The idea in one picture
Draw a bar chart with card cost along the bottom and the number of cards at that cost up the side. That shape is the curve. A deck with many cheap cards and few expensive ones has a curve that leans left; one with mostly expensive cards leans right. The curve is a summary of when your deck wants to spend its resources.
In many card games, the amount of resource available grows by about one each turn. If you have two resources on turn two, you want a two-cost play; if you have three on turn three, you want a three-cost play. A deck whose cards line up with that growth uses almost everything it has on each turn. A deck whose cards do not line up leaves resources unused, or holds cards it cannot yet afford.
Why unused resources matter
Resources that are not spent in a turn usually disappear. In many games they do not carry over, which makes each turn’s unspent amount a permanent loss. Over a game, a deck that wastes one or two resources a turn in the early game falls behind a deck that does not, even if its cards are individually as strong.
This is what players mean by tempo: how much useful action you take relative to what you could have taken. A smooth curve gives you good tempo, because you use what you have, while a clunky one does not.
What a typical curve looks like
There is no universal ideal, but patterns are common. An aggressive deck might have many one-, two- and three-cost cards and only a few at five or above. A midrange deck might peak at three or four. A slower deck might run a lower count of cheap cards but include some that stabilise the board, then several expensive finishers.
The relationship between price and power is also designed. Cost curves in card and unit design are what designers adjust to ensure a five-cost card is meaningfully stronger than a two-cost one. As a deck builder, you can use that: a cheap card that is unusually strong is a bargain, and an expensive card that is only slightly stronger than a cheaper alternative is often a poor use of a slot.
The link to resource count
A deck’s curve determines how many resource cards it should run. A low curve needs fewer, because you rarely need to cast expensive cards. A high curve needs more, because you want to reach the high numbers on time.
A rough, commonly cited starting point is that a 60-card deck uses around 24 lands. With 24 lands in 60 cards, a seven-card hand contains on average 7 × 24 ÷ 60 = 2.8 lands. The expected value is not a number you will actually see in any one hand, since hands contain whole numbers of lands, but it is a useful reference. The point is not that 24 is correct for every deck, merely that the number should be consistent with the curve.
Resource cards also have a cost: each land in the deck is a slot not used for a spell. This is the central tension of the topic. More lands help you cast your spells on time and reduce the chance of stalling, but make flooding (drawing too many lands) more likely.
Curve and the opening hand
The shape of the curve determines which opening hands are good. A hand with two or three lands and a mix of cheap spells fits almost any curve. A hand with only expensive spells depends on drawing lands. This is why curve planning and mulligan decisions are tied together: a deck with a low curve can keep hands with fewer lands, since it needs fewer to function.
When the curve is deliberately broken
Not every deck wants a smooth curve. Some decks plan to do very little in the early game and then play something much larger than their opponent expects, sometimes using acceleration to cast it early. Others rely on a particular card at a particular cost and fill the rest of the list around it. A curve is a means to an end. Judge it by whether it supports the deck’s plan, not by whether it looks tidy.
The same applies to tempo as a concept. Gaining tempo is not always better than gaining card advantage; each has a price, and how they compare is discussed in tempo and value. A deck that gives up some efficiency can still be correct if the extra power it gains is worth the price.
Checking your own list
A simple check for a draft or constructed list:
- Count your cards at each cost.
- Ask what you will do on each of your first four turns.
- For each turn, count how many cards in the deck would let you do it.
- Look for a turn where the answer is nearly none.
If your turn-two answer is “very few,” you will often do nothing on that turn, and the fix is to swap in cheaper cards or accept the weakness knowingly. If you have many cards at the top of the curve, count how many resources you need to cast them and whether the deck can reliably reach that number.
Some caution is needed. Curves give you averages, but individual games vary, and draws are random. A well-built curve reduces the number of awkward games without removing them. In the same way that streaks occur in any random sequence, a few poor starts in a row do not necessarily show that the list is flawed, as the topic of random clumping explains.