Friday, May 18, 2012

Creating the Test Skeleton

Last week, I created a set file for testing some changes in the Bond mechanic. I would like to describe what considerations I made at each step so that it may provide ideas to other people who develop a set they've made with friends. I also hope these ideas will also help the M13 team when we collectively try to settle on the shape of the skeleton.

In the card file for testing the Bond mechanic, I made many changes rapidly. The first objective was to make sure Bond mattered and wasn't outclassed by other common creatures. One of the ways Bond was supposed to matter was to create interesting dynamics in choosing your colors in Limited, but I felt that was not happening. I had to make a version of the skeleton that follows normal power level guidelines so that Bond can really shine, and see how the color constraints could affect game play when Bond creatures do outshine other commons.

Secondly, I wanted to try out some changes to the Bond creatures and the bonuses that Bond assigns. We had been experiencing some problems with the cards Windborne Pteron and Mastodon Calf.

Windborne Pteron was difficult to balance; it seemed too powerful as a 3/3 flyer since there are few commons that could deal with that, so it had been dialed back to a 4-drop 3/2 at times, where it would be an Assault Griffin with extra constraints attached. Mastodon Calf didn't seem that exciting to some people who gave us feedback. The criticism that "we use the set mechanic and all we get is an Assault Griffin and a Siege Mastodon?" could not be avoided.

Jay had talked to some people and thought a 1W 2/2 Pteron and a 3W 3/2 Mastodon (that turns into a 4/4 with a Forest) would be a good size, and I wanted to try that.

I also thought more Bond cards should offer the "multicolor experience" where cards like Deft Duelist or Branching Bolt have the traits of multiple colors to have an attractive combined effect that do things a mono-colored card couldn't do. Changing the Mountain bond bonus to haste was one of those — you have a decidedly Green-looking Wurm's body and a decidely Black-looking Undead's body, but once they have the Red keyword, they act completely differently. I thought that with these versions, the set mechanic would actually affect the card's play patterns and the way combat worked out for those creatures, rather than just provide a dry strategic advantage of a cost reduction. I felt it changed the cards in a way that can be experienced on a visceral level.

Finally, there were some draft archetypes I wanted to propose. These things were better proposed in the form of a set card file to show how a cluster of cards could work together and how it could fit, rather than in the form of individual cards.

So those three things were what I wanted to try out. Before I go to the details, I want to talk about the role of development.

One thing that development does is to take cards and build archetypes around the cards, adding cards to make that happen. An example of that is the token theme in Rise of Eldrazi. The Eldrazi Spawn tokens started out as "mana crystals" — floating pieces of mana that you can cash in on any time. They were turned into "manabots" much later in the process. Once they became creatures, the development team added support for attacking with a swarm of tokens instead of just using them for mana.

This is a mental image of what development does.

A card file before development:

A card file after development:

The images are of a landscape that a Minecraft player named Monkeyfarm came accross, and the building he made incorporating the terrain features. (For those not familiar with the game Minecraft, it's a game where you explore a world with amazing landscapes and collect building blocks to build houses, lego-style).

The reason I use this metaphor is because development should use the natural quirky features of the landscape and incorporate them as central parts of the set. Development isn't about painting over what design did with a different creative theme; rather, I imagine it as similar to Minecraft when you look at a landscape and plan, "this looks like a natural stairway! This outcropping could be an observatory! This concave can be an entrance hall the leads out to the top. What if I connect this spire and that cliffside with a bridge?"

Of course, if the Design stage had a very specific vision about the fast or slow pacing of the set, or about some archetypes that should be enabled, Development would follow that. But without such parameters, Development often starts out with questions like "with these cards, what do we want each color pair to play like?" or "what archetypes can sprout from these cards if we support them?"

With that in mind, these are the changes I made to the test file since the last time I posted "Playing Sudoku with the set skeleton"

Red: This went almost unchanged from the last time.

Black: This was like an endless maze for me. I would change one part of the creature curve and stare at it. The change would make me want to change another part to accommodate for the first change. It's really hard to settle on these things since every part is a potentially moving part.

This is the curve I was considering at the time of the last article:

One of the main problems with this configuration from last time was the 3-drop slot: There was Vaportrail Imp (3/1), Barony Vampire (3/2), and Scorched Shambler (2/2 that turns into a 4/2) all in the same cost. The similar size was a problem, but also the fact that conditional creatures such as the Mountain Bond creature and Island Bond creature are in the same mana range means that decks that don't run those colors will only have a selection of various similarly boring durdles in the 3-drop range.

Another problem in my case was that I was trying to fit in so many themes. I wanted a landwalking gimmick deck to be possible in draft. At some points I tried to fit Bog Tatters in there, but I already had too many clumsy high-power guys with similar sizes like 4/2, 4/3, 5/3.

One thing I felt lacking while looking at these cards was powerful-feeling creatures. Perhaps because of influence from M11, I felt that Black really wants to have creatures that reek of dark power (in terms of feeling). Cards like Nightwing Shade, Rotting Legion, Liliana's Specter, and Grave Digger are examples of that. Whispering Shade was like a compromise between a Shade and a landwalker, but I'm not sure it was as attractive as Nightwing Shade, and the need to have the clumsy Mass of Ghouls (as a vanilla creature that was different enough from the 4/2 and 3/2 in the color) wasn't helping.

Once I changed Vaportrail Imp to a 2/1 lifelink vampire that gains flying with an Island, and made Scorched Shambler into a 3B 4/2 that gains haste with Mountain, many of these problems started to solve themselves. I would make this change solely for the sake of making these cards play more interestingly anyways, but it was fortunate that it also worked out for the curve.

What would you put in the remaining slots?

For the 2-drop, I wanted a general-purpose one, preferably skewed towards control, since the other 2-drop was heavily skewed towards aggro. I also didn't want both 2-drops to have 2 power. Thirsty Bat was the perfect card for that, but I couldn't have that when there's another lifelink flying vampire in the skeleton.

Blood Seeker seemed too aggro, and would go in the same deck as Scourge of the High Seas. Bloodthrone Vampire was a good design, but I wanted an unconditional card for this slot rather than a card that only becomes really good in a particular archetype. (It also didn't seem like we had room to support a token-making archetype or a steal-and-sacrifice archetype).

Ravenous Rats seemed perfect as a non-aggro 2-drop, but I couldn't do that with Band of Theives in the set. As I discussed earlier, I'm trying to preserve and build around as many of the quirky features of the cards designed in the Design stage, and Band of Thieves was one of the cards that had more character to it.

I actually looked up every 2-drop black common ever printed on Gatherer, and was surprised to find that there was very little variety. (I was also surprised that there were a few 1/1s that turn into 3/1s while attacking!) So I had to make a hole-filling design, which I will discuss below.

For the 3-drop slot, I made a swampwalking creature for the landwalk deck. I also made it double as an Evil Presence. Evil Presence was in the set because of the land theme, but it shouldn't be because it just isn't a playable card.

The amalgamated card is probably too complex as a core set common though, but here I just wanted to experiment if messing with land with a death effect was fun. I felt at that time that if it turned out to be fun, we could try to figure out a way to work in a Swampifying death trigger.

For the 4-drop slot, I had a choice between Gravedigger and Drifting Shade. I chose the Shade because the other creatures all looked clumsy, and it seemed having it would be closer to the number of flyers Black usually gets. I also wanted as many cards as possible that could serve as potential answers to other powerful cards in the set, such as big flyers that might be put in the set.

I had a choice of using the more impressive 5-drop 2/2 Nightwing Shade instead of Drifting Shade. In that case, I could swap the 5-drop 5/3 vanilla with a 4-drop 5/1 vanilla. However, I felt that the 5/3 was important for trading with the 4-drop 4/4 Elephants which we were going to test with this file. The 5/1 could also trade with them, but the 5/3 seemed more main-deckable and we also had multiple cards in the file that hose 1-power creatures already, so that might prevent the 5/1 from trading with the 4/4. (Although maybe the lower cost 5/1s are better for stopping the Elephants?)

Once I chose the Shade to take the 4-drop slot, I wanted a Gravedigger variant somewhere else, so I made the 2-drop into one.

This card might be bad for variance, because it puts an effect for repetitive plays on a card that's very easy to include in your deck — it's good to draw in both in the early game and late game. A Disentomb, is a late pick because of the fact that you might draw it in your opening hand, but this one suffers no such problems. I later swapped it to a version that returns a random card from grave (including itself).

Now it provides a comical random mini-game where it might keep returning itself, which is sometimes what you want as a repeat blocker, sometimes what you don't want when you have a key creature you want to get back from the grave. But while I like how the card feels like an Imp that doesn't always do what you want, this slot could be any simpler card that fills a similar role, hopefully with no reminder text.

Finally, I wanted at least one of the high-end guys to feel powerful and dangerous, so I made the Band of Thieves into a 3/3 deathtouch creature. Deathtouch could make sense on a Thief since they tend to backstab and use poison in fantasy games.

Once again, one of the reasons I added deathtouch was because I felt it would help to have another creature that could trade with the 4-drop 4/4 Mastodon. Also, having removed the card advantageous Gravedigger, I wanted to make sure this card acted as a solid advantage for grindier decks. While I didn't want the entire set to be about card advantage, I wanted tight competition between aggro decks and control decks to occur. Finally, a 3/3 deathtouch would be able to attack into the many creatures with high toughness that were in the file.

On the whole, there are still too many complex cards for a core set, but many of them can be simplified, and with many of those complex cards I just wanted to try out if a particular theme was fun or not.

One thing that I would really like to do is add another creature, giving Black 11 common creatures. I would really like to add a 2/3 deathtouch creature at 3B. It would be like:

4-drop: 2/3 Deathtouch, 4/2 Haste
5-drop: Nightwing Shade, 5/3 Vanilla
6-drop: Cabal of Thieves as a 4/4 without deathtouch.

However, since we want to encouraging people to splash colors, it may be contrary to our purpose if we provide too many playables within each color.

Before I move on to the next color, I'd like to comment on the nature of this process. Creating a set configuration is extremely fun. I edit things in and I see potential strategies that can form. Ideas pop up that wouldn't come to me otherwise, as I'm constrained to think of filler designs such as "design a 2-drop gravedigger."

At the same time, it can be like an endless maze since everything is potentially a moving part, and the answer to every choice you make depends on the choices you made in other parts. There's no way to just go through each card one by one in alphabetical order and be done with it. You really have to try to identify what the main constraints are and identify any "either/or" choices that exist.

In the comments of this article Jay also posted another workable skeleton and he seems to have gotten it very quickly.

Finally, I described what choices I made, but I'm not a pro so my assessments on what works to affect what strategies might be wrong (such as a 5/3 vanilla being played to stop a 4/4 Elephant). But I hope that this shows some of the questions that should be considered while working on the skeleton.

I also wanted to show how various hole-filling designs would get added in during this process as a matter of necessity. They might look like they came out of nowhere, but they're not just flights of fancy on the part of the Developer (although that would be cool too). Rather, they could be the result of shifting one card to another mana cost like a Cadaver Imp being shifted to a 2-drop. It could be a way to handle powerful cards in other colors, such as the 4/4 Elephant. Or, it could be a way to build an archetype such as the "landwalk deck" off of the cards that already exist in the file.

I hope you enjoyed this, and I hope to get another post up tonight where I talk about what I did for White, which was actually the main thing I wanted to talk about.


  1. It's really interesting to see the process.

  2. The 5/3 is also more interesting than a 5/1 because it trumps the Elephant before it finds a Forest and it still just a 2/3. So that can create some dynamic board states.