Intellect and Emotion in Game Design: A Personal Cube Journey

I came to Cube design with a bit of a chip on my shoulder. If my first cube didn’t begin as a challenge to the status quo, my second one certainly did. But that half-joking project didn’t turn out like I expected, and is instead something novel and surprisingly fun.

Starting out, there was a lot I didn’t know about game design. There’s a lot we Magic players are not honest to ourselves about: why we play, what we value in it, and what we mean when we talk about “design”. Understanding ourselves better may help us have more fun in our hobby, and I think it would absolutely make us better curators of Cube.

In ~2015 the most vocal elements of the Cube design community shared a common focus: play the better cards. I didn’t get it.¹

Many Cube designers are players first. For them, Cube is a format where they can play with all their favorite cards. My curiosities were different. I have a background in design, have some experience building and programming games, and have always been fascinated with systems and generally how things work.

Behind all the aesthetic pieces — dice, cards, mini figs, boards with paths and positions — there’s a process that is “the game” in a pure, systematic sense. There’s a tree of decisions players navigate within the rule structure to advance the state towards some goal. The physical components of tabletop games² are just props to help track state, guide the understanding of rules, and provide sources of random inputs.

Mark Rosewater uses a similar framework to distinguish between games and other ‘activities’ or ‘events’ based on the importance of decisions.³ Without meaningful decisions, players are just accounting out the mechanics of a system. The decision tree of possible actions is a straight, unbranched line.

But critically, the distinction between game and activity isn’t purely objective. A system can include trivial branches where one option is unquestionably correct. In Candy Land, players move pieces along a linear path robotically, but occasionally have the choice to take a shortcut. Taking the shortcut is always correct. If a player doesn’t know that, they’ll still have the experience of making choices. As soon as they do understand there is only one right options, the trivial, irrelevant branches are pruned off, the game is solved. The decision tree collapses into a straight line. You may as well just be knitting.

Tic-Tac-Toe takes a little more work to solve. Players are given more options, more branches they could conceivably move down. But just like Candy Land, there is a single dominant strategy. Once a player learns it, they are no longer playing and participating is following the prescriptive strategy, merely doing an activity.

From this context, my ideas about Cube design developed. I wanted to look past the dice and cards and look at the system. My understanding was that a great game was all about creating a rich, meaningful decision tree. Minimizing the extra work and friction of trivial decisions or events that invalidate previous choices players made, and making every decision as meaningful as possible.

Magic players have borrowed the term "Spike" from Wizards’ design team to describe players primarily motivated by winning.

Ben Stark represented to me the ultimate Spike⁴ and is a useful model for thinking about the game from the player perspective. He is equally interested in digging down to the bare bones, not for the purpose of creating decision trees, but for pruning them down. You could replace all the illustrations and flavorful names with black and white text and I.D. numbers and you would take away nothing that contributed to his satisfaction in the game.

Most of us are not Ben, but we all fall somewhere in the multidimensional spectrum of a range of motivations that he represents one extreme of.

When I started designing Cubes, I was excited to make use of the parts of my collection which were irrelevant in other formats. I was also excited flip Magic upside down, unscrew the case, and tinker with it from the absolute inverse perspective from a player: not analyzing which pathways down the decision tree lead to the best chance of winning, but designing the systems that create those trees and making all the branches meaningful and fun.

I took to designing my first Cube, gradually at first and then with growing enthusiasm. I collected a list of cards I was interested in, pored over my collection, found trends and themes I thought were interesting, cut and padded and molded and eventually sleeved up a stack of cards.

This was at the time in the context of a culture that thought of Cube in a specific way: a set of the most powerful cards.⁵ The commonly accepted reasons to deviate from this standard were hard restrictions: pauper, peasant, plane themed, only cards with cats in the illustration. I wasn’t interested in the power direction — every other way to play Magic is already that, plus we already had a cube in the group following the ‘powerful cards’ archetype. And there was vast space to explore without extra constraints. I just wanted to make something fun that catered to the kind of Magic I liked.

A big part of that went back to the abstract decision tree. As a player, I struggled in the “all the powerful cards” Cube. Opponents would often do things I felt invalidated many of my decisions. I’d moved away from Commander already specifically because of this lack of agency I felt. I’d make decisions moving down one branch of the tree only to have an opponent lop off the whole limb incidentally. Ben Stark, our Spike paragon, has talked about rares in Limited with similar frustration. He’d prefer to throw out rares altogether to avoid games where a high-power card invalidates a whole history of decisions, instead maximizing the number of decisions that could move either player toward victory.

I gravitated towards a particular power level — somewhere around the upper tier uncommons and low tier rares of traditional Limited — a meaty slice of unsung cards full of rich designs.

I focused on ergonomics. I looked for particular types of complexity in cards that were individually easy to read and understand but created vast networks of emergent complexity when combined.⁶ This kind of strategic complexity is invisible, and doesn’t intimidate new players, but offers satisfying challenges to the experienced. The goal was to be as deep and re-playable as possible while still being accessible to our playgroup of mixed levels of experience. Even when it came to aesthetics, unless there was a good gameplay reason to make an exception, I wanted visually consistency. I wanted a crystal goblet for pure gameplay.

Since my unique priorities weren’t what drove card prices, this cube was incidentally cheap, making it even more functional. I could comfortably bring it to the shop, a bar, or lend out decks for a tournament.

I named the project the Regular Cube, in a bit of deadpan, self-effacing understatement. But the Regular Cube is also, quietly, a bit defiant. Contrary to what 2015’s self-described Spikes might say, my decisions weren’t gimmicky or out of bounds. I was building a Cube in a completely normal way in accordance with the letter of the rules that defined Cube, while trying to push for a completely different spirit.

Unsurprisingly, the Regular Cube had some issues at first. Matches were a little too slow, some colors didn’t have a clear identity, and players were unsure of what strategies made sense. A few individual cards dominated and others fell flat. I knew playtesting was the most important part of game design, so I made sweeping changes at first and then gradually smaller adjustments. Through playing, observing, and iterating, the list got to a place I was happy with it.

The initial reception to the Regular Cube was mixed. People were happy to try something different, but there was some confusion in the local playgroup. Why would I have included a particular card when clearly there was a “better”, similar option? This sometimes manifested in frustrating play experiences: “I could have won if my cards had just been better”. But to me that didn’t make any sense! Everyone’s drafting from the same pool. If that player’s cards were better, so would have all their opponents’. It’s zero sum. I wanted to tell people not to criticize the card choices, just focus on trying to win in this particular context.

I wasn’t trying to pick the best individual cards, I was trying to pick a system of cards that created the densest, highest agency decision trees. Avoiding the highest-power cards and choosing from the thick, middle chunk gave me a huge allowance of design flexibility. I wasn’t bumping up against the rough, irregular ceiling of powerful Magic cards, with only one direction to move in. As I iterated based on play experience, I could tune the Regular Cube down or up in power level.⁷

Gradually, reception improved. The Regular Cube got closer to my goals, and expectations around the gameplay started to gel. I think the second was a lot more important to people having a good time.

Outside the group, in the broader community online, it was more difficult. The suggestions to replace cards with strictly better options were common. Folks weren’t impolite, or incorrect in their intent, but their suggestions were presented as obvious and easy changes that would “improve” a particular card slot. My goals and the whole-system approach were difficult to communicate.

I designed some other Cubes, exploring the space from different angles, focusing on different aspects, imposing restrictions. Restrictions can breed creativity by forcing you to come up with novel solutions. Novel solutions could potentially be poured back into and improve Regular Cube.

I simmered in light frustration about the way Cube was talked about. Design, I thought, was how the components of the system internally related to each other. “Improving” components of that system made no sense when they all existed in relative tension against each other. Some players called themselves “Spike cube designers,” justifying their affinity for powerful.

They were following this analogy: A Spike plays the best cards in their decks, therefore they, as a designer, will include the best cards in their Cube. This analogy makes a lot of sense superficially, but loses a lot of usefulness by skipping over steps, and clouded community discourse about design. A “Spike” is a player defined by their primary motivation to win. Playing the best cards is an important step towards winning more games. “Winning” at Cube design, at least the sense I was interested in, is actually something totally different: creating a fun game.

This disconnect is more extreme if we step beyond of Cube and think about Magic card design altogether. Cube has the bumper rails up. You can add any Magic cards you want, and the rule system and cards work to make a functional game. If we took this “Player Spike” mentality to designing cards from scratch, it immediately falls apart. Lightning Bolt would be better if it did 4 damage, or maybe even 5. But why stop there? How about a clean 20? Why encumber it with a resource system at all? Let’s make it cost and give it Split Second just to be safe.

Spike players will always play “Better Lightning Bolt”, but the specious Spike mentality doesn’t work when applied to design. It just creates dice rolls. Being a Spike design means something completely different.

With me in this head-space, the Turbo Cube began as a wry demonstration of the absurdity of the “better cards” mentality.

The seed was planted through a “Turbo Draft” event on Arena, a special Ikoria bot draft with the twist that all spells cost less.

This sparked an idea. If everyone’s so sure more powerful cards are objectively more fun, why not let them have it? Why not take on that bigger role as the designer and change the rules of the game to make all the cards “objectively” more powerful. What greater way could there be to turn the spotlight upon and to trivialize the absurd fetishization of “power”?

Well, if I was going to do it, I was going to try to make it interesting. I drew up a plan: mana was too big a discount. Even in the Arena event, only a small number of cards in the set were relevant. A card that originally costed is likely to be wildly more powerful than one that costed , once the discount is applied. There are vastly more Magic cards at lower mana values, so a smaller discount would be more flexible. I figured would have the best balance of being a major change while retaining the most cards to work with. I expanded the discount to all effects players initiate, both spells and activated abilities, which avoided some awkward scaling and felt natural.

I laid out a few other guidelines: excluding cards that were individually broken (think Basalt Monolith), avoiding some confusing interactions,⁸ and favoring cards that get tweaked in interesting and impactful ways. Then I dove into Scryfall. I focused my search particularly on cards with costs of exactly and with a single colored pip, assuming (correctly) most things more expensive would be completely outclassed.

It was thrilling. I was doomed from the start. Prophetic Prism became not only a Mox but a rainbow mox that also drew a card! Eggs became cantripping rituals. The number of junk cards that made Goblin Guide look trashy would make anyone giddy. Sorry Counterspell, Convolute is embarrassing you. Wouldn’t it be cooler if Thoughtseize could Coerce you to discard lands too? Somehow there’s still no strictly better Lightning Bolt… but I’ll take Char and half a dozen other cards to deal 4 for 1 mana.

Like my previous cubes, I accumulated a list, found some trends, and mushed it into shape. Artifacts were an easy theme to emphasize, unsurprisingly. Land Cycling could create a novel suite of mana fixing. Red Aggro jumped off the page. Green presented some challenges… but when in doubt, I doubt you can go wrong with Saprolings, right!? I leaned into the Turbo aesthetic. Signets are powerful, don’t have the explosive vibe of eggs that, while ephemeral, dig through the deck. I added plenty of cycling cards and reusable abilities so players could really go ham.

The first playtests of the Turbo Cube were a little rocky. There were some broken interactions (in the literal R&D sense creating zero-decision, dominant strategies. Back to playing Candy Land!) We learned that the disparity between discounting spells at 2, 3, and 4 mana was even more extreme than I anticipated, so we brought the curve down. Green got a little turbo lands theme giving it its own explosive potential. The list improved, and, like Regular Cube, the experience improved tremendously with clearer expectations. Expectations that we could communicate to new players too: Heads up, the first time is a shock. Your opponent, on turn 1, just might attack you with 12 creatures, draw their whole deck, mill you out, or hit with a 50 pound laser. But enjoy the puzzle of it and try to figure out how to be the one to do that, or to be prepared stop them next time. Also try to win the die roll. That’s really, really important.

I softened some initial guidelines. I threw in some iconic cards that were already kinda broken under normal circumstances. Disregarding ergonomics, I added double faced cards, cards with irrelevant text, funny interactions, and all the best least readable cards Wizard’s new penchant ludicrous variants had to offer. For a bit, I even added my cycle of original dual lands, cannibalized from dusty commander decks. The Turbo Cube became a thing. It was a place to be ridiculous, luxurious, and decadent. I even filled the basic land box with nonsense to match.

In spite of my original intentions, Turbo Cube was fun. Really fun.

I don’t think I was wrong about my earlier, more formulaic ideas about what made good games.. The abstract model of game play as a decision tree is useful. It’s relevant to understand how games work in their bones without their aesthetic trappings. And, to be fair, I never said more powerful cards were less fun, I just said they weren’t objectively more fun.

But I was missing a lot. Games are not just an analytical enterprise. Even without aesthetics, the shape of that decision tree has a huge impact on the experience beyond just more meaningful decisions being better. It’s easy to think we’re just here to win and it’s all about good gameplay, not the artwork, or aesthetics, or personal stories, or excited tension of whether you’ll hit the right topdeck. It sounds more serious and like a justifiable demonstration of intellectual prowess to say all those things don’t matter. But we’re not really being honest if we do. Games are as much an emotional enterprise as an intellectual one, if not more so.

Maximizing player agency in the perfectly ramified decision space is a good start. But the irregularities, fast linear corridors, tangles, tense moments, dead ends, and the variety among these are part of an important emotional experience with a game.

We can call a turn 1 combo win, before every player even has a chance to draw a card, a failure of game design. We can also recognize the power of that experience: the frustration and disappointment that can draw a player in to trying to turn the tables next time; the curiosity over the complex puzzle that had to be unravelled to solve the game; the anticipation that can come from not knowing if a sequence will play out; the tactile satisfaction of taking prescriptive actions to reveal whether or not a plan will unfold; the joy when it does and disappointment when it doesn’t. The Turbo Cube has me wondering if we only play games to feel satisfaction and happiness. The disappointments are necessary to define success, but are they purely an unwanted byproduct of a zero-sum experience, or is the whole range of emotions that keeps us coming back to the table?

Players in the Turbo Cube might be aggressively knitting at each other to see who finishes first.⁹ Arguing whether that’s a “game” is missing a more important question: are people having a good time?

Not only was my Turbo Cube unexpectedly fun, there was another problem with my plan. The gimmick was, if power is fun, then here, just at this special table, these cards are hyper power. That only works if players buy into it. It was a flawed premise from the start because powerful isn’t fun because it’s objectively cheap and efficient, it’s fun because it exists in a context where it’s cheaper and more efficient than other things, and players get a rush from that contrast, which was the whole point I was trying to make while eating my own tail.

In the context of Regular Cube, there is still a most powerful thing you can do. If the game was a purely intellectual experience, then opening up the 5 out of 5 winningest card should be just as exciting as opening up the 10 out of 10 in the “classic cube”. But it isn’t. Players bring their X out of 10 scale that works in the larger context of the game with them.

Switching to the Turbo Cube is a dissonant experience. Opening your first pack full of 10s, 15s, and a 30 out of 10 Prophetic Prism is exciting, but not in the same way or the same degree as open up a real (and might I add much less impressive) Mox Sapphire. Turbo Prism is a more effective game-piece in the broad context, but it doesn’t really exist in that context. And it doesn’t have the allure, history, or recognition of real power that’s such a huge component of our emotional response to it.

The Turbo Cube fails to really incite the kind of reaction I intended, the intellectual realization of seeing the emperor with no clothes, because the experience it was trying to expose was never purely intellectual in the first place. But it does get people excited as they crack a pack and bend their expectations around degenerate new options.

Where the Turbo Cube is wild and exciting, Regular Cube is “regular” in just about every way. Everything is playable. Nothing sticks out too much. The sharper corners, where things have been too dominant or repetitive, have been chiseled off. The games don’t have the extremes of fast endings or protracted top-decking but follow a relatively predictable pattern. Even the basic lands are uniform, seamless, ergonomic. It’s successful in what it’s supposed to be. It’s useful as an intro to draft. It’s flexible in its homogeneity. It’s rewarding in its depth if you give it some time. It can sit on a shelf for years and come down when people are interested in playing some good clean games of Magic without needing the whole context of the latest constructed meta and newest draft sets. But what the Regular Cube really is not is playful or thrilling. It does just kind of live up to its name.

That might sound like the best option is somewhere between “comfortable regular” and “absurdist turbo”, but really I’m not sure there is a best option. People don’t just want variety, we want variety in our variety. Sometimes we crave the emotional extremes and other times want a more narrow, predictable experience, with smaller highs and lows, in a complex fractal emotional roller coaster.

There’s not a simple takeaway. In some ways I think the “Spike cube designers” really aren’t doing the design activity they think they are by just collecting all the “best in slot” cards. But that’s only according to my own definitions, however useful I might find them. They’re just pursuing what they feel is compelling, and they may be jumping straight to the experience they’re looking for while I’m laboring trying to derive it from first principles.

If anything stands out, it’s that the most important factor in creating a fun experience turns out to be just setting expectations right.

I tried to write parts of this article years ago: specifically the part about changing your perspective from the player to the designer, and how the numbers are all made up. An 11 isn’t objectively a better number than a 10¹⁰, all that matters is the way those numbers line up relative to others in the system. It turns out writing is hard and I had a difficult time communicating my point, and an even harder time doing it without being an asshole, telling people they’re doing it wrong. It’s probably good that I’m finally getting around to it now with a little more context and humility.

After all, it’s just a game. We shouldn’t take too seriously. Above all, we should be having fun with it. And that’s something I think we should be extremely serious about.

The community has changed drastically in the last half decade (with roots going back much further). Enough people have started going off the rails and doing their own thing for their own creative interests replacing a card with another with a bigger (or smaller) number is no longer the default mindset. I’m heartened by that. It’s not more correct, but for me it’s a lot more fun.