runaway Chariot's "drawback" makes you commit to picking up more and more Giant Growths and Auras, making you play in a different way than a typical midrange RG deck.

But I also like how Goblin Shortcutter helps Runaway Chariot, and how it allows aggro decks to get past defensive walls. In this set, I'd like to try out having more defensive options for control decks including high toughness creatures, balanced by more ways to get past them such as Goblin Shortcutter.

And finally, Dragon Fodder is also valuable since there is some kind of budding sacrifice strategy.

There are lots of ways to do this. We could replace Goblin Shortcutter with some version of Crossway Vampire. (We might need to reconcept the card since vampires haven't been Red in core sets.) Or we could replace both the Shortcutter and the Chariot with Fervent Cathar.

One of the cards I'd like to replace the power 2 guys is this:

While I want the RG aggro deck to be less explosive than it is now, I also want it to have ways to continue fighting after the opponent has started to stabilize. This card gives you a free draw every turn if your hand is empty, so it gives aggro decks hope in the late game.

Also, Dragon Fodder could be replaced with this:

Also, either of these could be costed at R; if the Swindler turns out to be too strong at that cost, the activation cost could be raised. That would push out the Goblin Arsonist; but perhaps Blisterstick Shaman could be put in at 3cc to maintain the number of pinging effects.

I went with no changes in the 1-drop range, and the 2-drops are Runaway Chariot, Goblin Shortcutter, and Goblin Swindler, with no token producing effects.

Red Curve B (1cc to 3cc)

Black Curve B (1cc to 3cc)

It may have been better to have a 2/1 vanilla in Red rather than Black. (Or to have 3 black 2-drops, one of them the Scourge of the Seas?) So what would happen in that case?

For black, I would keep Scourge of the Seas and possibly replace Eerie Nosferatu with Thirsty Vampire, or Bloodthrone Vampire if the sacrifice deck is supported. I would put the adapted Eerie Nosferatu in the 1-drop slot. We would be putting the vanilla 3/2 in Black, so something has to budge; probably the Cadaver Imp which could turn into old-fashioned Gravedigger.

For red, I would replace Goblin Shortcutter with the vanilla 2/1, then replace the vanilla 3/1 with a Crossway Vampire variant like this:

The ability could also be, "When this enters the battlefield, creatures can't be blocked by only one creature this turn."

Dragon Fodder gets replaced with Goblin Swindler. Unfortunately, I can't seem to find room for token generation unless I cut Goblin Arsonist's pinging effect or Runaway Chariot's hyper-aggro-ism. Perhaps the Runaway Chariot and Ruckus Raiser can be consolidated into Fervent Cathar so that a token-producing card can be in the 2-drop slot once more.

Red Curve C (1cc to 3cc)

Black Curve C (1cc to 3cc)

The Black curve still feels a bit off at some points — The 2-drop slot probably shouldn't be entirely evasive creatures. In the 3-drop slot, a turned-on Scorched Shambler and Barony Vampire look kind of similar. Perhaps the Scorched Shambler can be a 1B 1/1 Skeleton that becomes a 3/1 when you control a Mountain.

As I did this exercise, the effects also spread to other costs and into other colors like Blue, but I can't cover it all in detail. One of the things that spread the changes to Blue was that if I give one color a high-power evasive creature in a certain casting cost, I can't put a high-power evasive creature of the same converted mana cost in another color.

There are some constraints that I haven't satisfied, such as having a french vanilla version of most basic creature keywords. For example, Runaway Chariot doesn't qualify because it also has the "must attack" ability. There needs to be a creature with just "haste" on it, as such cards are good teaching tools. Perhaps we need to bend this rule for this set though, because the Bond creatures are already producing a lot of designs that are similar to French vanillas (in terms of what they do, not their complexity) that it would be a strain to have more such designs and less abilities such as enter the battlefield effects or tap abilities.

I hope this exercise demonstrates the nature of this process, namely:

• Things change a lot, by necessity.

Even if development is not trying to be overzealous, a lot of cards need to be shifted around or replaced by new designs. The only thing that design can hope to pass on into the final product is a vision of what kind of play they think is fun, and many, many candidate cards and replacement cards that development can take to fit together like a jigsaw puzzle. The specific surface details can change completely. It's not like just testing the individual cards one by one and fixing broken costs. For this set though, I hope the design team can join in to produce lots of replacement cards, especially for utility creatures and archetype-inducing cards.

• Every change spawns a whole chain reaction of changes. 

This has two implications. First of all, this requires a method of communication where people are either in the same room, talking through the necessary wave of changes in real time, or a method where people propose the entire batch of changes as one thing. We can't do this with a style of editing where one person proposes a change to one card, waits for approval, then proceeds to propose a change to another card. So anyone on the team who wants to propose their own version of the skeleton is more than welcome to.

Secondly, everything in the set depends on everything else. You can't look at a card and tell if it's going to be in the set just by itself; it could depend on something external like whether the vanilla 2/1 went to Red or Black. (Unless the card is central to the set in some way.)

So although time is running short, the focus of testing should still be on what archetypes are going to be supported, and how the influence of Bond can be made most meaningful. Any detailed tweaking that is done before deciding on that will be lost if the set structure changes (such as how many aggro 2-drops a color gets) based on how we want Bond to play out.