I liked this topic and I wrote about it. Some people have already commented on some of the same stuff it says already. Hopefully you don't mind reading it again. Maybe this is an article in the making, but I'm sure it needs (lots ) of editing.
Archetypical Thinking
It should go without saying that the purpose of a cube is to produce decks for playing games of Magic, but sometimes I wonder if people lose sight of that. Cards in Magic exist in relationship to other cards, but it is impossible to compare each of the potential decks that a cube can produce against the each of the other decks a cube can. As such, it is very tempting to look at easier to analyze metrics to decide whether a card is worth including, such as its abstract power level. While this analysis has some value for developing and focusing a cube, it is largely of secondary importance.
In order to compromise between the impossible task of comparing all possible decks and focusing on easier to see, but ultimately secondary factors, most designers organize the possible decks their cube can produce into a manageable number of categories bound by strategic similarities. These archetypes can be organized in a number of different ways, but possibly the most useful one looks for similarities in when, how and by what means deck tries to win.
People have been organizing Magic decks in this way for a long time. The most common purpose of these analyses was to group decks together for the purpose of defining which decks have the advantage against which other decks. An early article on the Dojo proposed five archetypes in constructed Magic: Beatdown, (disryptive)Aggro, Combo, Board Control, and Pure Control (excuse me if the terms are wrong, I’m working from memory of something written for Urza-Block constructed). Another model posed the aggro deck, the 2 for 1 deck, the counterspell deck, and the big spell deck. A common one is the aggro, midrange and control Roshambo. The important thing to know about all of these systems is that all of them are useful and none of them tell the whole story. They are tools to help you figure things out. However, if you take any of them too close to heart, you run the risk of having them become design constricting and self-fulfilling prophecy.
Take the Roshambo model. It makes a lot of sense. Fast decks beat slow decks before they can play the game, but slightly slower decks beat slightly faster decks because they play better cards. Now, if you take this to heart, you can certainly build your cube along three archetypes, where players can build a deck along this single axis. Every deck will have a deck it can beat and a deck that can beat it. This is the designers dream right?
I suppose if spending tons of design hours and money on cards to play rock/paper/scissors is your dream, then this is it, because that’s basically what you are aiming to create here. Of course not all games will play out in such a deterministic fashion thanks to the healthy amount of variance created by shuffling and drawing cards, but when you embrace this as your balance metric you are actively attempting to create a game with the depth of rock/paper/scissors. Under this conceit, if a slow deck has the advantage against a fast deck, there is a flaw in the design of the system. To me, this seems extremely uninteresting. With all the cards and the interesting actions they create, having them amount to a one dimensional system based on one factor of those cards in game purpose is wasteful. Why play a game as complicated as Magic to create something so simple?
So, at this point there are two competing points: there is some level of truth to the Roshambo model and that designing a cube based on that as a balance point is highly undesirable. The key to reconciling these points lies in the complexity of the game of Magic. Cards in Magic do specific things and interact with each other in specific ways. These specific interactions can override the general rules for the relationships between decks based on their general characteristics. While it might be true that in the absence of other factors, fast decks beat slow decks, there is never an absence of other factors because games are played with specific cards. Because cube design allows for complete control of the cards in the format, much can be done ensure that the format does not devolve into Rock/Paper/Scissors.
The prevalence of the Roshambo model in cube design is hard to miss, but the reasons for it make sense. With its focus on power level analysis, cubes often defer to the best in class of each type of effect. With everything turned up to 11, role flexibility is drowned out by sheer efficiency. Decks can’t tune themselves very far because the tools that they use exist in as the superlative of their singular function. When a deck can be filled to the brim with nothing but cards that are exemplars of the specific archetype, those extreme decks will almost certainly become the best decks in the format because they are so focused. So, of course the Roshambo happens because it was designed specifically to happen. When you push design in a direction where you only run the aggroiest aggro cards, the controliest control cards and the midrangiest midrange cards, those cards are going to congeal into three basic decks that exist to counter each other.
That doesn't make this division inevitable, because again, specific cards interact specifically. If one opens up the power level band to allow for a real toolbox of design choices and takes a step back from hierarchical archetype balance, there is really a lot of space to create more complex relationships between cube decks. The speed at which a deck wins the game is not the only relevant concern in balance. Decks also win the game by different means and they may directly interfere with each other, all of which can be utilized through card selection to adjust the matchups in a cube.
One need only look at the history of both limited and constructed formats to find countless examples where favorable matchups are the reverse of what one would predict based on general theory. Archetypes , particularly those attempting to pick favorable matchups, are of limited is use in the area of environment design because the design of the environment decides which archetypes are in it, not vice versa and and even if it is not intential..
Still, it is too daunting to attempt to consider every possible deck and matchup into decisions about card selection, so simplifying the thought process by grouping these decks into archetypes is critical. The tricky part deciding where to group those decks in a manner that makes the most sense. Circular groups don't work well because they end in linear matchups. The archetypes utilized should be able to be paired with each other archetype and be flexible enough to allow either side to have the advantage depending on which tools the deck decides to utilize. Granted some archetypes will have consistent weaknesses against others due to the way key cards or concepts intersect and that is fine. It's not desirable, but if you are striving for variety its inevitable. The key is to avoid it where ever possible. The drama of the matchup should be about the unique composition of each deck and how it pairs up with the opponent's deck. Part of why I love Westchester Draft is because since each player knows exactly what everyone else is drafting, they can attempt to make subtle adjustments to push themselves to the best possible position. In a Rashambo model, the best you can do is to try to be rock when four opponents are playing scissors.
So, what exactly should archetypes be based on? The ways in which a deck interferes with others is important, but it is much too vague as it gives no idea by itself what the deck does. This interaction is very important of course, because this is what changes Magic from a solitaire race to a game where you actually have an opponent. How a deck interacts with others is something that a deck can tune, but it gives no real insight into how other decks will match up against it because it tells us nothing of what the decks actual function is. The means by which a deck wins is probably the best place to look for this, as winning the game is the goal a deck is striving for. If a deck isn't made to win the game, it is irrelevant for the purposes of balancing it into the environment, so we can freely ignore that case.
The means by which a deck wins the game is really the fundamental basis of the deck. The end of the game is the end of the conversation. A deck has to win before the opponent wins, either by simply reaching its win condition faster or interfering with its opponent’s ability to win the game. In Magic, interfering need not be directly through "take that" mechanics, because most Magic cards have multiple purposes this can also be done indirectly. Blocking with a creature, for example, interferes with the opponent’s ability to win the game, but usually a creature isn't included in a deck for the expressed purpose of blocking. These incidental interactions are usually where favorable and unfavorable matchups are determined, though speed is still a factor, most often when two decks have the same means of winning. In these cases, both players cards are of similar function, so most of the interaction is extremely straight forward and the player with the better, but slightly slower, cards will win unless one players cards are faster to such a significant extent that they are able to win before the opponent can deploy their card quality. The Roshambo model comes into play in these cases when decks are very similar just like they do when decks are defined by extremes and the deckmaking interest comes from attempting to build the same archetype in a slightly slower version then the opponent (or trying to build the fastest version possible and hope the opponent gets stuck with the slow cards).
When defining these archetypes, care must be taken to dig deep enough to really understand how the deck wins. Most decks win by hitting people with creatures, but clearly hitting people with creatures isn't an archetype. How the creatures are amassed, why the creatures can hit the opponent in spite of their resistance and why your creatures will kill the opponent before their creatures kill you are in combination how the deck wins. Some means are relatively firmly tied to a timeframe for winning the game, others are much less so. Often times archetype will be linked with a particular mechanic that the deck utilizes to gain an advantage, but that mechanic can be utilized in a number of different ways that creates a large variety of decks that come from it. The key takeaway is the important implication that comes from this: how the deck fails to win. Having decks able to make each other interfere with each other in interesting ways is possibly the most important thing for the designer of a cube to do, because this is what separates Magic from two people playing solitaire. Once we know how decks fail to win, we can includes tools archetypes can use to ensure this happens, hopefully without resorting to blunt "take that" style mechanics. If players fail to interfere with each other or only do so in rote or superficial fashion, there isn't much of a game going on and as with the Roshambo model, you are probably better off playing a different game if that's what you want to do.
Creating the web of interesting interactions is complex. Once you get through the more mundane basics of designing a cube, it is basically the entirety of the process. As such, it is impossible to simply write an article telling someone how to do it, but there is definitely some general advice.
Make sure everyone can play in every phase of the game: Overspecializing decks and defining their cards by costs is going to reduce interactivity in games. Regardless of what a deck is trying to accomplish, it should be able to do something on almost any turn. Due to the way mana works, this means your mana curve needs to be low. Even if you think your mana curve is low, your cube would probably improve by making it lower. The only turns guaranteed to be played in a game are the early ones and these are the turns where you have a full hand of cards that should give you plenty of options, but you can only have plenty of options if they are in your cube to be drafted. Particularly since many low cost cards are going to be redundant in function, having a wealth of options available at mana costs 1 to 3 is vital for allowing a variety of decks to function. This is important for basically every type of deck. Some decks use early turns to play threats, some use them to assemble synergy and some just attempt to stall through them so they can deploy more powerful cards later. Players need the resources to do these different things as well a multiple avenues for doing so or else archetypes will be inflexible and there matches will become circular because there is no way to adjust them.
A similar, but more subtle extension of this is giving every deck things to do on later turns. The goal here is not to force players out of aggressive decks, but rather to give tools to all decks, including aggressive ones to compete in a longer game. By controlling the power level of cards, the all-in aggro deck is not invalidated, but the need to play all-in aggro in order to fill a role in the Roshambo is gone. Decks can hedge alternate plans that give them play against a variety of decks while still supporting their main tactical angle. Utility land drafts are an amazing tool for getting long term plan cards to early game decks and I cannot recommend the concept enough.
Play cards that cross borders: Try to find cards that play roles in different archetypes and include as many of them as you can. You will find that cards can help multiple archetypes in extremely different ways. The simplest example is how red burn provides an aggressive deck with late game reach, but it provides a controlling deck with early game defense. Cards that fill this category sometimes have explicit multiple functions, like charm cards, but often it is much more subtle then that. This take a lot of experimentation, but the best way to find these things is simply to constantly try new cards and see what decks they find their way into. As smart as you think you are, you are no match for the combined creative force of a group of people playing your cube. Just throw things in and see what happens. You will be surprised.
Avoid cards that destroy distinctions: While having individual cards that cross archetypes is important, archetypes should still be strategically distinct. Care should be taken to avoid cards that destroy these distinctions. The cube designer makes a promise to a player, in a way, when they support an archetype. You do not want to break that promise. Consider in a Roshambo balanced cube, a player drafts Red/Green Zoo. They take the Goblin Guide, the Kird Ape, the Chain Lightning, the Flinthoof Boar and all of the other tools that go with that. The Roshambo Cube makes the player a promise: You are the fast deck and will have an advantage against the control deck. You will be the beatdown. So, after shuffling up, they start playing and their control opponent plays tinker on turn 2 of a mox or sol ring and brings out Sphinx of the Steel Wind. The promise is completely shattered and they feel like an idiot for believing it. The game is a shit game that wasn't even worth playing.
Situations like this need to be avoided because they create a sense of futility in the drafting process. Players gravitate towards an archetype with an expectation for a certain kind of gameplay. Cards should improve the quality of gameplay by adding decisions and choices, not remove the need for either. If an archetype is supposed to fulfill some niche in your cube make sure that you don't include cards that do the same thing, but better, that are usable by another archetype.
Your aim as a designer is to able to look at what each archetype does and ask, "What does this other archetype do against it?" If the answer is "lose" or "pray to topdeck a specific card", there is work to be done with regard to how those archetypes relate, because neither of those are qualities of fulfilling games.
The main culprit in this area are cards that provide mana or avoid mana costs, because mana costs are the key control in the game for deciding the timing of effects. The timing of when effects come down is critical to cube design. Changing the timing of powerful effects requires the entire environment to adjust to it. Entire archetypes can cease to function when effects become available turns earlier.
After going through these generalities, the specifics aspects of individual archetypes and how they relate come in. Carefully tinkering your selections will make your matchups more intricate and interesting. But the decisions will be tricky. You will make mistakes. There will be some non-games. You will laugh at yourself. However, your cube will get better and as you actively pursue these improvements, you will have more and more fun cubing, because you will be playing fewer games of solitaire and rock/paper/scissors and more games of Magic.
Also thoroughly enjoying these discussions we've had this week!
