More games of Magic are decided before the first spell is cast than most players want to admit. The mulligan is the highest-leverage decision you make all game, because it is the only one where you get to reshape your resources for free. Yet it is also the decision players think about the least, usually resolving it with a gut feeling about whether a hand “looks fine.” This article replaces that gut feeling with a framework.

The London Mulligan, briefly

Since 2019, Magic uses the London Mulligan. You always draw a fresh seven, and each time you mulligan you draw seven again, then put a number of cards on the bottom of your library equal to the number of mulligans you have taken. Mulligan once, keep, and you bottom one card. Mulligan twice, and you bottom two.

The practical consequence is enormous: you are digging through seven fresh cards every time, so the odds of finding a specific effect improve sharply with each mulligan, and you only pay for what you keep. This makes aggressive mulliganing correct far more often than under the old rules. A hand you would have grudgingly kept in 2018 is often a snap mulligan today.

The two questions every hand must answer

Before anything else, ask two things:

1. Can this hand cast its spells? This is the mana question, and it is the one players misjudge most. A hand with the wrong number of lands, or the right number in the wrong colors, is a mulligan no matter how exciting the spells are.
2. Does this hand do something my deck wants to do? A keepable mana base attached to cards that do not advance your plan is still a weak keep.

A hand needs a yes to both. One without the other is usually a ship.

The mana filter

Start with lands, because mana failures are the least recoverable. For most 60-card decks, the safe keep range is two to five lands, with three being the comfortable middle. The edges are where judgment lives:

• Two lands: keepable if your curve is low and at least one spell is castable now or next turn, ideally with cheap card selection to find the third land. Two lands with a hand full of four-drops is a trap.
• One land: almost always a mulligan. Even with a one-drop cantrip, you are betting the game on drawing lands, and the math is against you. The land count and colored source odds show just how against you it is.
• Six or seven lands: a mulligan more often than people think. A hand that is nearly all lands does not pressure the opponent or defend you; you will draw into spells, but you spend the early turns doing nothing.

Then check colors. A two-color hand that can only produce one of them, holding spells of the other color, is functionally screwed even at three lands. Count sources, not just lands.

The plan filter

Once a hand can cast its spells, ask what it actually does. The useful mental exercise is to play the hand forward three or four turns in your head, assuming average draws. Does it establish a board, protect a plan, or pressure the opponent? Or does it durdle, playing lands and reactive cards while the opponent builds toward something?

Match the standard to your deck’s role. An aggressive deck needs a hand that applies pressure early; a slow, controlling hand from an aggro deck is a mulligan even if the mana is perfect. A control deck needs a hand with interaction and a path to the late game; a fistful of threats and no answers may be a ship. The single most common keep mistake is holding a hand that is fine in the abstract but does not do the thing your specific deck is built to do.

Play, draw, and matchup

Good mulligan decisions are not made in a vacuum:

• On the draw, you can keep slightly greedier hands, because the extra card improves your odds of hitting land drops and finding action.
• On the play, tempo matters more, so lean toward hands that use your mana efficiently from turn one.
• Against a known matchup, the standard shifts. A hand with no early interaction might be a fine keep against a slow control deck and an automatic mulligan against a turn-three-kill aggro deck. If you know what you are racing, keep hands that win that specific race.

The mistakes that cost the most

Three errors show up again and again, and all of them are expensive:

• Keeping two-landers on hope. Players fall in love with the spells and keep a two-land hand with no way to find the third. The London rule made this worse, because people remember the old greedy keeps working.
• Shipping keepable hands out of boredom. A hand does not need to be exciting to win. A functional three-land hand that casts on-curve spells is a keep, even if it lacks fireworks.
• Ignoring colors. Counting to three lands and stopping there, without checking whether those lands cast the cards in hand, loses more games than any flashy misplay.

The short version

Filter every opening hand through two gates in order: can it cast its spells, and does it do what your deck wants. Use the two-to-five land band as a starting frame, count colored sources rather than raw lands, and remember that the London Mulligan rewards decisiveness. When a hand fails either gate, ship it without ceremony. A six is not a punishment; it is often the better hand.