Building a Commander deck: the complete step-by-step guide from scratch

In my Commander guide, I promised to go deeper into deckbuilding. Here it is. Playing a precon is fun, but choosing which cards work together is what truly makes a deck yours.

This guide focuses on multiplayer Commander with room to build a strategy. Discuss in advance with your group which tempo, budget, and type of endgame you enjoy. Casual does not automatically mean every game takes eight turns. Also, high-power and cEDH are not the same: in the official bracket system, bracket 4 is called Optimized and bracket 5 is cEDH. You can read more about that in our guide to Commander Brackets.

Step 1 — Choose your commander and define your plan

Start with a commander that appeals to you and whose text you understand. Sticking to one or two colors keeps your mana base straightforward for your first deck. You don't have to choose the strongest or most popular commander right away: it is more important that you are excited to execute the associated game plan.

Write that plan down in one sentence. For example: “I create creature tokens, build up a wide board, and win with an empowered attack.” That provides much more direction than “I’m playing good green and white cards.” In the guide on deck archetypes, you will find other playstyles.

Also, decide how you want to win. Multiple cards that support the same endgame plan give you backup if one is removed. You don't need to put three completely different strategies into one deck to achieve this.

Step 2 — Distribute your 100 cards

With this starting point, you can start collecting cards without getting lost immediately. The numbers below are our practical initial distribution for a first deck, not official rules and not a mandatory recipe for every commander.

Card slots and functions are different counts. A creature that provides extra cards can fit into both your creature plan and your draw engine. Give it one primary role for the total count and use multiple tags in your deckbuilder if necessary. This allows you to see the overlap without accidentally building a 108-card deck.

Step 3 — Build the foundations

Lands: start around 37–38 and test your land drops

For many first decks, 37–38 is a usable starting point. An expensive deck or a strategy with extra land drops might need more. A low curve and plenty of cheap card selection can support a lower number, but don’t lower it just because you want to add another fun spell.

The probabilities below have been recalculated for seven random cards from a library of 99 cards. No mulligan has been taken, and no extra cards have been drawn.

Calculation: hypergeometric distribution, without replacement. Only the number of lands counts; colors, entering tapped, and the rest of your hand were not factored in.

The last column is intentionally labeled “0–1 land” and not “unplayable.” A one-land hand with cheap mana and draw can sometimes work; a hand with four lands in the wrong colors can be difficult. Use the numbers to understand the risk of running few lands, not as an automatic keep rule.

In multiplayer Commander, your first mulligan is free. For subsequent mulligans, you draw seven again and, if you keep the hand, you put one card on the bottom of your library for every mulligan after that first free one. In standard multiplayer Commander, the starting player also draws a card in the first draw step.

MDFCs: one card with a choice

Some modal double-faced cards have a spell on one side and a land on the other. You choose which side to play. That makes them flexible, but you don't get both functions at once: if you played the card as a land, you no longer have that same card as a spell in your hand.

With Shatterskull Smashing, for example, you can choose the land Shatterskull, the Hammer Pass. It enters tapped unless you pay 3 life. You can count such a card as a land option in your deck planning, but keep color, tempo, and how often you actually want to keep the spell in mind. Not every MDFC has a land side. For things like searching your library, the front side determines the characteristics; therefore, a spell with a land back is not simply a land you can find with a land tutor.

Ramp: start with about 10 cards

Ramp helps you use more mana than your regular land drop per turn provides. Preferably, choose a significant portion that you can play early. A two-mana mana rock can produce four mana on turn three with three land drops, provided it is usable then and makes the necessary colors.

Llanowar Elves is cheap, but as a creature, it is normally affected by summoning sickness. Arcane Signet can produce mana as soon as it hits the battlefield. Cultivate searches for two basic lands: one goes to the battlefield tapped, the other to your hand for a future land drop.

A ritual like Dark Ritual gives temporary mana. While it is mana acceleration, it fulfills a different role than a permanent source. For this starting template, we primarily count sources that also help you later. Ramp can still be useful if you miss a land drop; having enough lands primarily makes it easier to play your ramp on time.

Probability of seeing at least one card labeled ramp, not necessarily being able to play it. Seven cards is the opening hand; nine cards is the opening hand plus two normal draw steps in multiplayer, without mulligan or extra draw.

Card draw: combine repetition with immediate replenishment

Start with about 10 cards that replenish your hand or playable options. Many decks aim for 12; a commander who reliably generates cards can take on part of that task. Do ensure your deck can keep playing if that commander is removed.

A cantrip that only replaces itself is primarily card selection, not net card advantage. Returning cards from your graveyard and being able to play cards from exile can provide extra options. Self-mill only helps if your deck can do something with the cards in your graveyard.

Beast Whisperer rewards you for casting creature spells. Creating tokens doesn't trigger him, so he fits better with actual creature cards than with a deck that creates tokens primarily with instants and sorceries. Return of the Wildspeaker can draw cards all at once based on your largest non-Human creature, OR empower your non-Human creatures. Here, too, you choose one mode.

Repeatable draw is attractive in long games but is not automatically better than “draw three cards” at the right moment. Combine a few engines with direct replenishment, so you can rebuild after a board wipe. With cards like Sylvan Library, you must also read the conditions: keeping the extra cards can cost life.

Targeted answers: about 10, with different goals

You want to be able to do something against a dangerous commander, an annoying enchantment, or a decisive spell. Therefore, distribute your answers across creatures, artifacts, enchantments, and, if necessary, graveyards. Counterspells, bounce, and exile solve different problems; playing only ten creature-removal spells leaves gaps.

Swords to Plowshares exiles a creature and gives the controller life. Generous Gift can destroy any type of permanent but gives the controller a 3/3 Elephant. A broad answer sometimes has such a trade-off. Bounce usually removes a permanent temporarily; the owner can play it again later.

Ensure that multiple answers operate at instant speed so you can intervene outside your own turn. When you can do that, you can read in the explanation about the stack and priority. Don't hold onto everything indefinitely, but don't automatically use your removal on the first creature that appears, either.

Board wipes: start with 3 and look at your own board

A board wipe clears many things at once. Wrath of God, for example, destroys all creatures and prevents regeneration; indestructible creatures normally survive that. Other wipes exile, give negative stats, or return cards to the hand. So, read what kind of problem you are solving with it.

Three is a starting point. A controlling deck might want more resets; a creature deck might choose fewer or effects that spare their own board. An aristocrats deck sometimes benefits from deaths, but that doesn't mean five to seven wipes are automatically good. You also need creatures to rebuild your engine.

Keep protection and disruption separate in your mind. Teferi’s Protection protects you and your permanents, but it does not remove the threat. Graveyard hate like Bojuka Bog handles a different zone. Thalia, Heretic Cathar specifically slows down creatures and nonbasic lands of opponents, not all their permanents.

Step 4 — Choose cards that support the same plan

With the starting distribution, 28 slots remain for your strategy and protection. Your ramp, draw, and removal can also contribute to that plan. Look for exactly that overlap, as long as the cards continue to perform their support duties reliably.

Example: from tokens to an attack

Raise the Alarm creates two creature tokens. Trostani, Selesnya’s Voice can create a copy of a creature token you already have using populate. Overrun gives your creatures +3/+3 and trample temporarily to turn a wide board into attack power.

This is a simple build example, not a complete decklist or a guarantee of victory. Overrun is much less impressive without creatures. Trostani needs an existing token to populate. That is exactly what you should look for when building: do you have enough cards that get things going, or primarily cards that are only good once you are already ahead?

Adjust the ratio to your playstyle

Step 5 — Check your mana curve and colors

The mana curve shows how many spells you have at each mana value. Ensure you have meaningful choices early on. If your first real action usually costs five mana, the rest of your deck and your playgroup need to provide room for that.

You don't need to hit an exact number of one-, two-, or three-drops. Look at your first four turns instead: can you build mana, do something toward your plan, and potentially hold up mana for an answer? An average mana value says less than whether you can actually play your cards on time. X-spells, alternative costs, and your commander can skew that average.

A good mana base is more than just a land count

• Count color sources. A deck with early green ramp needs green mana before that ramp can solve your other colors.
• Pay attention to double symbols. Trostani costs two green and two white mana. Four mana in total is therefore not enough if the distribution is off.
• Limit tempo loss. Many lands that enter tapped can be affordable, but they make early turns slower.
• Be frugal with colorless utility lands. A fun ability helps little if it prevents you from casting your colored spells.
• Check your search cards. Cultivate only finds basic lands; other searchers have different limitations.

Set a budget in advance. Opt for a working plan and reliable mana first. You don't have to automatically copy expensive staples from an online list. Capture your concept in Moxfield or Archidekt, for example; use Scryfall to compare card texts and alternatives.

Step 6 — Testing, goldfishing, and tweaking

Goldfishing means playing through the deck without opponents. This lets you test the sequence of your actions, land drops, and colors. It doesn't tell you yet if you can withstand removal or how your deck performs at a real table.

1. Test ten opening hands, for example. Note how many you would keep and why. That's a first impression, not conclusive statistical evidence.
2. Play out a number of starts for four to six turns. Note missed land drops, missing colors, and expensive cards that remain unplayed.
3. Test without your commander as well. Can you still set up or draw cards if they are removed twice?
4. Play real games. Look at recovery, interaction, and whether you can actually close out your game.
5. Change a small group of cards at a time. This way, you better see what a modification achieves.

What if your deck isn't running well?

The most important thing remains that you and others have fun in the game. Align expectations using the unwritten rules of the table. A deck doesn't have to be fully optimized to be enjoyable to play.

Frequently Asked Questions

How many creatures should my Commander deck have?

That depends on your plan. A creature deck needs many, a spellslinger deck may need few. Furthermore, token creators aren't always creature cards. Therefore, count functions first and then see if you have enough creature spells or creatures on the board for the cards that depend on them.

Can my commander count as ramp or card advantage?

Yes, for your deck's functions. For card slots, it counts once as the commander. See how early and reliably it fulfills that function and keep alternatives on hand for removal and increasing commander tax.

Should I always play Sol Ring and other staples?

No. A popular card can be strong, but your plan, colors, and table agreements remain leading. Also check if a card is on the current Game Changers list when playing with brackets; Commander-legal doesn't automatically mean it fits every bracket.

Can I start with a precon?

Certainly. A precon gives you a complete foundation to learn which cards you like. Check out our Commander precon buying guide. The same checks for mana, draw, interaction, and game plan help with upgrading afterward.

Watch the deckbuilding template from The Command Zone

In this video, The Command Zone shows how to build a Commander deck using categories. It is a useful additional approach. The numbers in this guide are our own practical starting distribution; there is no mandatory universal Commander template.

Watch the video on YouTube.

Want to spar about a commander or your first list? Feel free to send me a message. Commander is my favorite format, so I'm happy to think along.