This article is long and will be posted in two parts. This is part 1.

For years, I looked for ways to describe why I liked certain cards, certain decks, and certain play patterns. I used generic and non-committal words like cool, fun and interesting and terms that beat around the bush like good gameplay and interactive. I did isolate some aspects that I value about cards, like elegance and resonance, but only recently I’ve found a word that explains my instinctive preferences and my love for cards like Fact or Fiction: agency.

Agency is a player’s ability to make informed decisions that impact the game meaningfully. It is at the core of games; without agency Snakes and Ladders is as good as they get.

In the early days of my cube, I unthinkingly favored high agency cards. One of my design goals was to challenge the notion that “there is a single line that maximizes your win rate in a given game, and every time you stray from that line is one mistake you make.” In pursuit of this goal, I loaded up my cube with cards that offered options and difficult choices, such as modal cards and tutors, so that the options were so broad that finding that perfect line by brute force was impossible, and playing a game felt more like painting a picture than navigating a minefield.

Over time, my priorities shifted away from that design goal, and though the draft experience improved, the gameplay became shallower.

Playing a lot a variety of limited formats recently made me notice that I was neglecting gameplay agency in my cube. In Kaldheim, I had fun playing aggro because equipment, runes and boast created many lines and mana sinks even late in the game. I saw a profusion of options in the Afternoon Delight cube, built into the cube to fill the void between 2, 3, and 7 drops, and felt the gameplay was very good since there were always decisions to make. I learned some things about psychology, and how a sense of control is a thing our brains crave in general.

I turned to The Elegant Cube and noticed I had too many linear decks in it, which simply tried to do something synergistic and powerful. I had focused on agency over which decks to play, but not over how to play them.

High agency cards are fun

Agency is the hidden umami flavor that makes cards like Lightning Bolt and Counterspell favorites among players.

A 1-mana, instant, that can target players, creatures and planeswalkers alike is as flexible as a burn spell gets.

The instant speed allows Bolt to be cast on any step and opens up many lines of play to which sorceries do not have access. Bolt can blow out a double block, a combat trick, destroy a Creeping Tar Pit, trigger Monastery Mentor precisely when necessary. Knowing when to main phase it and when not to is a small but meaningful edge to gain.

The low mana cost of Bolt allows for more combinations of plays with the same amount of mana. Being mana efficient matters less - one rarely Bolts something just because it can’t be squeezed into the rest of the curve in a future turn.

The ability to point Bolt at a creature or to face opens up lines in which the caster sacrifices board state to go for the win. A risk/reward scenario, one of the most common ways to create interesting decisions. Burn is only “dumb” when the deck is built to point all spells at face and hope for the best.

Let’s shift to the most refined instrument of control: Counterspell. The original Counterspell is a feat in elegance. “Counter target spell.” has to be the more efficient way of phrasing: “You may keep mana open instead of playing your spells. If you do so, during your next turn, if you haven’t done so yet, each time your opponent plays a spell, you may cast this to deny them permission to resolve that spell. This card can be played from outside the game even if you don’t own it, but if you do not have it in your hand, you have to give permission to your opponent every time they ask.”

Keeping mana open for Counterspell is a decision in itself as hard as actually casting it. In fact, a common heuristic to decide what to do is to list out the combinations of spells that your available mana allows you to cast. “Counterspell” is an option in your combinations even if you don’t have it. You can represent one by keeping two islands untapped and thinking for a bit before letting spells resolve until the theater falls apart, depending on the matchup, your acting skills, and your opponent’s boldness to test your claim.

Getting past the decision of whether to keep two islands up or not, now the ball is in the opponent’s court. Do they play their best threat? Do they try to bait out the counterspell with a lesser threat? Do they pass back, trying to flash in something or use their mana to activate an ability instead of playing into open mana? Depending on the context - and the uncertainty of whether you actually have it makes the context fuzzier - the decision will be different in each game and different players of a similar skill level might make completely different choices.

Assuming the opponent does cast something, should you counter it? Often it’s obvious, but the opponent is incentivized by game theory to cast the spell that exerts the smallest amount of pressure that would make you counter it or the best spell that they do not believe you would counter. They are incentivized to mislead you in the evaluation of how important a spell is for them. In other words, the opponent is rewarded for creating a good subgame for you.

Of course, if the opponent is unaware that you are running counterspells, or cannot afford to play around them, counterspells reduce the opponent’s agency. Hence, counterspells are particularly unfun to play against in a number of situations: for new players, when they are a surprise like Mana Tithe, or when a deck runs so many of them that it will counter anything and everything. Counterspells are fun to play against only when you feel like you have control over the interaction.

Cryptic Command is another example of a high agency card. There are six ways the modes can be combined, and all combinations are relevant in different situations. Being an instant gives it a lot of possibilities in terms of timing: upkeep Time Walk, precombat cantrip Fog, anytime Dismiss, Repulse in response to equip.

For contrast, take Grave Titan, for example, a card shunned by many cube designers as a “dumb” threat. There are not many decisions or interesting questions with Grave Titan. Its controller almost always attacks, because making more zombies and swinging for 6 with a body that’s hard to double block is always good. If the opponent has removal, they’ll play it. The situations involving the card play themselves.

I don’t mean to say that agency equals fun. Fun has many components, like your memories of playing the card and having a good time, how good the art of a card is, how much control over the game it feels like you have when it’s sitting in your hand, if someone doubted the card was playable and you are proving them wrong. Still, I believe that agency is one of the most important components - and perhaps the most important component - of how “fun” a card is.

Agency for you, agency for me

Agency is not one, but two dimensions.

Thoughtseize is a high agency card for its caster, but it removes options from the opponent. Discard, land destruction, True-Name Nemesis are particularly bad offenders to the opponent’s agency since they drastically narrow down their options.

Removal checks like Baneslayer Angel are also low-agency cards for the opponent, while not necessarily low agency for the controller. If the opponent has removal, they have to use it. If they don’t, they have no agency over the game anymore and will likely lose. That’s a matter orthogonal to whether the card has a high win rate or not; it’s a matter of the players’ choices not dictating the outcome of the game. Yes, one could argue that the opponent should have held a Doom Blade to deal with something like that, and because they ran out of removal, they should lose. But games can be decided by threats that have more options of counterplay, such as blocking or racing.

On the positive side, some of the most interesting cards are those that do not reduce, but actively increase the opponent’s agency. Fact or Fiction is an example - the opponent is given more agency than the player - in fact, the opponent is trying to create a decision as interesting as possible to the caster, which is a small minigame of game design.

Another positive example are Morphs, in formats where there are lots of them and they are viable. The opponent can interact with them in a number of ways, and the hidden information adds a thicker layer of risk management to decisions about removal and combat.

Complaints that are, really, about agency

Variance is deplored in Magic when it results in a lack of agency. Losing a game in which you were forced to mulligan to 4 is the classic example of shrugging a loss off as “variance”, but the issue wasn’t that an improbable combination happened. Having varied games is nice. Games in which all lines of play lead to a loss are not. The real problem was a lack of agency.

When players say they like interaction, they don’t mean Hymn to Tourach. They mean agency over the opponent’s plans, by choosing what to spend spot removal on, what to block and trade with, or whether to leave mana up for a counterspell.

Drafting on rails is a problem in cubes with insular themes because there is little agency after the first few picks. In these cubes, abandoning a lane is bad, your picks are automatic and you are forced to make uninformed decisions with little information from signals.

Topdecking wars are uninteresting due to the lack of agency. Whatever is drawn likely will have to be cast. The emotional high of flipping a card over is only fun sparingly.

The reason some mechanics are more liked than others is usually linked to agency. Cycling, flashback, kicker, adventure, treasures, evoke and investigate are all well-liked mechanics whose common feature is giving agency to the player that plays those cards. Disliked mechanics like miracle, affinity, hexproof, cumulative upkeep, epic, fateful hour and tribute are often mechanics that actively remove possible lines or make one line of play so much better that it is effectively forced. Of course there are some exceptions: cascade and proliferate are some mechanics that are well-liked and low agency. And then there are divisive mechanics, like dredge, storm and buyback, which usually give agency to the caster but remove it from the opponent.

Part 2 of this article will delve deeper into player agency:

• Agency seen from a higher level of decks and archetypes instead of individual cards
• What kinds of decisions are "interesting"
• Environments with too little or too much agency
• Timmy/Tammy, Johnny/Jenny and Spike
• Correlation between power level and agency
• Practical ways to promote player agency