I love Commander, but right now, I think I’m most excited about Cube. Recently, while inspired by another one of my favorite formats, Vintage Rotisserie Draft, as well as some offhand comments by good friends nearly a decade ago about doing exactly this, I built a carbon copy of the Magic Online Vintage Cube that I plan to maintain for the foreseeable future. It’s well-known, well-curated and produces fast, exciting games. Sure, sometimes you Bribery someone’s Archon of Cruelty, and sometimes you get hit with a Channeled Eldrazi Titan on turn one, but at least those games are over fast (yes, those were both personal experiences from the debut in-person draft of my copy of Vintage Cube).

Since I now have a finished product, I need a new project, and the more I think about it, the more I want that to be another Cube. I’ve been toying with the idea of Commander Cube for some time, and with Commander Legends: Battle for Baldur’s Gate still fresh in my memory, I think it’s time to make this happen. Before I start digging through boxes of cards, I need to generate a list, and I’m going to start that process today and share my thoughts with all of you. That way, I can give you a blueprint for how to build your own Commander Cube in the future, even if you disagree with some of my decisions at the micro or macro level.

First, let’s think about how we want the Cube to be drafted. Most Cubes are drafted with three 15-card packs per player, but Commander Cube probably shouldn’t work exactly like that. After the two Commander Legends releases, people have internalized Commander Limited as a 60-card format, and I think that’s the right call for a format that will feature more protracted multiplayer games as well as one that has variability as a feature. With that in mind, drafting only 45 cards is unlikely to be correct. I think drafting a total of 60 cards per player, just like in the Commander Legends sets, is the right idea, but exactly how should we do it?

Three packs of 20 cards each with two picks per pack is the default retail option, likely due to players being well-used to left-right-left three pack drafts, but that doesn’t mean it’s the only way. I’ve seen some folks talking about drafting four packs of 15 cards each (still with two picks per pack). My main thought, after drafting a lot of both Commander Legends sets, is that 20-card packs are pretty difficult to dissect in the moment without feeling like I’m taking forever, and I’m one of the faster drafters/players I know. I think less experienced players are more likely to tunnel vision when overwhelmed by a ton of cards, and especially in a Cube format with less filler and more killer than your average set of booster packs, I think 20-card packs are asking for trouble. Therefore, I’m borrowing an idea from other Commander Cubes I’ve seen and going with four packs of 15 cards each.   

When I Cube with nonstandard numbers of players, I borrow from the playbook of former ChannelFireball editor and cubemaster extraordinaire Andy Cooperfauss – with eight players, I use the standard three packs of 15 cards each, but with six players, I move to four packs of 12, and a four-player draft actually works just fine with five packs of nine. With a full eight players, I think three 20-card packs is fine if your group prefers that, but at six or even four players for Commander Cube I’d drop down to four 15-card packs right away. With four players I might even go with five packs of 12 – that’s an interesting experiment to file away for the future. 

Let’s check our assumptions on one thing quickly: why are we drafting two cards per pack? Well, a few reasons stand out to me. First is speed – we’re adding 15 cards to the draft pool above the normal amount, requiring players to build a larger deck with color identity restrictions in effect, and we’re probably going to be playing some lengthy games, so finding some ways to cut a little time is good. Second, I think it makes it easier to focus in on your color identity and even switch things up if necessary. Finally, it lets you grab two synergistic cards that happen to be in the same pack without trying to wheel one, and with synergies on both macro and micro levels being so important to the spirit of Commander, I think that’s a huge positive. If you recognize those first and third points in favor of two cards per pack, well, they’re both things Gavin Verhey called out when he introduced the first Commander Legends set back in 2020 – they’re not my own original arguments, but I’ve had nearly two years to think about those arguments, and I’ve come to agree with them.

Okay, so we’re settled on four packs of 15 cards each, for a total of 60 cards per player, with each player drafting two cards at a time (with the notable exception of the final card from each pack.) How many cards should be in this cube? 

The Magic Online Vintage Cube features 540 total cards, with 360 cards placed randomly into the pool in each draft. With a full third of the card pool absent from any given draft, you can’t count on particular cards showing up, but you can assume that supported archetypes will likely have enough cards present to succeed. If we wanted to mirror that structure, we could easily reverse engineer the numbers. Eight players drafting 60 cards each will end up with a total of 480 cards – if we wanted to leave a third of the card pool out, the Cube would need to be 720 cards in total. But is that a good idea? I think it doesn’t matter as much in a Commander Cube as it would in a 1v1 Cube setup, since you’re probably playing two separate games with four players each and therefore not having to interact with the exact same set of cards every time, but given how important I think variability is to Commander, I’d like to preserve some of the mystery. I’ll start at 720 and consider scaling higher if the Cube starts to feel stale.

One really important consideration in Commander Cube that isn’t part of the regular cube ecosystem is, obviously, the commanders. How many commanders should be in the Cube, and how should they be treated? 

Let’s start by talking about some of the special rules I’ve seen for Commanders in Commander Cubes while doing research online.

Single Color Commanders have partner

I am overall okay with the partner mechanic, but I think it’s a much better fit for Limited environments featuring actual sealed booster packs rather than Cubes. I’m also personally not a fan of Cube rules that flout the normal rules of Magic or change the text of cards. Magic is complex enough as it is, and this additional layer of complexity probably won’t help the experience of newer drafters. 

Commanders with partner have “Choose a Background” and vice versa

I currently plan to avoid adding partners and backgrounds into this Cube, so this is not relevant. Just a matter of preference.

“Partners with” pairs are represented by one card and cost one pick

I’m not deeply in love with any of the partner pairs to the degree that I’d want to introduce this level of complexity. 

The decision not to use the special rules above is all based on my personal preference – if I didn’t care about your preferences, I wouldn’t even introduce these ideas in this article! The more different concepts for Commander Cube you’re exposed to, the better you’ll be able to decide how to build your own. I don’t think doing or not doing any of these things is objectively right or wrong.

If we’re skipping adding partner to monocolored commanders, I think I’ll be leaning away from monocolored commanders. Sure, there will probably be the odd one-color legend in the list, but I’ll let the drafters know up front that they’re mostly intended as role players rather than commanders themselves. We’ll get into color distribution in a minute, though – first, let’s talk about how many commanders to put in the Cube.

I like the idea of each player seeing two commanders per pack. That number is largely based on expectations generated from playing Commander Legends sets – the number of available commanders felt good, and I’d like to try mirroring that. But didn’t we just say we were going to be doing four 15-card packs per player? We sure did, and that means players will therefore be opening eight commanders in total. I think having a higher density of commanders is good when we’re skipping partner – that way, players will have some more flexibility in their color choices. 

So how do we get the “as-fan” number of commanders per pack up to two with 720 cards? One option is simply to manually seed packs, which seems to be what a lot of people choose to do. What that would mean in this case is that, before drafting, we’d separate out eligible commanders from all of the remaining cards, shuffle each of those two sets of cards separately and then make packs with two random commanders and 13 random non-commanders each. This is a ton of work, but some people are happy to do it. Some folks even do this with regular Cubes in order to seed packs with a certain number of cards per color. I’m not super into this method, partially because I think it’s a lot work, and partially because I’d rather actually have some of the clumping that makes for tough decisions! 

Instead, I’d rather leave it up to chance – to some degree, anyway. When you’re not seeding packs, the key is simply to include enough commanders that people will see roughly two per pack. On a basic level, we are looking to solve for the variable “commanders” in the following equation:

Commanders / 720 * 15 = 2

A little algebra gets us to 96 commanders needed. For the sake of round numbers (with the mild benefit of erring very slightly higher), let’s go to 100. As usual, I loaded up the hypergeometric calculator and asked it to tell me the odds. With 100 commanders, each of our players should open at least eight commanders in their four 15-card packs about 61 percent of the time. What about a nightmare scenario where a player opens no commanders in their first two packs? How likely is that? Well, it’ll happen just over one percent of the time. They’ll certainly get passed some commanders, though, and the odds of opening literally zero across four packs are vanishingly small – .008 percent. 

Let’s get back to talking about color distribution. I want to avoid decks descending into “good-stuff” piles, and I also want to reduce the number of unplayable cards in the average pack, so I’d like to make two-color decks the norm, with three-color decks being a possibility. Four and five-color decks are right out for me, though if there is a clamor from my playgroup for something different, I’d consider a change. I still want to include a few monocolored commanders for people looking to try that route, though I’d make sure there was sufficient incentive to include them as pieces of the puzzle in two or three-color decks. With that in mind, I think the following distribution is reasonable:

• Monocolor commanders: four per color (20 total)
• Two-color commanders: six per color pair (60 total)
• Three-color commanders: two per color combination (20 total)

But what if the unthinkable happens? What if someone really wants to draft a particular color combination but gets totally wrecked during the process and ends up with no matching commander? Well, a lot of folks have worked out that having each player start with two copies of The Prismatic Piper in their card pool is an easy fix for this. You can play your chosen color pair, but your commander is a pair of silly 3/3s. I’m fine with that as an edge case fix, though I doubt it will come up often. 

Speaking of starting with cards in your card pool, a lot of Commander Cube curators seem to want to emulate the Commander experience by letting players start with important cards in their card pool – usually ones that help with mana issues. Other than The Prismatic Piper, I’ve seen Cubes where players get free copies of Command Tower, Arcane Signet, Commander’s Sphere and even Sol Ring. I understand where this urge comes from – having just one copy of any of these cards in a Cube feels a little weird, but having more than one copy might cause a player to scoop up multiples, and depending on how you feel about the singleton “rule” in Cube, that could be bad as well (multiple Sol Rings in one deck would just be bad all over.) As usual, you should do what works for you and your playgroup. I’m not super attached to the “getting wrecked by Sol Ring” experience, so I’d personally leave that one out entirely.

After thinking about what I want, I’m all the way out on starting with any particular card in your pool other than the emergency backup Pipers. If you want mana fixing, draft it – I’ll make sure there’s enough of it available in the Cube. That said, if players feel like it’s too hard to get their mana working, I’ll start with Command Tower and move to add Commander’s Sphere if that’s not enough. 

I’ve seen a few other special rules in some Commander Cubes, like easing restrictions on hybrid mana, letting monocolored commanders splash a color of their choice, letting planeswalkers be commanders and more. As you might have predicted, I’ll be avoiding adding or changing rules in this way, but as with so much of this, that’s just my preference. 

Most importantly, I’m willing to change my mind on most of this as I go through the eventual process of building and playing this Cube. One of the most important things as you build a Cube is being willing to take in new information, including the opinions of your playgroup, and thus being open to making changes that could improve the play experience. 

That’s all for now – next time we revisit this, we’ll get into discussions of themes, commanders, and hopefully specific card choices!