The Cube Has Six Faces

Posted in Feature on April 30, 2012

By Tom LaPille

Tom LaPille makes things. Some of the things he makes are card sets, like Dark Ascension and Born of the Gods. Sometimes he makes stories, too. Sometimes he makes unexpected things, like 16th-century Japanese clothing. He's probably making something right now.

Hello, my name is Tom. You may remember me from such sets as Magic 2012 and Dark Ascension, or such columns as Latest Developments. Before I was a real Magic developer, though, I learned a lot about Magic development while working to popularize Cube draft. I wrote about how my work on Cube drafting got me my job in a Latest Developments column, and you can still find some of my original work on Cube here. Given my past, I hope you can imagine my glee when I found out I would have the opportunity to be the lead designer on the first release of Cube draft on Magic Online.

Several people have picked up the Cube torch and ran with it since I joined Wizards, and it's become more popular than I ever imagined it would when I was inside the movement. However, it's still something of a niche activity. If you've never Cubed before, here's a quick overview. A Cube is a collection of Magic cards, usually sleeved, with enough similarly sleeved basic lands to use with them. Traditionally, Cubes contain no more than one of any given card and are built with some kind of theme in mind. The most common theme I've run into is the "powerful cards I like from Constructed" theme. You can play any kind of Limited format you want with a Cube. The most common mode of play is Booster Draft, but I've played Winston draft, Winchester draft, Sealed Deck, and even three-on-three team Rochester draft with my paper Cube before.

A few weekends ago, Cube booster draft made its debut on Magic Online. Our Cube was true to tradition and went with the "powerful cards we like from constructed" theme. You can see the list of the current Magic Online Cube in the announcement Arcana.

I've been updating my physical Cube since GenCon 2006, and in the process I've learned a lot about what makes Cube work. I've also learned a lot about what makes different Cubes work differently. Almost any wisdom we've ever shared about designing or developing normal Magic sets applies, so I would suggest that potential Cube architects go on a deep dive through our design and development columns if they want to learn about how to make their Cubes better. However, I realized early on that the Magic Online Cube wanted to be quite different from any paper Cube I had ever encountered in a few important ways, and I'll talk about some of them here.


Cube designers often feel pressure to make their Cubes as small as possible. A significant part of the point of playing with the kind of Cube we decided to make is that you get to play with all of the sweet cards you like from all of Magic's history. The bigger the Cube is, the less often you see each individual card, so you want to keep the thing fit and trim so all the fun cards you like show up often enough.

Snapcaster Mage | Art by Volkan Baga

I started my Cube at 360 cards, which meant that in an eight-player draft every card was always in play. That's as high-concentration as you can get. On the other hand, my Cube was at 720 cards in the winter of 2007, and that was about as big as I felt I could get while keeping the frequency of the cards I liked high enough. It's possible you could go bigger than that now, as we have five more years of cards since then, but I haven't tried and I'm not sure I'd want to.

On the other hand, draft requires variety to be fun for long periods of time. In normal Magic, we've found that 100 commons is about as many as we need for a set to have a long enough draft lifespan when drafted by itself. We don't normally ask you to draft three packs of a small set because we know that's not enough variety. In fact, during the leadup to changing the order of multiple-set draft, we tried drafting DKA-DKA-ISD instead of DKA-ISD-ISD, and we found that there still just wasn't enough card variety in that model with the current set sizes.

Over the time I've been interacting with Cubes, I've found that 540 cards—enough for two simultaneous three-on-three drafts—is about the minimum amount that satisfies me if I'm playing once a week or less. When I was drafting multiple times per week at 540 cards, though, I started to get bored. That prompted the move to 720, which was great until I got hired at Wizards and Cubed less frequently again. My physical Cube now sits happily at 540.

I still remember what repeated drafting on Magic Online feels like and I knew people were likely to Cube draft many, many times in a weekend when Cube was available. Therefore, I went with the biggest size I felt was also small enough for the cards we all love to show up enough.


When I learned I would be designing a Cube for Magic Online, the first thing I got excited about was that Cube was going to be on Magic Online. The second thing I got excited about was being able to include sideboard cards.

Ancient Grudge | Art by Ryan Yee

Both as a player and a developer, I enjoy sideboarding in Limited. As a player, I love changing my deck around to make it stronger against a particular opponent. It makes me feel smart, it gives me more options, and it makes Magic a deeper game for me. As a developer, I love that I can put cards like Ancient Grudge, Ray of Revelation, or Mask of Avacyn into sets that are especially powerful against particular kinds of decks without feeling the pressure to make them strong enough to usually play main deck. These put nice checks on strategies that use tons of Equipment, Pacifism effects, or spot removal without requiring me to make a bunch of overpowered cards. All in all, sideboard cards in Limited make life better.

Ancient Grudge
Ray of Revelation

Unfortunately, people aren't great about remembering their sideboards while they're Cube drafting in paper. I've found that most people just leave them around, which causes issues when they either remember they want to sideboard and don't know where their sideboard is, or when you are playing in a public place. I eventually reverted to just taking everyone's unplayed cards after deck building for security reasons.

I still wanted people to be able to kill problematic permanents, though, so, I included cards like Wickerbough Elder, Ravenous Baboons, and Indrik Stomphowler that you could legitimately main deck. This was fine, but it was no substitute for actual sideboards. I dreamt for years about having cards like Red Elemental Blast and Deathmark in my Cube, but I knew I could never make them work. Magic Online, however, is the perfect place for them.

Celestial Purge
Vulshok Refugee

The cool thing about sideboard cards in Cube on Magic Online is that they are very low-cost to include in the Cube and to draft. The cards are all so good that you can almost always find a reasonable twenty-third card, so taking a Celestial Purge or a Vulshok Refugee early doesn't cost you nearly as much as it might in a normal draft. However, when you get to sideboard one of those in, you get rewarded with a very powerful card in the matchup.

Niche Decks

Another consequence of Magic Online's much faster draft repeat rate is that we needed to include more niche strategies. The Cube that the 2007 Magic Invitational used included a few pairs of cards that only worked with each other—most notably Illusions of Grandeur and Donate. This makes sense when the goal is to make the most compelling theater, but it's not the best for repeated play—especially when those cards have no use aside from each other. However, I also knew we needed to come closer to this kind of idea than most physical Cubes do, as people would be drafting it over and over again. Keeping the experience satisfying requires deep fractal complexity, and a few narrow cards provide that in spades.

My design file included a few whole-deck themes, like mono-black. It also included several smaller and more portable themes, like reanimation and the Wildfire effects. Max McCall, who led the development of the list, did me one better by finding some individual cards that are highly specific but still open-ended. I think my favorite of those is Loyal Retainers. In general, the card is quite specific, but it still works with every legendary creature in the Cube—including the Eldrazi. It's also quite saucy in particular with Survival of the Fittest and Fauna Shaman. I enjoy these emergent strong interactions between particular cards much more than the kind of engineered specificity that things like Illusions-Donate create, and Max did lots of good work here.

Fauna Shaman

Not all of my themes survived, though. One of the ones that regrettably did not make it was artifacts. My handoff contained Metalworker, both Tezzerets, Master of Etherium, and all the cards you expect to go with them. Unfortunately, that never worked out in practice because of how big this Cube was. Exactly one player forced that deck in each of the playtests that included it, and none of those players ended up with reasonable decks. I was sad to see it go. I'm fairly certain that in a smaller Cube it would have been fun, but we had to be big, and that meant artifacts had to go.

Max and Zac Hill suggested replacing it with the small storm theme that you see in the Cube now. They pitched the deck to me as a midrange sort of deck that works similarly to how storm decks worked in Time Spiral Limited. That deck doesn't kill people on turn three, but it can make twelve Goblins and drain them for 10 on turn four. That sounded cool to me, so I let Max do his work. Both of them have drafted it successfully in other Cubes, although I haven't done it myself. Your mileage may vary, but at least one person managed to do this last weekend, as evidenced by this deck-building screenshot of the deck that went 3–0, according to its pilot.

Mana Fixing

The last bit of the design I've received questions about is the mana fixing. I went with all ten fetchlands, all ten original dual lands, all ten Ravnica dual lands, all ten Magic 2010-style dual lands, and the five Worldwake "manlands." I didn't include Ravnica block's Signets or "karoos" because both of those cycles have the opposite effect on Limited gameplay that I wanted. I like for green to have the best mana acceleration, and I also like for control decks to have to prioritize taking artifact mana acceleration if they desire it rather than letting late Signets float to them. The "karoos," on the other hand, are so strong that it's often correct to take one even if you only have cards of one of its colors. I wanted people to take lands because of the actual spells they had, not play more colors because they took a very strong land.

Glacial Fortress

I chose to put the total amount of mana fixing where it is because I wanted color commitments to mean something. I found when I had a higher ratio of mana fixers to total cards, strong players could take cards of tons of colors early, then table all the mana fixing they needed past weaker players who were drafting one or two color decks. This left the strong players with all the best cards and good mana on top of it. When I used a lower amount of mana fixing, the strong players who went for greedy mana bases couldn't do that anywhere near as much, so that keeps everyone honest. In general, I like when the natural course of action is close to correct, and most Magic Online drafters are used to playing two colors with a possible splash. Cube tends to be a bit more multicolor-friendly than normal Limited, so I hoped to make two colors with a splash of a third the norm. Based on what I've seen, that's where we ended up.

The Card Selection

The actual design process was fairly simple. I've been happy with how my own Cube plays for a year or so, and I just scaled it up by 33% by including more cards that help the themes I already included. When I ran out of strong support for the themes I wanted, I put in other niche cards or sideboard cards. This is, for example, how I fit in the original artifact theme.

Fiend Hunter | Art by Wayne Reynolds

The other cool thing I got to do is apply some of our modern Limited development methodology to a Cube for the first time. We have developed guidelines about the percentage of opened cards of each color that we want to be creatures, noncreatures, removal spells, and a few other categories, as well as some guidelines about each color's mana curve. The numbers I naturally had weren't too far off of those guidelines, but I did some adjusting to make them closer. Just having the skeleton framework gets things close enough for development to take it the rest of the way.

The Handoff

I spent most of my time in Magic R&D on development teams, so it was a little weird to hand my work off to Max McCall for development. There were times Max did things that scared me, and I got a taste of what it's like to be on the other side of the process. Max changed some individual cards, but looking back, most of them are a matter of personal preference and I can't say with any certainty that the cards I liked were any better than the cards he liked. He was also nice enough to come back to me near the end of the process to check if he had removed any cards I felt especially strongly about. In the end, though, I stand behind about 98% of the choices in the final Cube, and that's a much higher number than most designers get.

I hope you all had fun with Magic Online's Cube debut. It was an honor to be part of its creation, and a dream come true to watch people playing my favorite Limited format on Magic Online.

Latest Feature Articles


September 20, 2022

(Almost) Everything to Know About Unfinity Boosters! by, Mike Turian and Adam Styborski

Welcome to the lights and excitement of Unfinity! Magic's latest Un- set release is packed with attractions (and Attractions) to delight players—from the fresh-faced thrill seekers ready ...

Learn More


September 20, 2022

Unfinity Mechanics by, Matt Tabak

Welcome to Myra the Magnificent's Intergalactic Astrotorium of Fun! It's the galaxy's most awesome entertainment complex. For just a few credits, or whatever currency your planet uses, th...

Learn More



Feature Archive

Consult the archives for more articles!

See All