How many creatures should a Commander deck run?
Any page that gives one creature count for every Commander deck is describing one deck. The real answer depends on what your creatures are for. A deck that wins in combat, a deck whose engine triggers off creatures, and a spellslinger deck that needs a few blockers and a finisher are asking three different questions. Here is what each one depends on, with the exact odds behind the ranges.
Start with what your creatures are for
Creatures do three different jobs in a Commander deck, and most decks run them mainly for one:
- They are the win condition. The deck wins in combat, with a typal army, a wave of tokens, or one big commander.
- They are the fuel. An engine like Guardian Project or Beast Whisperer turns each creature into a card, so every creature you cut makes the engine worse.
- They are utility. Mana dorks, value creatures and blockers, in a deck that wins some other way.
The count follows the job. The first two groups need density, and the third only needs enough.
A creature payoff is only as good as the creature count
Beast Whisperer draws a card whenever you cast a creature spell, so it is worth exactly as many cards as the creatures you have left to cast. That makes its value roughly proportional to the count. By turn 8 you've seen 4.2 creatures on average in a 30-creature deck and 2.1 in a 15-creature deck, so the same payoff fires about half as often. Kindred Discovery and every other typal payoff work the same way, and typal decks usually also need several creatures out at once. Here is what each count delivers, on the play in a 99-card deck, using exact hypergeometrics (the same math as this site's calculators):
| Creatures in the 99 | ≥1 in opening hand | ≥3 drawn by turn 5 | Average seen by turn 8 |
|---|---|---|---|
| 10 | 53.7% | 8.0% | 1.4 |
| 15 | 69.6% | 21.7% | 2.1 |
| 20 | 80.5% | 38.9% | 2.8 |
| 25 | 87.9% | 56.1% | 3.5 |
| 30 | 92.8% | 70.9% | 4.2 |
| 35 | 95.8% | 82.2% | 4.9 |
A typal deck that wants three of its creatures in hand or on the battlefield by turn 5 gets there in about seven games in ten at 30 creatures, and barely more than half at 25. Those last five slots are the difference between a typal deck and a deck with a typal theme, which is why typal sits at the top of the ranges below.
Plan for two boards, because the pod will wipe one
Your creature count also has to answer to everyone else's removal. Suppose each of your three opponents runs the same number of board wipes. Here is the chance that at least one of them has drawn one by each turn. It counts every opponent as if they were on the play, the conservative case, so the real figures are a little higher:
| Wipes per opponent | Someone holds one by turn 6 | By turn 8 |
|---|---|---|
| 1 | 32.1% | 36.7% |
| 2 | 54.1% | 60.1% |
| 3 | 69.1% | 75.0% |
| 4 | 79.3% | 84.4% |
At two or three wipes each, the range the removal guide suggests for a midrange deck, someone at the table is holding a wipe by turn 8 in 60.1% to 75.0% of games. Holding one isn't casting it, and they'll pick their moment. Still, a creature deck should treat a mid-game wipe as the expected case rather than as bad luck.
That changes what the count is for. The creatures you commit before the wipe are gone, and the ones still in your hand and library are the rebuild. A deck that wins with creatures needs enough of them to field a board twice, or recursion and card draw that do the second job for it. It also argues for committing a few at a time rather than everything at once, since a held creature is the one a wipe can't touch.
A low count is a defensive choice too
The low end has its own logic. A spellslinger deck with ten creatures leaves the table's creature removal with nothing worth aiming at: Swords to Plowshares sits in an opponent's hand, and a board wipe costs that deck almost nothing. When your threats are spells, planeswalkers or enchantments, a low creature count decides which of your opponents' answers line up against you. What the deck gives up is defense, so the few creatures it does run should block, draw or finish, ideally two of the three.
Count bodies, not creature cards
Token decks break the usual count, because their bodies come from spells. A deck with 15 creature cards and a dozen cards that make tokens can field more bodies than a 25-creature deck, so count the token makers alongside the creatures. Check the payoffs before you rely on that, though. Guardian Project only counts nontoken creatures, and Beast Whisperer only counts creature spells you cast, so a token deck wants payoffs that trigger on any creature entering. The wipe table applies to tokens just the same.
Voltron, and the creature in the command zone
A voltron deck runs one creature that matters, and it never has to be drawn. The rest of the deck is equipment, auras, ramp and the protection that keeps the commander alive (that's the interaction budget), so its creature count is whatever utility those slots need. Any deck with a creature commander has a smaller version of the same thing: the command zone is one creature you always have, and it's worth counting before you count the 99.
The ranges, by what your creatures are for
| Deck | Creature cards | Why |
|---|---|---|
| Typal, or an engine that draws per creature | 30 to 38 | Payoffs scale with every creature, and 30 draws three by turn 5 in 70.9% of games |
| Creature midrange that wins in combat | 25 to 32 | Enough to rebuild after the wipe the pod is probably holding |
| Tokens | 12 to 22 | The bodies come from spells, so count token makers alongside creatures |
| Voltron | 12 to 20 | The commander is the threat and the rest of the deck protects it |
| Spellslinger or control | 6 to 15 | Blockers and finishers, and the table's creature removal stays dead in hand |
| Combo | No range | Whatever the combo needs, plus a few that buy time |
These are judgment calls anchored by the tables above, and the edges are genuinely arguable. A creature deck can win with fewer than 25 when its commander does the heavy lifting, and a control deck can run more than 15 when its creatures are its card draw. What the tables settle is the shape. A deck whose creatures are the plan needs density and a second board. A deck whose creatures are support needs only as many as their jobs require.
Check your deck's own numbers
Enter your creature count and how many board wipes you expect each opponent to run. It opens on 25 creatures against two wipes per opponent.
With 25 creatures you've drawn three by turn 5 in about 56% of games. If each opponent runs 2 board wipes, at least one of them is holding one by turn 8 in about 60%.
On the play, no mulligans. Opponents are each counted as if on the play, the conservative case, in a four-player game.
FAQ
How many creatures should a Commander deck run?
It depends on what the creatures are for. Typal decks and decks with a payoff that draws a card per creature want 30 to 38, creature midrange that wins in combat wants 25 to 32, and token decks want 12 to 22 creature cards because their bodies come from spells. Voltron decks run 12 to 20, and spellslinger or control decks as few as 6 to 15.
How many creatures does a typal deck need?
About 30 or more of the chosen type. Typal payoffs scale with every creature you draw, and the deck usually needs several out at once. At 30 creatures you have drawn at least three by turn 5 in 70.9% of games on the play, while at 25 it is 56.1%, so the last few slots buy a lot of consistency.
How many creatures does a spellslinger deck need?
Six to fifteen, chosen for what they do rather than to reach a count: blockers, finishers, and creatures that draw or copy spells. A low creature count also leaves your opponents' creature removal with nothing worth aiming at, and costs you almost nothing when someone casts a board wipe.
Do tokens count toward my creature count?
Count the cards that make tokens alongside your creature cards, because both put bodies on the battlefield. Check your payoffs before you rely on tokens, though. Guardian Project only counts nontoken creatures and Beast Whisperer only counts creature spells you cast, so a token deck needs payoffs that trigger on any creature entering.
Should I play fewer creatures if my playgroup runs lots of board wipes?
Not fewer, but plan on losing them. If each opponent runs three wipes, at least one of them has drawn one by turn 8 in about three games out of four. A creature deck should commit a few at a time, keep enough in hand and library to rebuild, and run recursion or card draw so a second board is possible. A deck that wins without creatures can lean into a lower count instead.