Instead of having a “Keep or Mulligan” article this week, I’ve decided to write a bit about mulligan theory in general.

Mulligans are, as we all know, a very inexact science. There is some math you can do, and there are some right or wrong answers, but a lot of it is a big grey area that comes from experience and gut feeling. That said, over my many years of MTG, there are a couple of misconceptions that I’ve seen repeated over and over, even by competitive players. In my head, I think of these as the “mulligan traps”—hands that look like they should be kept, but that for one reason or another just aren’t good enough. Today I want to talk about four of those traps.

Keep in mind that my goal here (as well as with the Keep or Mulligan series) is not necessarily to argue any particular example, but to convey the idea behind it. If you disagree with my final decision but you understand why I’m writing the things I’m writing, and you think you can apply it to future hands, then that’s fine.

Trap #1: “I just need one more land”

The most common misconception I see in mulliganing are the one-land hands that need to draw one more. Now, I don’t have a problem with keeping lands that need to hit a land to get there—I’m generally a risk-averse person, but I still understand the basics of probability. The problem I have is that, in general, the statement “I need one more land” is just not true—you need one more land to not lose the game immediately, but you need way more than that to actually win.

Imagine a hand of land, 2-drop, 2-drop, 3-drop, 5-drop, 5-drop. This hand will need one land to operate in the early game, but unless one of your 2-drops is a Stoneforge Mystic, you are not going to win unless you draw several more in a reasonable time frame (four more, to be exact). When doing math in their heads, people will usually figure out a rough estimate of how likely they are to find a land, but they will assume that this is strongly correlated to their possibilities of winning the game, and I don’t think you can do that.

In “Keep or Mulligan week 4,” I had two similar hands:

(vs. Zoo)

I said I would mulligan the first hand, and I would keep the second one. A lot of people asked me why I thought the hands were different, since they both need to draw a land fairly quickly or they would lose on the spot.

To me, the difference is that in the draft hand I’m taking a huge risk for not that big of a reward. You need one land to play a fair game of Magic—the downside is you die without playing a spell, and the upside is you have a decent curve. You could calculate the odds of hitting immediately—roughly 75%—but this does not tell you enough about your chances of winning the actual game because you still need more things to break your way.

If you don’t hit land, you lose, but if you hit land, you don’t necessarily win—you need several lands, and then the upside is, again, you play a fair game of Magic. If you hit and then brick twice, are you actually in a good spot? If you hit twice but they’re both Mountains, is that enough?

Now, the Modern hand is very different. With this hand, when I say you need a land, I actually mean that. If you draw one land, you’re set—your hand is great and you don’t need anything else. There is great reward in this hand, so I think it’s worth the risk.

The moral of the story is: when you think to yourself “I only need a land,” make sure you only need a land to turn your hand into something great, instead of only needing one land to not die immediately. If you don’t die immediately but fall too far behind in the early game, you’ll still probably lose the game regardless—playing an off-color morph when the opponent just played a 5-drop is probably not going to win you the game, even if you did have a 2/1 on turn two.

Trap #2: “It’s already a mulligan”

Sometimes, people will say that their mulligan is free because their hand is “already mulliganed anyway.” This is a flawed perception.

People will think their hand has already mulliganed for two reasons: The first is that they have a card they won’t be able to cast for a very long time. For example, a second Dig Through Time, or an Ugin in a hand with two lands. Those cards are blanks in the short term, yes, but they still have value over the course of the game. I would choose to have any other card in my deck over Ugin in my opening hand, but I’d still prefer having Ugin over nothing, because some games I’ll get to 8 mana and then that card will be great. Mulliganing away your Ugin is not “free.”

The second reason is that they have a card that will actually never do anything. If I’m playing the Esper mirror and I have Ultimate Price in my hand, then that’s a mulligan—it does nothing and it will always do nothing. If I have two copies of Hammer of Purphoros in my hand against a mono-red deck, then that’s a mulligan, since the next copy will never do anything. In those spots, the “it’s already a mulligan” argument is more reasonable, but still flawed, because it fails to take into account that you can just draw those cards again.

Imagine I have a hand of 6 cards + Ultimate Price in the Esper mirror—that’s effectively a 6-card hand. But it’s effectively a 6-card hand without Ultimate Price. What if I mulligan to 6 because a mulligan is “free,” and then I draw Ultimate Price again? Now I’m at effectively 5 and my mulligan cost me a lot. Even if I don’t draw it immediately, it could be my third or fourth draw step. There is marginal value in having that Ultimate Price in my hand as opposed to shuffled it into my deck.

The moral of the story here is: it’s true that some hands are “already mulliganed,” but that does not mean that taking a mulligan is free, because your 6-card hand could just contain the same type of card.

Trap #3: “Lands and spells”

It’s very easy to identify that you should mulligan a spell-heavy hand—almost all zero-landers and most one-landers will be thrown back by the majority of the Magic population. It’s also easy to identify that you should mulligan land-heavy hands—all seven-landers and most six-landers are going to be evidently not good enough. The problem arises when you have a mix of lands and spells, creating the illusion that the hand is good when it actually isn’t.

One of the main issues with competitive Magic is that, sometimes, you just don’t get to play. It sucks to be stuck on one or two lands, and it sucks to draw ten lands and two spells. We want to avoid that. When we see a hand with three or four lands and three or four spells, we know we dodged a bullet and we’re relieved. I think our fear of having no-landers clouds our judgment a little bit, and leads to keeping hands that shouldn’t be kept, solely because they have both lands and spells.

It’s the case with the most controversial hand from the last Keep or Mulligan:

Being on the draw against an aggressive GW deck, I don’t think you should keep this kind of hand, because the potential to be overwhelmed before you can cast your spells is way too high. You have lands and you have spells, and your spells aren’t even bad, but you need some sort of early action and this hand doesn’t have any.

I think people in general overestimate the amount of time they have to do something. Sure, Limited is slower and you’re not going to be dead by turn five, but neither of your cards are very good at playing catch-up, so if you fall behind too much, you might as well have lost the game already, even if you get to cast them eventually. Against an aggressive GW deck, it’s very easy to fall behind too early.

The moral of the story here is that a hand might have a healthy mix of lands and spells, but if none of those spells are the right spells, then it’s OK to mulligan it.

Trap #4: “If I go to 5 I lose anyway”

I’m going to be honest with you—every time my opponent keeps 7 and I have to mulligan down to 5, I feel like I’ve already lost the game. I feel like, at this point, playing the game is just a formality. But I don’t give up, because I know that my feelings do not correspond to reality.

Mulliganing to five is awful, there’s no denying that, but it’s not a game loss. You absolutely can win on a mull to 5, and people assume they can’t, which leads to them keeping all sorts of bad hands on 6 on the assumption that “five-card hands won’t win anyway so I might as well try with this six.” I think people mulligan a lower percentage of hands than they should mulligan on 6 because they underestimate how much a five-card hand can actually win (it’s not a lot, mind you, but it is a real number).

Sometimes your deck has a certain card that is so powerful and key to your strategy that it’s worth mulliganing more to try finding it. It’s the case with Bitterblossom in Faeries—most 5-card hands with Bitterblossom were better than 6-card hands without it, so the correct play was to mulligan all bad 6s. The same was true with Stoneforge Mystic in Caw-Blade, or with any sideboarded game where you have a card that is very hard to beat (such as Leylines).

At the last Legacy GP I played, I was playing game two of the Miracles mirror and I found myself with a bad 6-card hand. I figured that it would be better to search for my sideboarded Stoneforge Mystics, so I went down to five cards looking for them. I didn’t find any, and, instead of keeping a bad 5, I decided to go down to 4, because I knew I had to find a way to win the game, and my 5- and 6-card hands didn’t look like they were up for the job unless they topdecked a sequence of lands and Stoneforge Mystic. If I require my top two cards to be land + Stoneforge, then I might as well mulligan and give myself four extra chances to hit those. On 4, I found one land and a Stoneforge Mystic, which I resolved, and which singlehandedly won the game.

The moral of the story here is that 5-card hands are bad, yes, but they aren’t as bad as they feel, and sometimes it’s preferable to go to 5 rather than keep a bad 6. If you have a specific card in your deck that is very powerful, then that makes it even more likely that mulliganing to 5 is right.

[Editor’s Note: The hand in trap #3 originally mistakenly pictured Soul Summons instead of Ojutai’s Summons.]