Tuesday, March 27, 2012

Making Mythic Creatures

This draft has been moldering in the editing pile forever. I decided to pull it up again when I caught this comment from Metaghost in the dragons discussion: "...I think that the prominent thing this exercise has revealed to me is that part of why recent WotC dragons are boring is that the most resonant dragon tropes are boring. Or at least played out and stale." For some reason I was a little hurt by this. I think it's partly because I secretly love playing giant, terrible cards, especially dragons. Another factor may be that I'm a relentless fan boy, and I want to love everything Wizards does. But frankly, many of the mythics in the last two years (especially dragons) have been really disappointing. I think I assumed as a casual designer that making mythic creatures would be easy (surely they must be easier than red commons), but maybe it's worth reviewing what makes a mythic creature good.

When I look at any mythic card, there's always one factor that I consider first.

It should be awesome on its own.
A mythic creature should be able to display its full potential without depending on other creatures or the ability to play other cards. Consider Bloodlord of Vaasgoth. Bloodthirst is already a tricky ability for a mythic to get right, since you need some other card to turn on the ability. Then for the 'lord to work, it needs to survive a turn, have another vampire creature card to cast, then deal damage again for it to turn on bloodthirst for that vampire, then you need to successfully cast that vampire. This is just too many hoops to jump through. At the same casting cost as Bloodlord there's Skithiryx, the Blight Dragon, which is a design I really like. Skithiryx is an impressive clock (even without any other infect creatures in your deck), is resilient, has fun creature types to build around, and it can do stuff the turn it enters the battlefield. It also generates some spicy controversy -- a few players loathe infect, so a card like Skithiryx makes a good enemy. Bloodlord of Vaasgoth merely makes for a good snooze. I'm not sure how to emphasize this enough: a mythic creature must have an exciting pop to it. But mere awesomeness is not sufficient to make a good card. There are a couple other factors to consider.

Mythics need compelling flavor.

Mythics are one of the big selling features of a set, and as a result, all of the creative elements (name, art, and creative-mechanical connection) have to come together on each and every one. This puts them in a very different category from the other rarities, where sometimes you can gloss on card flavor to fix a design hole (like Crushing Vines), or meet a player expectation even when it doesn't quite fit the world setting (Balefire Dragon). I am starting to realize this is my major designer weakness (especially when it comes to names), so I'm not going to spend as much time on it, especially since I have a lot to say on the next topic.

Players should get play value out of their mythic.

This element is a little more complicated. Playing a mythic (especially in limited) should be a special event. This has several consequences. First, since opening a mythic is an unlikely event, players in limited need a fair opportunity to actually enjoy it. This means the card needs to be playable in limited, which imposes some important restrictions on acceptable casting costs. Obviously, not every sealed pool will accommodate a certain color, and sometimes drafts will take you away from your cool first pick. I also understand that for some cards, the sheer size is part of the appeal (like Blightsteel Colossus). But generally, if a card's colors (Phyrexian Obliterator), or its high converted mana cost (Iona, Shield of Emeria), make a card unplayable in limited, then that card is a wasted opportunity. The other consequence of the rule that players get value from their mythic is that the creature must have a reward that can't be nullified by the sorcery speed answers that are available in every color at common. So, mythic creatures need some way to generate value even if they die or are otherwise disabled. Before I get into ways to make that happen, a brief counterpoint to a common critique:

"Dies to Doom Blade" is probably the most common complaint about mythic creatures (ok, maybe second to, "OMG!!!1 So broken!!1), but this is the wrong test. Doom Blade is functionally similar to a counterspell like Mana Leak or Essence Scatter; the only difference is that removal and counterspells have different timing restrictions. So, by printing death-resistant creatures, you push constructed formats in ways that promote blue. It's a subtle problem, but it is an issue that has significantly skewed constructed play in the last two or three years. The goal should be to make a creature that a player can have fun playing, but that an opponent could still reasonably answer from any color (assuming they have enough cards in hand to answer it with).

On top of that, implementing the Doom Blade test gives you the wrong answer to some problems. For example, let's say you want to make an amped up Broodmate Dragon like Dragon Bro. If you focus just on instants, it seems like a fine card; a player lays Dragon Bro on the table, and even if the opponent kills half the player gets to keep a token. But if you think about available sorcery answers, Dragon Bros looks a lot worse. In constructed, it's easy to clear the board with Day of Judgement, and in limited, it still might fall prey to an Act of Treason (your opponent steals the original, and forces you to block with the token for a trade). A sweeper or stealer can be harsh against any creature, but it creates cases where you want to design a mythic creature that gives you value even against a sorcery-speed answer, and trying to impose the Doom Blade test gets you the wrong answer. Here are the options that are actually important when designing a creature so that it gives its owner value against sorcery speed answers:

It does something special the turn it enters the battlefield.

Strictly speaking, every creature does something on the battlefield, since it changes the board state and combat logic, but mythic creatures need to do more. And this isn't necessarily the same as having an "enters the battlefield" ability. For example, Avenger of Zendikar technically has an enters the battlefield trigger, but it doesn't actually do anything special the turn it comes into play (other than clog up the ground). Conversely, the New Phyrexia praetors make a major impact their first turn on the field, despite their lack of strict enter-play triggers. The biggest downside here is that since the M11 Titans were immensely popular and have been in circulation for a couple years now, these abilities seem stale. Still, I think there is a lot of open space here, and I think this is the best way to design a mythic.

It does something special when it dies.

I was expecting to see more of these in Innistrad and Dark Ascension. The change from "when CARDNAME is put into a graveyard from the battlefield" to simply "dies" frees up a lot of design space, and Innistrad is a death-focused block. But there still aren't many of these; the only mythic creatures so far with death triggers are Wurmcoil Engine and Child of Alara. (I see Vorapede and Mikaeus the Unhallowed as having resilience/resistance abilities.) Again, this doesn't absolutely refer to death triggers, it can be something like the "leaves play" template found on Sundering Titan. There could be other ways for you to use graveyard-focused mythics, like self-contained resurrection effects or abilities that only trigger from the graveyard (like a mythic Firemane Angel). Looking at the cards that have seen print so far, this is totally open space.

It has special resiliency.

This seems to be the direction R&D chose during recent block design with cards like Thrun, the Last Troll, Geist of Saint Traf, or Vorapede. Honestly, I am still uncertain about how to evaluate these. I would say generally that I worry about any kind of card that limits interaction as much as these do, and hexproof has not seemed like a fun mechanic. On the other hand, Thrun and St. Traf don't have much practical impact in limited, and they aren't breaking the current constructed meta. I feel like this is an opportunity to knock together a bunch of designs, and play them out to see if this is an area worth more exploration.

It's valuable in combat.

This category of abilities needs to balance the high risk that the owner will never get to use it with a high expected value if the creature lives a turn. A creature designed with this in mind shouldn't be able to win the game in a single hit, but it should create a dramatic swing in the game. The problem here is that the creatures built along these lines so far tend to create feel-bad moments for both the player and his or her opponent. Baneslayer Angel is a good example of this. Baneslayer is almost impossible to attack through profitably, and if she deals damage to the opponent, you probably win. Other times, your opponent rips a top-deck Pacifism or Day of Judgement, and your fancy angel is no better than a Siege Mastodon. These kinds of swing moments can make for memorable moments, but usually they are the wrong kind of memorable. I think this is the weakest of the acceptable methods to build a mythic.

It has a combination of these abilities.

This was referenced several times by R&D writers when discussing the titan cycle. Basically, when considering mythic creature abilities, instead of asking either/or, the answer is both. This answer has the same inherent problem as designing strong enters-the-battlefield abilities — the success of the titans has set the bar very high, while leaving a lot of players feeling jaded towards anything titanesque. On top of that, designing mythic creatures using this approach is unsustainable in the long run. I suspect that over time, combination platter style creatures will start to seem lame and tacky. Designing elegant abilities might be more difficult in the short term, but will yield greater rewards in the long run.

A few closing thoughts:

This started as a much longer piece, and I have cut it back down to a more focused essay of about 2000 words. I'm sure in the process I forgot some very relevant pieces, so as always, I'm curious to hear your feedback. I mentioned above that one of my biggest weakness is building strong, creative descriptions; my other big weakness is finishing what I start. I began writing this in June 2011, so it's taken about nine months to finish. I hope this is me turning over a new leaf -- there are several other posts languishing in our archive that I'd like to revisit in the near future.


  1. These are all interesting ideas about what makes a "good creature", but I'm not sure they're necessary for a "good mythic". After all, mythics don't need to be Spikey to be fun, resonant and enjoyable. I think those are the main criteria for mythics, not any necessity to be "tournament worthy" by having any immediate effect on the board.

    For example, Hero of Bladehold is a great mythic. It fails most of your criteria (Doomblade test, no immediate effect, not huge), but it's efficient, powerful attacker. And it tells an interesting story of being the hero leading an army into battle. It's also the most popular creature in standard according to the only reasonable measure of these things (http://gatherer.wizards.com/Pages/Search/Default.aspx?sort=rating-&action=advanced&type=%20[%22Creature%22]&format=|[%22Standard%22])

    The next most popular is Glissa, who also largely fails your tests.

    The least popular mythic creature EVER is Sphinx Sovereign. Big, fairly resilient, unique, does something the turn it enters the battlefield. And it's hardly unloved at 3.27. But still, it's ranks at the bottom of the bin.

    This isn't to say that these criteria aren't good: they define what makes for a "good" creature. But they don't account for the WOW factor that makes for a good mythic. I mean, none of the Mythic creatures verge on Archangel's Light reviled mythic status at a 1.465. Even the Bloodlord exists in the top half of the pack of Magic 2012 at 3.94.

    TLDR? Mythicness is in the eye of the beholder.

    1. I strongly believe that mythics need to have at least some appeal for each psychographic profile. It's fine for some to skew one way or another, but there needs to be a minimum level of appeal for everyone.

      I'm skeptical that the Gatherer ratings provide a reliable measure of popularity, mostly since they're not a randomly selected survey.

      Hero does work on plan D; if you get to swing with her, you're probably not going to lose.
      I was really excited when I first saw Glissa (and a little sad at the hero's downfall); but I haven't been able to fit her into any decks. So no, I don't think she was a successful card.

    2. Gatherer rating aren't ideal or anything near scientific. But they are just about the only polling on cards available anywhere.

      I think it's both interesting and important when thinking about game design to try and empathize with your audience. Cards that you might have never been able to fit into any decks might be beloved by some niche audience out there. So while there will always be some people out there take don't consider a card successful, it can be more useful to listen to the 200+ people that liked it.

      The only way I know of measuring a cards "casual appeal" is by their rating on Gatherer. I have no, just zero, idea who out there takes the time to actually vote on cards on gatherer. But they do, and it seems to reflect some of the decisions Wizards makes regarding reprints in products like FtV or Foil decks. That gives me hope that the rating look at least a little like the Godbook studies that Wizards uses.

      For reference, the quartiles of Gatherer ratings break down like this:
      4.960(Demonic Tutor YAY!)-4.046
      2.706-0.764(Viashino Skeleton BOO!)

      I mean, sure, ideally you want the Mythic creatures to be in that top bracket. And by and large, they all are. Out of the 145 mythic creatures (and creatures reprinted as mythic), 22 fall in the second bracket and 1 falls in the third (Sphinx Sovereign, what happened?). All 122 of the rest appealed to some subgroup out there somewhere. So, a batting average of about .841. It's a little off because of popular cards reprinted as mythic, but that's cut a bit both ways.

      In Standard, that's fallen to a .784. It's a pretty big drop, and you're right to notice it. But I'm not sure that these factors alone are enough to explain it. Vorapede, Falkenrath Aristocrat, heck, even Blightsteel Colossus meet some or most of your criteria but don't break through that top teir.

      I wonder why?

    3. Every Mythic should be modeled after Storm Crow.

    4. Maybe I should have said this first, but I'm pretty sure there's a big difference between cards that are popular, and cards that are well-designed. Otherwise, Demonic Tutor and Ancestral Recall would be some of the best designs ever.

      I agree that with mythics you want your designs to be popular, but I'm not not sure one necessarily implies the other.

      (I think that's the point Jay's trying to make with Storm Crow.)

    5. Storm Crow is popular because it's hilariously awful, has a in-joke name and sweet art. It's why Storm decks have dominated every format in which they're legal. It might not have been /intentionally/ well-designed, but in the context of the Magic sub-culture its spawned its pretty awesome.

      If you had a way to make every Mythic as popular, I'd be all for it.

      Demonic Tutor is WAY up there in terms of best designed cards, IMO. Top down flavor, innovative ability, promoted unique game play (tool box, combos) and a bit too powerful for its cost. It'd likely be one of my top choices for best designed card in the early years of Magic.

      Ancestral Recall is also insanely popular. It's one of the most powerful cards in Magic, it holds a legendary place in the Power 9, and its power is more easy to grok than Black Lotus or Moxen. It's also the only piece of power in the top 10, which speaks to the fact that Gatherer ratings aren't just a list of the most broken/expensive cards in the game. Heck, it's beaten by two commons, one of which is an Aura.

      Good designs are both popular and flavorful. Gatherer's pretty good at identifying that blend. They don't all have to be Ancestral Recall, but good Mythics should be looking like the top of this list: http://gatherer.wizards.com/Pages/Search/Default.aspx?sort=rating-&action=advanced&type=%20[creature]&format=[%22Vintage%22]

  2. Thanks for sharing your thoughts on this topic. Do you believe that more attention should be paid to mythic creatures being playable in limited than rare creatures, even though they will come up significantly less? Do you consider Iona and Obliterator to be mistakes?

    For me, I like to think if you're going to print a card that is particularly aimed at having a niche role in constructed, mythic can be an alright place for it (assuming it's awesome enough in its merit/temerity).

    And how much of this should apply to planeswalkers? Obviously you're likely to pick one out of the pack in draft for its value alone apart from your deck, and many of them are just amazing/excellent cards. But as printed planeswalkers get more and more niche, I have to imagine their limited playability going into steep decline.

    The other question I would ask would be, if we are to print humongous creatures, CMC=9+, what rarity *should* they be printed at? Is there a (mechanical) reason why it would be better to print that at rare than at mythic?

  3. Glad you revisited the ol' archives, Dan!

    While Duncan's correct that gatherer statistics would suggest the popularity of a given mythic has little to do with many of the metrics you've outlined, there's still value in reviewing these concepts and letting them guide you to a potentially superior design. Your ultimate point that "elegant abilities" are superior in the long run is a truism I'm sure MaRo would support, though I still hate Drogskol Reaver.

    One thing that hasn't been directly addressed in any of the recent mythic discussions is that the majority of modern limited formats have been fast, ranging from the blistering speed of ZZW to the "surprise, you're dead!" of triple Scars. This speed has distorted the non-casual player's perspective on many mythics, as they start to see Balefire Dragons wheel on a consistent basis. So, while I don't know where you stand in the spectrum of "Kitchen Player to MODO Grinder", it's been important for me to realize, as a designer, that the target audience for most mythics has no idea that Sphinx Sovereign is effectively unplayable.

  4. I'm glad you wrote this, Dan, since it is a subject we've been dancing around for a while on the site. I agree with your first two points, that a mythic should be awesome and that strong flavor is even more important than normal for a mythic, but I don't agree with your final point. A mythic needs to be splashy and exciting enough to help sell the set. It needs to feel more rare than a rare. It doesn't necessarily need to be strong, or playable in any specific format.

    There would be no legendary Eldrazi if they all had to be Limited playable. Limited—my favorite format by far—is not how most Magic games are played, so to say that every card of a rarity needs to be relevant in that format shows no heed for the many and various constructed formats. Most notably, hard-to-cast mythics have a home in EDH where 'expensive' is just another word for 'playable'.

    That we would assign this playability requirement to the mythic rarity and no other is another concept I take issue with. There are rares that are good in Limited but bad in Constructed and vice-versa. The same is true of uncommon and common. What about a rare that we see only a third as often makes it something we should guarantee is always relevant to your deck? If anything, the reduced frequency with which players will open a mythic in Limited gives us more room to try concepts that only work in unusual situations.

    1. I'm not saying every mythic critter has to be capped at 7-8cc (though I think it's a fair rule of thumb for core sets).

      But I think the legendary Eldrazi support my point. Rise of the Eldrazi was a slow, slow limited format (especially compared to ZZW), and if you put together a defender ramp deck, there was a fair chance to actually play Kozilek or Ulamog (Emrakul was still pretty impractical, but it was sort of the definition of "Awesomeness defined by size.")