Part 1 of this article talked about the concept of agency, how individual cards measure high or low on it, how agency for the player and for their opponents are separate dimensions, and common problems that become more obvious when viewed with agency in mind.

This is part 2.

Agency in Decks and Archetypes

Agency can be measured for decks and archetypes as well as cards. It is up to the cube designer to manage the levels of agency in archetypes - the players’ goal is to win, and they will try to do that by playing the best decks, whether they give them options or not. Below are common patterns and pitfalls of macro archetypes (aggro, control, etc.).

Combo decks are a mixed bag for their pilot’s agency. Storm, for example, requires difficult choices of when to go off, which color of mana to float, how to sculpt a hand for “the big turn”, when to Wheel of Fortune. These decisions make Storm a high-agency deck for the player (though it is quite linear to draft). For the opponent, though, playing against Storm is low-agency, since its plan can only be interacted with by counterspells and discard. Many decks have to just ignore it and race it.

In general, combo decks are low-agency for the opponent, due to the limited window and ways of interacting with them. Parallels from constructed aren’t always useful: Splinter Twin can be reasonably high agency deck in Modern for both players, since the combo is disruptable in multiple ways and it is more of a combo finish to an interactive deck. In Cube, however, it is rarely correct to play around it without knowledge of the opponent’s deck, making it much worse when they flash in a Pestermite then win on turn four, than in Modern, when it was a mainstay in the metagame.

Ramp is a common deck that contrasts with Storm for being low-agency for the player and not for the opponent. The deck’s lines tend to be pretty obvious, but the opponent’s lines do not, since they can bolt the birds, use discard, counterspells, keep removal up, race, and so on.

The version of aggro that I call Hyper Aggro is often low-agency since it’s about dumping 2/1s quickly and turning them sideways. More interesting versions of Aggro are a bit slower but include ways to refuel, disrupt the opponent’s plans, and constantly present a threat in the long game.

Midrange is the beatdown in some matchups and not in others, so decisions of tempo vs card advantage will be common. Recognizing who’s the beatdown in each game and in each point of the game is an important source of agency and strategy. Two dangers with midrange are running:

• Cards that are low agency for you, like threats that are low-risk to cast
• Cards that are low agency for the opponent, like removal checks.

Control can have very lopsided matchups, for good and for bad. It can be either oppressive or vulnerable to aggro, depending on how it is built and actual card composition, or it can be conversely oppressive or vulnerable to midrange. It’s easy to build control with too many effects that hose a particular type of deck but leave it soft to another, like Pyroclasm and Desertion. Requiring control to set something up and take some risk by tapping out to get card advantage is an important source of agency for it.

The most common pitfall for micro archetypes (artifacts, tokens, counters, etc.) is to reward them for focusing 100% on their own synergies and ignoring the opponent’s plans. Supporting that archetype by maximizing its power when ignored (like the cards above do) creates less interesting games that are just about racing. It is useful to support these archetypes in the form of answers that are reward going into archetype (Galvanic Blast), resilience so that its plan is harder to disrupt (Moorland Haunt) and non-multiplicative forms of scaling (Blastoderm).

Making sure synergy decks are slow enough to “go off” that it is in their best interest to run removal or disruption instead of laser focusing on their own plan is another way to introduce agency into micro archetypes.

Tribal archetypes are usually low-agency, since their plan is to create a large board and let the lords and synergies grow the power level of the cards quadratically. Decisions are mostly sequencing and risk/reward of playing around board wipes and removal. The versions that are all-in in the plan are the least interesting ones, while the ones that have a less linear game plan, like Counter Slivers or Goblin Bidding from constructed, lead to better games.

The nature of interesting decisions

Interesting decisions, as Sid Meier describes them, are the ones in which some players would choose an option, and other players of the same skill level would choose another. Context influences those decisions and having sensitivity to the situation to make the right one is what sets a strong player apart. Interesting decisions require information to be available so the player is making an informed choice - choosing which number to bet on a roulette isn’t interesting, for example. Interesting decisions often are decisions about tradeoffs of risk vs reward, short-term gain vs long-term sustainability, tempo vs card advantage.

Some types or cards, like tutors, are excellent at providing a huge amount of options, where most players see a few good candidates that are 90% efficient without much analysis paralysis, but great players see outlier lines that are far from the other efficient lines, but are the optimal ones. These seemingly bad lines that turn into a winning one are beautiful lines in all games, Chess, Go, and Magic alike.

Planeswalkers are a good case study in interesting vs non interesting decisions. Jace and friends often bring agency to the player who runs them, since their loyalty abilities represent an extra choice each turn, and to the opponent as well, since they have an extra target to attack. But planeswalkers are also a source of uninteresting decisions. When you’re already close to victory, they broaden the decision tree with many possible paths, almost all of them leading to a win, so you spend a lot of energy on decisions that are nearly irrelevant, just making sure you are not making a mistake.

When agency is missing

A few years ago, I severely cut down on card complexity and officially made elegance the defining characteristic of my cube, even renaming it “The Elegant Cube”. You can read about its design principles in this thread.

There was an unintended consequence, which is quite predictable in hindsight, and though I was warned by a friend not to “dumb down” the cube, I didn’t understand what I needed to watch out for. There is generally less agency in simpler cards. I say “generally” because some mechanics like burn spells, instants, mana sinks, creature combat, and flashback allow for a surprising amount of agency with very few words, and I started this with the examples of Lightning Bolt and Counterspell, cards that are both elegant and high agency. Still, favoring cards that don’t have added modes over ones that do, cutting down on activated abilities, selecting straight power/toughness boosts as payoff to cut down on words, trimming the amount of modal spells, and lots of other changes made in the name of elegance, in conjunction, decreased the amount of gameplay agency significantly.

Some high-agency cards I cut for complexity reasons

As always, a balance is best. The best cards are like Lightning Bolt: easy to learn, hard to master. Apart from those, tradeoffs must be made between agency, elegance, and other aspects like resonance, power level adequacy and archetype needs.

Too much of a good thing

When too many options are presented to a player, they have to spend more energy to take in all the information, process it, and pick a line of play. Options have a way of combining exponentially with each other, and though players are quite good at shortcutting, modelling, simplifying the problem and applying heuristics to tame universes of options, if the decision space is too smooth, analysis paralysis sets in, as the player navigates a plateau. Rotisserie draft is an example of a very high-agency draft (it is effectively a one 360-card booster draft) that has this risk.

Even when it’s not even a matter of situations that are too complex to understand, too much agency has the effect of rewarding skill too much, such that a better player rarely loses to a less skilled one. Chess is a game like that. Grandmasters do not lose to a bad draw (even Fischer Random Chess is relatively low variance) and that contributes to making chess less accessible than Magic, in which a beginner has a real chance to win at least a game in their first draft. Caw-Go was a dominating Standard deck around 2010/2011 which was perhaps the peak of consistency. Some experienced players enjoyed the high-skill mirrors, but to less experienced players, and to players interested in playing different decks, the environment felt oppressive. This skill-rewarding effect is undesirable in many playgroups, particularly ones that have players of varying skill levels, though it might be desirable in groups where players have similar skill levels.

Defining psychographic profiles in terms of agency

It is useful to categorize players over what exactly in the game they prize control over, what kind of agency matters to them. Note that though players often label themselves as one of the classic player psychographics, the concepts are more valuable as reasons to play Magic, varying in degrees of importance to each individual, than as an tagging mechanism for people.

Timmies/Tammies value agency over what cards they play and how they win. They want to play exciting, powerful spells, seeing unique and hilarious situations and in general having a good time. They don’t mind so much a low gameplay agency environment, as long as their strategies are viable. Timmies and Tammies benefit from an environment where a wide variety of effects are viable and unusual choices are sometimes correct.

Johnnies/Jennies value agency over how they build their deck and how it works. They benefit from having lots of options of how to build their deck and create synergy between the cards in the draft, especially if the synergy wasn’t intended. Johnnies and Jennies benefit particularly from cubes that are not overdesigned, and from high agency in the draft portion.

Spikes value agency over their win rate and how they draft and play. Outsmarting their opponent, and winning a game most other players would lose is what gives them a high. Spikes like pretty much everything in this article, but particularly gameplay agency.

Power level and agency

Though agency is not power level (well, not as I’m defining them in Magic, but power might very well be agency in real life), high power is correlated with high agency. There isn’t a main reason for this, but many factors added up:

• Options are good, and versatility is often undercosted
• Low mana values are correlated with power and also provide more agency
• Value generation like repeatable card draw is powerful and leads to higher agency
• Higher rarities are correlated with power and with complexity, and complexity is correlated with agency

This effect is what made Cube both start and still be commonly regarded as a “collection of the best cards ever printed.” The corollary is that the lower power your cube is, the more you need to worry about agency.

In practice, how do I increase agency?

• Keep most cards within a reasonable power band. Outliers on the high power side force players to run them. Outliers on the low power side punish players for trying to make them work.
• When evaluating cards, think about how they affect your options. For example, Aerial Responder usually flies over blockers, doesn’t need to block to race, and doesn’t need to forfeit an attack to block. It’s a low agency card. Flickerwisp also flies over blockers, but you have to choose between attacking and blocking with it, and it’ll always die in combat, so you’re giving something up to get it to trade with another creature. Sometimes you have a great target to blink when you first are able to cast Flickerwisp, but in other situations you have to choose between playing your clock on curve versus waiting to get value from the blink ability. The blink ability has many situational uses, so sometimes you’ll have to choose between blanking a blocker and reusing an ETB ability. Flickerwisp is a high agency card.
• When evaluating cards, think about the options that they allow or take away from your opponent. Stone Rain allows them fewer combinations of spells and abilities each turn and renders expensive cards in their hand dead. Great Sable Stag cannot be dealt with by Dimir decks. Evasion and hexproof prevent the creature from being removed.
• Run enough fixing that your spells can be reliably cast in time to be impactful. The faster the environment, the more this matters. The more multicolor cards you have, the more this matters as well.
• Run cantrips and smoothing mechanics, like scry and cycling.
• Run lots of cheap spells, especially at 1 and 2 mana value, even if your environment is slow. Make sure those cards support not only aggro, but other types of decks too.
• Make both tempo and card advantage matter in your environment. The choice of pushing for a quick win versus building a value engine for the long run is a classically deep choice. Environments that are too fast skew towards choosing tempo, and environments that are too slow skew towards choosing value.
• Run an appropriate amount of removal. Ideally for agency, in most games players need to choose which threats to destroy and which ones to leave. Too little removal and the game is about having the biggest creature or carrying out your plan first. Too much removal and it is a grind until players run out of removal or threats.
• Run enough mana sinks, such as:
  • Activated abilities
  • Recursive mechanics like flashback and embalm
  • Scaling spells
  • Utility lands
• Run modal spells such as [MDFCs|Kabira Takedown], [Commands|Witherbloom Command] and [split cards|Fire // Ice], but do find a balance with wordiness and complexity adequate to your playgroup’s experience level.
• Make sure your aggressive decks have ways to refuel, like Bomat Courier.
• Avoid cards that are bad topdecks, or need to be cast in a certain sequence to be good, such as Champion of the Parish.
• Consider more forgiving mulligan rules, like one free mulligan to 7.
• Evaluate your archetypes in terms of agency both from the player’s point of view and from the opponent’s. Having a few low agency archetypes is not a problem: with enough agency in the drafting section, one can choose to play a low or a high agency deck. However, make sure your archetypes have a good amount of agency on average, and if your group leans towards Spikes, try to get all your decks to be high agency.
• Pay particular attention to micro archetypes since they tend to be low agency more often than macro archetypes.

Conclusion

Managing agency is a powerful tool to add to your cube design toolbox. While a good level of agency comes naturally to high-powered cubes, maintaining that level is a challenge in others, such as tribal cubes, low-powered cubes, and synergy cubes. Designers of cubes of all types can benefit from the concept by considering agency at the level of cards, decks and matchups.