General Defining & Optimizing Complexity

landofMordor

Administrator
What does a successful game share in common with a catchy song, an arresting painting, or a delicious gourmet dish?

Hm, that's an ill-defined question. Let's try another: Which of the four following cards is most appealing to you?

Shock One {R}: Instant -- Deal 2 damage to target planeswalker. Flavor text
Shock Two {R}: Instant -- Deal 2 damage to any target. Flavor text
Shock Three {R}: Instant -- Kicker {4}; Deal 2 to any target; if it was kicked deal 4 instead. Flavor text
Shock Four {R}: Sorcery -- Flash; Kicker {W} and/or {R}; Select target opponent and up to one other target. That opponent mills 20 cards, then gains 6 life, then puts the top 20 cards of their graveyard on top of their library. If you chose a second target, that opponent loses 6 life and you deal 2 damage to that target; otherwise, that opponent loses 4 life. Note all untapped permanents you control, then tap and untap them. Empty your mana pool. Shuffle all the cards in your library that you haven't looked at this game. Put up to three cards from outside the game into your graveyard, then phase them out until the game ends. Exile all World Enchantments. Flavor text

... Let's talk complexity.

What makes a decision interesting?​


A game is "a voluntary attempt to overcome unnecessary obstacles" (Suits' definition). Unlike other areas of human life where we are faced with unnecessary obstacles (e.g. participating in the democratic process, or flying internationally), a game is fun, and no small part of that comes from the interesting decisions we make during gameplay.

Sid Meier has done a decent job of summarizing what an interesting decision entails:
- Player has sufficient information about the choice and its effects
- Decisions have persistence, cost-benefit tradeoffs, or clear causality
- Decisions incentivize players to choose differently under different circumstances

To borrow from another philosopher of games, C. Thi Nguyen: games are the art form of agency. Games distill human agency, the act of doing, in the same way that a painting distills the act of seeing or a song the act of hearing. Just like the subtle, complex textures of Van Gogh's impasto technique lend his "Sunflowers" a heightened visual weight and grandeur, complexity in a game can yield a more compelling agential landscape.

Not all complexity is created equal.​


There are several kinds of complexity, some of which I owe to a LPR episode (which itself derives from MaRo, as all Magic theory does ;)):

- Visual complexity (difficulty for the brain to parse what it's seeing, be it rules text or graphic elements)
- Comprehension complexity (difficulty to correctly use a game piece)
- Interaction complexity (MaRo calls this "board complexity")
- Tracking complexity (difficulty of mental math, assessing the game state, etc)
- Strategic complexity (difficulty of making strategically sound choices)
- Metagame complexity (amount of meta-game needs like: tokens, dice, Companion-like deckbuilding restrictions, dexterity requirements, etc)

These types of complexity are all somewhat convoluted -- to return to my toy examples at the top, "Shock Four" offers a lot more strategic complexity than the alternatives, but also has more of all the other kinds of complexity! Even weirder, these latter complexities are partly what generate the strategic complexity in the first place! Complexity is extremely... well, complex, but it's still possible to sketch the contours of the tool using a few heuristics.

Heuristic 1: Complexity facilitates and hinders interesting decisions​

Complexity exists to permit more interesting decisions. "Shock Two" above usually offers more decisions than "Shock One" -- its rules text will interact with more cards, and the card itself offers a larger decision tree. This is a primary reason why complex cards exist. Complex cards just do more stuff, which will create emergent interesting decisions without needing to be designed explicitly. A related use of complexity is to clearly establish the decision space of the players (e.g., unambiguous rules/reminder text, a systematic rule set, or game pieces like tokens & counters to track stat changes). This kind of complexity aids in the clarity of rules, adding to players' decision-making capabilities.

However, if the amount of interesting decisions remains identical between two designs, complexity cannot add mechanical value. If you only have one land in play, there is no longer a meaningful mechanical difference between "Shock Two" and "Shock Three". Though this example is highly specific and somewhat unlikely, it illustrates two further properties of ludic complexity: 1) the additional decision-making that a complex game piece offers over a simpler one is contextual, and 2) if two designs offer identical decisions, additional complexity can't be a value-add. I'll return to the former point in much greater detail later on, but the latter property begs the question -- what's the harm in complexity, even if it adds only marginal value?

Complexity that adds negligible value should generally be avoided, because complexity can spoil decisions which are usually fun. As a toy example, consider why Wizards doesn't print reminder text on every French Vanilla they print -- it adds visual and comprehension complexity that isn't necessary for gameplay. As another example, I loved Skaab Wrangler's play patterns in MID along with decayed Zombie tokens, but occasionally I found myself with too many Zombies and the decisions became drudgery. "Shock Four" suffers from this same over-saturated decision tree to resolve what is usually "just" 2 damage. Eventually, one's appetite for making an initially interesting decision is satiated, and the novelty fades into tax-season tedium. To return to C. Thi Nguyen's definition of games as the "medium of agency," this "complexity satiation" is the same kind of tipping point as when an overplayed pop song's catchy hook starts to annoy the listener, or when the third slice of pizza isn't as satisfying as the first two.

Even though complexity adds a lot of fun to art (including games), there is value in moderation.

Heuristic 2: Complexity depends on audience.​

Just like two people will have different tolerances for prog metal or Impressionist paintings, different players will have unique thresholds for complexity in a game.

Much of this threshold depends strongly on the player's familiarity with Magic. For greenhorns, even cards with full reminder text prove difficult to comprehend, much less play properly. And even with a decade of Magic under my belt, I still get fatigued more quickly during my first draft of a new set, since I have to pause and read every card slowly to avoid misplays. (It's also why I rarely accept invitations to fully-custom cubes, chaos drafts, or cubes with a high density of illegible old-border {W} cards.) Repetition breeds familiarity. Uniqueness breeds complexity.

Even among players with equal exposure to Magic's rules and game pieces, their motivations will shape the complexities they find most interesting. My friend who built a "Dinos Driving Cars" commander deck may be less interested in comprehending or tracking my game pieces than fiddling with Crew costs on his own playmat. Tammy may be more interested in growing the stats of their 100/100 creature; Jenny may prefer the interaction complexity of a build-around; Spike may desire the strategic complexity of a hard-to-master interactive card; Vorthos may draft with stipulations to pair Tourach with his Hymn (true story!). Unless you know your group very well, assume that you'll have a wide set of motivations among your players.

Luckily, many well-established Cube design heuristics already accommodate a lot of this diversity: divergent macro- and micro-archetypes that will naturally have different decision trees, build-arounds and bombs to widen or simplify decision trees; deck customization through Draft... Most importantly, the cube's drafters have agency to opt in to the style of deck that appeals most to them. But, when practical, Cube designers can also incorporate complexity into their cube design process.

Heuristic 3: Cube complexity is a design resource.​

Selecting a card for a cube is a balancing act between (at least) flavor, mechanics, power, and all the kinds of complexity. This is a very context-sensitive process, so to illustrate, I'll choose some case studies from my main cube (whose names definitely did not describe my mental state while writing this article).

Terror: I still run Terror over several strictly-stronger options. Terror comes with a nonzero cost to complexity -- it has extraneous rules text which tracks card attributes (I don't run any regeneration, and run very few artifact creatures). I don't care, because I adore Adam Rex's 10th Edition art, such that a decision to pick this card is made more "interesting" by the aesthetic pleasure of the object itself. Moreover, Terror has a readily available heuristic in my audience's minds, which shortcuts the comprehension complexity of the game piece. I don't think I can completely saturate my format with meaningless lines of rules text like Terror's, but as an exception, it's a-okay.

Fury: This Elemental comes with a lot of strategic complexity. Do I Evoke? if so, what card to pitch? what things to target? There's some rules complexity, too, as Fury doesn't have reminder text for its fairly rare keyword. And Fury may interact in new ways with a huge variety of other cards. However, the strategic and interaction complexity map neatly onto the traits identified by Sid Meier -- Fury's effect is high-agency, persistent, dynamic, and costly -- which is exactly what I want from a power outlier in my Red section. Fury's complexity is precisely the benefit of including it! (As an aside, the card's high rate of Constructed play will do a lot to minimize the comprehension complexity.)

Duress: I'm currently on two versions of Duress, the Urza's Saga version and the Japanese alt-art from the Mystical Archive. The effect is grokkable and simple, relatively speaking, and the card is very well-known. My English copy alleviates some burden on the Japanese one. And the strategic complexity of Duress is really nice, even compared to other hand-hate effects: the "noncreature" rider limits any analysis paralysis, the choice I make will depend dynamically on the game state, and the choice is persistent. There's one other advantage my JPN Duress has: running a second copy of it means I don't need to run the much-less-elegant Dread Fugue. Fugue offers some real mechanical upside, but it's lesser-known, and so that mechanical upside has to be weighed against the cost of the additional complexity. In my case, in a cube where many important spells sit at 3 MV and/or are noncreature, Fugue doesn't offer enough upside to be worth removing a well-known, perfectly functional, simple design like Duress.

At the risk of being too reductive, here's what I learned in this exercise:
- All the types of complexity can matter.
- Repetition breeds familiarity. Uniqueness breeds complexity.
- For complex designs, find what they offer to your design over against alternatives.
- If the marginal utility of a complex design is small in your context, then the most salient feature of the card is its complexity, not its utility.
- Non-singleton designs have a natural complexity advantage over singleton cubes of the same size.
- Hidden or one-time complexity is more beginner-friendly than the alternatives.



Please post below with thoughts and feedback! Are there types of complexity that I missed? What's the tipping point for your complexity threshold? What cards are worth the complexity in your cube formats? Thanks for reading, and cheers.
 
Last edited:
Complexity in all of its forms is one of my favourite things to discuss! The goal of a cube it probably the main driver for decisions about complexity: currently I have three cubes, with very different goals and therefor very different relations to (different kinds of) complexity. This is my interpretation and implementation of complexity, with some of my own observations:

Visual complexity on a very basic graphical level is very important to me; I don't like to include old-border or alternate frames in my cubes at all. This includes frame aspects like watermarks (which I'd love to avoid) or the legendary crown (which I'd buy a card's new version for). It also creates a graphical unity that I personally really like.

Interaction complexity might be the only type you mentioned I'm not actively considering. I consider this a kind of complexity that can be conquered by skill and knowledge of cards. The target group for which you need to aim for lower interaction complexity are players new to TCGs, to avoid creating information overload.



My Jeskai Cube
goal: provide a boardgame-like experience for 2-4 players familiar with Magic rules (on any level).



Comprehension complexity (aimed low)
A few of my friends know Magic's rules and play enough Hearthstone to appreciate en enjoy Magic's gameplay, but just don't want to buy into it. My cube's comprehension complexity is aimed to provide an enjoyable experience for them. I'm limiting the amount of rules text and keywords in the cube with a lot of returning mechanisms and themes, as if it was a premier draft set. For example, while Adventure might not be the simplest mechanism, it is present on 14/180 cards (an asfan > 1) and becomes faster to comprehend for every instance you encounter.

Trinket text (and flavor words) is another form of comprehension complexity I try to avoid. The most egregious remaining example in my cube is Daretti, Scrap Savant's "this card can be your commander". Two other examples are Ethereal Investigator (although the cube can be played multiplayer it's not the main aim) and Teshar, Ancestor's Apostle (it mentions Saga's, which aren't in the cube (yet)). I'm afraid this could create confusion for players not as familiar with the ins and outs of Magic as we are ("what is a commander?")

Word count can be comprehension complexity, although that relationship isn't liniar: Surely Badgersaur is a far simpler card than it's word count would suggest, as it boils down to "whenever you discard a card, do a thing". Especially in draft you don't have to comprehend the entire card before you know whether it's relevant for your deck or not. Godhead of Awe might actually be more difficult to understand, especially with stat-altering effects.

Tracking complexity (aimed low)
For tracking complexity I focus on tokens and counters. I want to see in one glance at the battlefield what the situations is. For counters, that means that each permanent type can only obtain 1 kind of counter, which is especially relevant for creatures: no -1/-1 counters, no ability counters, no shield counters. Everything is trackable with a single die.

For tokens, it means that I severely limit the kinds of tokens that the cards in this cube can produce. A 1/1 Servo artifact token is interchangable with a 1/1 Myr artifact token, although no cards are should care about (a token's) creature types. For Master Splicer and its kin, this means preventing the presence of another kind of 3/3 artifact token.



My 1-life Cube
goal: provide a 2-player draft in 15 minutes, portable in an 80-card deckbox.



Metagame complexity (aimed low)
Because the goal of this cube is to be easily portable, I don't use cards that need/produce tokens or counters. The only exceptions are Sacret Cat (which can use itself as a token) and Sentinel Dispatch (which is sleeved in a transparant sleeve with the token on its backside). This entire cube is build around this complexity, although I never realised it is a for of that.

Strategic complexity (aimed high)
The target group for this cube is opposite of that of the Jeskai Cube. This means that comprehension complexity is completely thrown out the window and this cube is made as strategic as possible. This can best be seen in Basking Rootwalla and Grisly Sigil, both of which have a huge word counts with lots of room for comprehension mistakes, but also a lot of strategic depth. This cube might prefer the Shock 4 variant over the others.

Tracking complexity (aimed high)
The size of this cube is 54 cards, which are grid-drafted in 6 grids. This means that each player has a very good idea of the (max 18) cards in their opponent's card pool. Cabal Therapy is an example of using this trackable information to your advantage, as you have a very decent chance of guessing a correct card in the opponent's hand if you've paid attention. This goes further of course, as you could also be aware of all possible answers and threats in your opponent's hand at all time. This is a feature, not a flaw.



My Horde Module
goal: provide a 1 player option for any cube.



Tracking complexity (aimed low)
Tracking a second deck/player is engrained in the whole gimmick of this cube. In order to make it fun for the player, that tracking has te be intuitive and fast. The biggest challenge is creating a rule set that facilitates this easy tracking, but that's a little bit outside the scope of this discussion. However, a related factor is that the Horde should never have to make choices: the cards in this cube hand over all agency to the actual player through effects like Dredge the Mire.

The simplicity of cards for the Horde was a second way to actually lower tracking complexity. All cards should be as simple as possible to avoid confusion or mistakes on the part of the player, so a card like Cower in Fear is preferrable over Festering March: even if the latter provides more interesting gameplay, it doesn't make the game playable.



It's a fun exercise applying the different kinds of complexities to your cube's design goals. Reading back, I wonder how much of my ramblings actually foster discussion, but there you are :)
 
Great read and I very much appreciate the condensed bullet points to drive home the take aways.

- For complex designs, find what they offer to your design over against alternatives
- If the marginal utility of a complex design is small in your context, then the most salient feature of the card is its complexity, not its utility
These two points will help me take a critical look at my cube and see if the most complex designs are worth it or not. Thanks!
 
Regarding card choices I feel like "visual complexity" is one of the hardest things to avoid sometimes.
Only yesterday I faced the situation of - even one of my regular drafters - having to re-read Moon-Circuit Hacker. The card has just a lot more going on but is at the same time much more complex (even ignoring the typing). That being said, once you understood what it does it is really not that complex a card. It can be explained in one sentence despite the size of the textbox.
It's just the initial burden that makes the card a lot harder to parse than its older brother. For most cubes the Hacker would be the far superior choice, with it's cheaper cost and better floor but it is a much longer read than its predecessor. I guess the question one has to ask is: does the playability outweigh the increase in complexity.
 
Last edited:
Hacker vs. Deep Hours is also an interesting one because Ninja of the Deep Hours is actually simpler than it initially looks if you're an enfranchised player, since it falls under the general umbrella of "card you've seen many times before, except it has a set mechanic now". Moon-Circuit Hacker is a more unique design, so it ends up causing more cognitive load despite being a very similar card.

I think it's also important to consider that not all complexity is created equal. Targeting restrictions are a good example — Power Word Kill ultimately has a more complicated restriction than Doom Blade, purely because checking for specific creature types is harder than checking a card's color.

...

You know, I just thought of a better example of a "complexity scale" than the Shock one (which has a ridiculously massive jump in complexity between Shock Three and Shock Four):



EDIT: I actually think looking at why Light Up The Night is so complicated compared to the other two cards might be useful.

1) It's wordy as hell. Blaze is one line of text, Devil's Play is technically two, and LUTN is six lines of text. Ugh.

2) The basic effect is conditional. With Blaze and Devil's Play, you dump X mana into the spell and it deals X damage, regardless of what you're targeting. LUTN does that... and then goes "hey, add 1 damage if you aren't aiming this at a player!" It's easy for one player or the other to forget about that over the course of a game.

3) Flashback itself is a cognitive load, since you need to remember that you have a card you can cast in your graveyard, which might not actually be visible. It's entirely possible to forget that you cast a card with Flashback earlier in the game.

4) Even with that in mind, though, LUT'N one-ups Devil's Play by having a completely new way of determining X when you flash it back. Suddenly you need to care about how many counters you have on a certain type of permanent instead of just how much mana you have available.
 
Last edited:

landofMordor

Administrator
Some quick responses --
I just want to say, that all magic theory derives from Richard Garfield lol. But MaRo is doing most of the modern theorizing.
Poe's Law strikes again! (I made that statement jokingly.)
Fun fact: Gaddock Teeg doesn't actually affect Shock Three, since kicker doesn't change the mana value.
Touche.
Potential Shock Three replacement that both offers more decisions and complexity than shock 2 while being affected by Teeg:
It was simply an illustration of the principle. The actual choice of card shouldn't matter as long as you get the point I'm trying to make. If you like, substitute "if you only have 4 lands" instead of "if Gaddock Teeg is in play".
Great read and I very much appreciate the condensed bullet points to drive home the take aways.


These two points will help me take a critical look at my cube and see if the most complex designs are worth it or not. Thanks!
Thanks! The only editors on my RTL posts are me, myself, and I, so I try to inlcude the tl;dr to counteract those loonies' innate tendency to talk too much :)

I'll circle back later today with some more responses! Thanks for the thoughts, all!
 

landofMordor

Administrator
It's a fun exercise applying the different kinds of complexities to your cube's design goals. Reading back, I wonder how much of my ramblings actually foster discussion, but there you are :)
Thanks for your reply! I'm glad you enjoyed the essay. There were several neat insights I gleaned from your response:
watermarks (which I'd love to avoid) or the legendary crown (which I'd buy a card's new version for)
I'd never even considered that watermarks or legendary crowns could be considered a type of complexity! But you're definitely right that they create visual noise. The difference you imply here is that the legendary crown correlates to a mechanic (kinda like the artifact-creature frame), whereas watermarks are "purely" decorative. This reminds me of my "all-terrain battlebox", where I faced a similar choice: whether to limit non-evergreen mechanics, flavor text vanillas, frame variance, etc. Here, I actively wanted some small amount of visual noise to spark my players' curiosity about Magic. Questions like "Wow, all of these cards have text from 'The Theriad'! Are there more like this?" or "Woah, this frame is so different! What's the story with the mana cost over on the left like this?" or "What's this demon-face watermark?" are definitely successes for this environment.
The target group for which you need to aim for lower interaction complexity are players new to TCGs, to avoid creating information overload.
Good point -- I think I agree, but it also applies to a playgroup that has a large rotation of games -- if Magic is competing against thinky boardgames like the excellent Inis or Terra Mystica for a playgroup, then interaction simplicity can be an advantage.
Trinket text (and flavor words) is another form of comprehension complexity I try to avoid.... Teshar, Ancestor's Apostle (it mentions Saga's, which aren't in the cube (yet)). I'm afraid this could create confusion for players not as familiar with the ins and outs of Magic as we are ("what is a commander?")
I agree that there's risk here, but returning to my battlebox, I realized that Veteran Swordsmith's Soldier tribal, though mostly-irrelevant in my format, implies greater depth to the game that can be really attractive. It's fun to feel like there's more to discover.
Word count can be comprehension complexity, although that relationship isn't linear: Surely Badgersaur is a far simpler card than it's word count would suggest, as it boils down to "whenever you discard a card, do a thing". Especially in draft you don't have to comprehend the entire card before you know whether it's relevant for your deck or not. Godhead of Awe might actually be more difficult to understand, especially with stat-altering effects.
Excellent point. And conversely, there are other cards, like Dead Ringers, where a low word count leads to immense complexity. In general, I think the two are correlated, but there's not a strict causal relationship.
This cube might prefer the Shock 4 variant over the others.
Definitely a good reminder that complexity is contextual!
Tracking complexity (aimed high)
The size of this cube is 54 cards, which are grid-drafted in 6 grids. This means that each player has a very good idea of the (max 18) cards in their opponent's card pool. Cabal Therapy is an example of using this trackable information to your advantage, as you have a very decent chance of guessing a correct card in the opponent's hand if you've paid attention. This goes further of course, as you could also be aware of all possible answers and threats in your opponent's hand at all time. This is a feature, not a flaw.
This tracking of the opponent's probable hand to generate strategic information is an aspect of complexity I overlooked, and is a great addition. Definitely not right for every cube, but the cubes that do want this type of complexity are likely to be defined by it in very novel ways.

Thanks again for the response! :)
 

landofMordor

Administrator
Regarding card choices I feel like "visual complexity" is one of the hardest things to avoid sometimes.
Only yesterday I faced the situation of - even one of my regular drafters - having to re-read Moon-Circuit Hacker. The card has just a lot more going on but is at the same time much more complex (even ignoring the typing). That being said, once you understood what it does it is really not that complex a card. It can be explained in one sentence despite the size of the textbox.
It's just the initial burden that makes the card a lot harder to parse than its older brother. For most cubes the Hacker would be the far superior choice, with it's cheaper cost and better floor but it is a much longer read than its predecessor. I guess the question one has to ask is: does the playability outweigh the increase in complexity.
There's also several additional open-ended interactions on the Hacker. Enchantment type, discard trigger. If those synergies matter, then they might outweigh the additional complexity of the first read. I totally agree that this question is key to assessing complexity.
You know, I just thought of a better example of a "complexity scale" than the Shock one (which has a ridiculously massive jump in complexity between Shock Three and Shock Four):
I respectfully disagree about the adjective "better" :) My goal for Shock Four was to illustrate that it's trivially true that complexity is not an unmitigated good. I also had a "Shock Zero" at one point, whose sole rules text was "Choose a target", to illustrate the trivial truth that complexity is simultaneously required to generate decisionmaking. However, I really appreciate the complementary insight of your Blaze comparisons!


EDIT: I actually think looking at why Light Up The Night is so complicated compared to the other two cards might be useful.

1) It's wordy as hell. Blaze is one line of text, Devil's Play is technically two, and LUTN is six lines of text. Ugh.

2) The basic effect is conditional. With Blaze and Devil's Play, you dump X mana into the spell and it deals X damage, regardless of what you're targeting. LUTN does that... and then goes "hey, add 1 damage if you aren't aiming this at a player!" It's easy for one player or the other to forget about that over the course of a game.

3) Flashback itself is a cognitive load, since you need to remember that you have a card you can cast in your graveyard, which might not actually be visible. It's entirely possible to forget that you cast a card with Flashback earlier in the game.

4) Even with that in mind, though, LUT'N one-ups Devil's Play by having a completely new way of determining X when you flash it back. Suddenly you need to care about how many counters you have on a certain type of permanent instead of just how much mana you have available.
The cognitive load of tracking other zones is a component of complexity that I missed, as is the complexity of doing mental math to determine X (not just on Blaze variants, but also everything from Flunk to Tarmogoyf). Your succinct breakdown, especially #3-4, really draws for me the line between an elegant design (Devil's Play) and an inelegant one. Really cool stuff.
 
I feel like the jump in complexity between Shock Three and Shock Four is undercut by the fact that Shock Four is so needlessly complicated that you get distracted trying to figure out what it actually does, to the point where I mentally checked out for a bit. :p I also feel like showing people actual cards is helpful, since no duh is Shock Four too complicated... but is Light Up The Night?

Another fun comparison I just thought of:



It might not be immediately obvious, but Saw It Coming is way more complicated than Admiral's Order. I'll write out what they do, and bold the added bits of complexity:

Admiral's Order - 1UU
Counter target spell. If you attacked this turn, you can cast this spell for U instead.

Saw It Coming - 1UU
Counter target spell.
Instead of casting this spell, you can pay 2 and place it face-down. If you've done this, you can cast it for 1U instead as long as it isn't the turn you placed it face-down.
 
Thanks for your reply! I'm glad you enjoyed the essay. There were several neat insights I gleaned from your response:

I'd never even considered that watermarks or legendary crowns could be considered a type of complexity! But you're definitely right that they create visual noise. The difference you imply here is that the legendary crown correlates to a mechanic (kinda like the artifact-creature frame), whereas watermarks are "purely" decorative. This reminds me of my "all-terrain battlebox", where I faced a similar choice: whether to limit non-evergreen mechanics, flavor text vanillas, frame variance, etc. Here, I actively wanted some small amount of visual noise to spark my players' curiosity about Magic. Questions like "Wow, all of these cards have text from 'The Theriad'! Are there more like this?" or "Woah, this frame is so different! What's the story with the mana cost over on the left like this?" or "What's this demon-face watermark?" are definitely successes for this environment.

Good point -- I think I agree, but it also applies to a playgroup that has a large rotation of games -- if Magic is competing against thinky boardgames like the excellent Inis or Terra Mystica for a playgroup, then interaction simplicity can be an advantage.

I agree that there's risk here, but returning to my battlebox, I realized that Veteran Swordsmith's Soldier tribal, though mostly-irrelevant in my format, implies greater depth to the game that can be really attractive. It's fun to feel like there's more to discover.

Excellent point. And conversely, there are other cards, like Dead Ringers, where a low word count leads to immense complexity. In general, I think the two are correlated, but there's not a strict causal relationship.

Definitely a good reminder that complexity is contextual!

This tracking of the opponent's probable hand to generate strategic information is an aspect of complexity I overlooked, and is a great addition. Definitely not right for every cube, but the cubes that do want this type of complexity are likely to be defined by it in very novel ways.

Thanks again for the response! :)
Please, do not ever mention Dead ringers again. How it is worded is an abomination: slightly better would have been: destroy two target non-black creatures that are the same (or share the same) colours. They cannot be regenerated.
Better yet, forget it ever existed because even if you understand the card, it is weak, especially in invasion block which has way to many two for ones as it is.
 
It
... the excellent Inis ...
Good to find another Inis fan here. Possibly my favourite board game, let down only by the fact that it seems to be ludological marmite(1) in my game group. As well as the drafting each round that might be popular with magic players, it creates a rich multiplayer experience that Commander can only aspire to.

There are two areas of complexity that irritate me the most as a cube designer. The first is the variation in templating for similar effects from different eras. For example “Removed from game” and “exiled” mean the same, but create unnecessary inconsistency. The second is where similar effects function slightly differentl, for example immediate and delayed blinking, making it harder to learn useful heuristics about how cards function.

(1) a savoury spread produced as a byproduct of the brewing industry in Britain, commonly used as a metaphor for something either loved or hated.
 
+1 to Inis

-1 to marmite


Actually, I think that what Humpty_Dumpty mentioned is actually a larger type of complexity: interactive complexity. This complexity doesn't result from one card but rather from having multiple similar cards within a larger set that have similar yet mechanically different abilities.

ex. this set of four cards are all individually very easy to parse, yet their combination starts to get rough



These could all realistically show up in the same deck and play just fine individually, but the Venn diagram (okay, it's no longer a Venn diagram, yes, but work with me here) of what abilities apply to what get hard to track. While this complexity isn't that important when you're just looking at single cards, it's a crucial element of the layer cake that is a synergistic cube. Too many strings on a card and you overwhelm your drafter--too few and it's not as flavorful as you may like.
 
Last edited:

landofMordor

Administrator
I also feel like showing people actual cards is helpful, since no duh is Shock Four too complicated
I agree. But I do think it's useful to show that there is some limit where it's trivially obvious that complexity is unhelpful. It might be overkill for you, but if you're a person who happens to really love Light Up The Night, that point wouldn't come across.

But, more to the point, I really like your counterspell example, too! Cancel itself could fit somewhere on that spectrum. Actually, comparing all 3 suggests one more design goal of a complex card: complexity can be used to tune power level. Cancel is too weak for many formats, including often Retail Limited, so they add a exile clause to get Dissipate, which hits at about the right power level. Additional complexity, sure -- but a simpler version maybe would have been irrelevant on power level in a format like INN or MID.

Please, do not ever mention Dead ringers again.
lol consider it done!

Good to find another Inis fan here. Possibly my favourite board game... As well as the drafting each round that might be popular with magic players, it creates a rich multiplayer experience that Commander can only aspire to.
It's so good! The more I play that game, the less likely I am to ever settle for Commander. If only I could get Kemet to the table as often...!

Actually, I think that what Humpty_Dumpty mentioned is actually a larger type of complexity: interactive complexity. This complexity doesn't result from one card but rather from having multiple similar cards within a larger set that have similar yet mechanically different abilities.
Hm yeah or maybe like "heuristic-resistant" complexity. I once overindulged this type of complexity when I went up to 8 or 9 dual land cycles, including Llorwyn filterlands, checklands, creaturelands, shocklands, fetchlands, horizonlands, and painlands. Eventually it felt like governmental bureaucracy just to see which spells I could cast. It was awful, and I went to 3 fetch/3 ABUR dual/2 assorted without looking back. I felt similarly when playing with 7 Cryptic Spires in a 4-color 2X2 draft deck, and it was exacerbated by the super vague grey/black border and text box on that land that didn't give graphic clues as to the land's color.

Your example of token types mattering is also very good. If I could add three more cards to that list, it'd be Usher of the Fallen, Elspeth, Sun's Nemesis, and Champion of the Parish. Good luck keeping all those straight, haha!
 
I'd never even considered that watermarks or legendary crowns could be considered a type of complexity! But you're definitely right that they create visual noise. The difference you imply here is that the legendary crown correlates to a mechanic (kinda like the artifact-creature frame), whereas watermarks are "purely" decorative.

You and I know this, but for new players (or returning players) that might not be obvious. I'm actually unsure how they perceive this visual noise, but once they learn the legendary crown has a mechanical implication, they might be looking for one in watermarks as well (and sometimes watermarks have those, like with the Foretell watermark).

So yeah, that's why I'd like to include one but not the other. It's a very minute form of complexity though, and is just as important for me for aesthetical reasons as for mechanical ones :)
 

landofMordor

Administrator
This morning I've been bingeing Zach Barash's "Mechanics In Review" series on Hipsters of the Coast. Zach brings up some great nuances r.e. hiding complexity, which I'll paraphrase here:

Twinshot Sniper and Mirrorshell Crab are complex for their rarities. Moreover, the heuristics that apply to one don't apply unilaterally (Sniper has quasi-Evoke, but Crab does not). However, these cards are much less burdensome than equivalently wordy designs. Much of their complexity is only present before the spell is cast/activated, yields a temporary and predictable effect, and their complexity is also in a hidden game zone. New players don't even have to think about their opponent's possible Channel abilities, and they only have to think about their own until the cards leave their hand.

Another way to hide complexity is to remove player agency. A NEO Saga like Kumano Faces Kakkazan has a lot of words on it, but the player gets very little agency in how the card resolves, and they don't even need to fully grok what Chapters 2-3 do at the time they put it on the stack. The Saga's rules just chug along and resolve each line of text one at a time with minimal additional strategizing from the player. Contrast this to a card like Plargg, Dean of Chaos who has two wordy cards that must both be conceptualized prior to casting it, and whose abilities require strategizing and state-checks every time.

And on a minor note, for tracking and interaction complexity, it's often easier to make a one-time check of the game state (Cryptic Serpent), instead of a persistent check (Tarmogoyf).

I don't think my initial essay did enough to address that complexity's downsides can be attenuated using other design tools. I'm glad I ran across Zach's articles to help provide this perspective. They're certainly worth a lunch break of reading time!
 
This morning I've been bingeing Zach Barash's "Mechanics In Review" series on Hipsters of the Coast. Zach brings up some great nuances r.e. hiding complexity, which I'll paraphrase here:

Twinshot Sniper and Mirrorshell Crab are complex for their rarities. Moreover, the heuristics that apply to one don't apply unilaterally (Sniper has quasi-Evoke, but Crab does not). However, these cards are much less burdensome than equivalently wordy designs. Much of their complexity is only present before the spell is cast/activated, yields a temporary and predictable effect, and their complexity is also in a hidden game zone. New players don't even have to think about their opponent's possible Channel abilities, and they only have to think about their own until the cards leave their hand.

Another way to hide complexity is to remove player agency. A NEO Saga like Kumano Faces Kakkazan has a lot of words on it, but the player gets very little agency in how the card resolves, and they don't even need to fully grok what Chapters 2-3 do at the time they put it on the stack. The Saga's rules just chug along and resolve each line of text one at a time with minimal additional strategizing from the player. Contrast this to a card like Plargg, Dean of Chaos who has two wordy cards that must both be conceptualized prior to casting it, and whose abilities require strategizing and state-checks every time.

And on a minor note, for tracking and interaction complexity, it's often easier to make a one-time check of the game state (Cryptic Serpent), instead of a persistent check (Tarmogoyf).

I don't think my initial essay did enough to address that complexity's downsides can be attenuated using other design tools. I'm glad I ran across Zach's articles to help provide this perspective. They're certainly worth a lunch break of reading time!
There is also something called complexity for the opponent. The saga commands full reading to know whether you have to answer it while Plargg reads for the opponent like tap, cast something from the top of your library.
Furthermore, for drafting/including you need to know the full card. That’s what makes saga’s complicated.
 

landofMordor

Administrator
There is also something called complexity for the opponent. The saga commands full reading to know whether you have to answer it while Plargg reads for the opponent like tap, cast something from the top of your library.
Furthermore, for drafting/including you need to know the full card. That’s what makes saga’s complicated.
Good point. I agree, with a few caveats:

Sagas are usually Enchantments, which are one of the toughest card types to remove. So the opponent won’t often need to know whether to kill a Saga, because usually they can’t anyways. (The creature side is typically the simplest side, and the easiest to kill, for NEO.) This is an additional way that reducing agency hides complexity.

Also, during draft, Sagas (especially in NEO) follow the heuristic “a 2-for-1 on the front; a French vanilla on the back”, which really compresses/hides the complexity of the individual designs. To misuse an old saying, “there’s safety simplicity in numbers”. Plargg is a one-off with no heuristics available for his play pattern, by contrast.

And of course, like I mentioned in my original post, complexity is context sensitive. It may well be that Sagas are only less complex than Plargg in the context of a set stuffed full of Sagas that I’ve played a lot of. Maybe that doesn’t hold true for you or your group, which is a-okay.
 
Good point. I agree, with a few caveats:

Sagas are usually Enchantments, which are one of the toughest card types to remove. So the opponent won’t often need to know whether to kill a Saga, because usually they can’t anyways. (The creature side is typically the simplest side, and the easiest to kill, for NEO.) This is an additional way that reducing agency hides complexity.

Also, during draft, Sagas (especially in NEO) follow the heuristic “a 2-for-1 on the front; a French vanilla on the back”, which really compresses/hides the complexity of the individual designs. To misuse an old saying, “there’s safety simplicity in numbers”. Plargg is a one-off with no heuristics available for his play pattern, by contrast.

And of course, like I mentioned in my original post, complexity is context sensitive. It may well be that Sagas are only less complex than Plargg in the context of a set stuffed full of Sagas that I’ve played a lot of. Maybe that doesn’t hold true for you or your group, which is a-okay.
Well, plargg may be one off but it is also not. There are many creatures which draw or do something like that. Mentally, it is easy to parse.
 
Having watched someone play Pauper recently, a thought has occurred to me:

One thing to consider when thinking about the complexity of your format as a whole is the relative power level of your enablers vs. your threats.

If you look at Standard, you have an environment where your threats tend to be way more powerful than your enablers (with the exception of wild outliers like Expressive Iteration). As a result, games tend to be pretty simple despite the complexity of individual cards — your game plan of "find and cast my big bomb" tends to be straight-forward, and interaction tends to be pretty limited.

Pauper, on the other hand, has enablers that are wildly more powerful than the available threats. As a result, actually winning tends to be pretty convoluted, despite your card pool being pretty mechanically simple.

...

I'd honestly suggest checking out Pauper if you aren't familiar with it, because the decks straight-up feel like they came from an alternate dimension. Like, this is a format where Wx Gates is an exciting new deck, Ux Faeries is the top dog in the format, and where Chromatic Star and Ghostly Flicker are archetype-defining cards.
 
Top