As an Amazon Associate I earn from qualifying purchases and sometimes recommend products from other sellers at no extra cost you. For more details see my disclosure policy and privacy policy.
You’ve got seven cards on the mulligan screen, a clock ticking in the corner of your eye, and a decision that will shape the next fifteen minutes more than almost anything else you do this match. Most players make this call on vibes, half-remembered folk wisdom, and a general sense of dread about giving up card advantage. That’s not a framework, it’s a coin flip with extra steps, and it’s costing you games you should be winning.
🃏 What A Keepable Hand Actually Needs
The “always keep two lands” rule that gets repeated in Arena chat is only true by accident, when it happens to be true at all. A two-land hand with two one-drops and a two-drop in your colours is a snap keep in an aggro deck. That exact same two-land hand in a control deck with three six-mana spells is a disaster waiting to happen.
Land count on its own tells you almost nothing. What actually matters is the intersection of three things: how many lands you have, whether those lands produce the colours your early spells need, and whether your curve gives you something productive to do on turns one through three. A hand with four lands and zero plays before turn four is not safer than a three-lander with a clean curve, it’s just a different kind of bad.
Colour access is the piece players skip most often. Two lands that are both the same colour when your hand needs both colours by turn three is functionally a one-land hand. Before you count lands, check whether they actually cast your spells.
🎯 How Arena’s London Mulligan Actually Works
MTG Arena uses the London mulligan, the same rule that’s been standard across paper and digital Magic since 2019. Every time you mulligan, you shuffle your hand back in and draw a fresh seven, not a smaller hand. Once you decide to keep, you then put a number of cards equal to how many mulligans you took on the bottom of your library, choosing which ones.
Arena’s interface lets you drag unwanted cards to the bottom and arrange whatever remains in whatever order you like on top. Plenty of players call this “scrying to the bottom,” and functionally it feels similar, but it isn’t a true scry. You’re not looking at your library, you’re selecting which of your seven drawn cards to keep and which to bury, then sequencing the survivors. That distinction matters less for decision-making than the practical result: a one-mulligan hand still gives you seven cards to evaluate before trimming to six, which is a meaningfully better starting point than the pre-2019 rule that just dealt you six cards outright.
One quirk worth knowing: Brawl and Standard Brawl on Arena grant a free first mulligan, so you don’t lose a card the first time you take one in that format. Constructed and Limited don’t get that exception, so every mulligan after your opening seven costs you a card on the bottom.
⚖️ Bo1 Versus Bo3: Why Risk Tolerance Changes
The single biggest Arena-specific variable in your mulligan decision is whether you’re playing best-of-one or best-of-three, and our full breakdown of Bo1 vs Bo3 on MTG Arena covers the wider strategic gap between the two. For mulligan purposes specifically, the key difference is that Bo1 gives you no sideboard and no second game to correct course. If you keep a clunky hand and get run over, that’s the match, full stop.
That reality should push your Bo1 mulligan standard slightly higher, not lower. A borderline hand that you might keep in game one of a Bo3, trusting that you can adjust after a loss, deserves less benefit of the doubt in Bo1 because there’s no adjustment coming. In Bo3, once you’ve seen a game against your opponent, you also have information about their deck that can make an otherwise risky keep more justifiable in games two and three.
Arena’s Bo1 queues also use a hand-smoothing algorithm on your opening seven, which slightly reduces the frequency of true non-games without changing how the London mulligan itself functions once you’re deciding whether to keep or send back.
🧠 The Framework, Deck By Deck
A repeatable framework only works if it flexes for what you’re actually playing, so here’s how the same four checks shift by archetype.
Aggro decks care most about turns one and two. If you’re leaning on the fundamentals covered in aggro on MTG Arena, a two-land hand with a turn-one and turn-two play in your colours is often better than a three-land hand that does nothing until turn three, because tempo lost early is rarely recovered.
Midrange hands need to survive to their curve’s sweet spot around turns three to five without folding to aggression first. The midrange strategy guide framework of flexible answers and proactive threats means your mulligan check should weigh whether the hand can both apply pressure and stabilise, not just one or the other.
Control decks flip the priority entirely: early plays matter less than having interaction to survive to your card advantage engines. Per our guide to control on MTG Arena, a control hand with four lands and a removal spell but no proactive play before turn four can be a perfectly fine keep, where that same hand would be a mulligan for an aggro deck.
✅ Your Step-By-Step Mulligan Checklist
Run this mentally at the mulligan screen, in order, and stop as soon as a step gives you a clear answer.
First, count your lands against your deck’s actual land count, not a generic number, remembering that three or four is the sweet spot for most 17-land decks.
Second, check colour access specifically for the spells you’d want to cast in the first three turns, not just the colours present somewhere in your hand.
Third, map out what you can actually do on turns one through three, since a hand with lands but no action is still a mulligan candidate.
Fourth, weigh your archetype’s priorities from the section above, because a “playable” hand for a control deck can be unplayable for aggro and vice versa.
Finally, factor in your format: tighten your standard in Bo1, loosen it slightly in Bo3 game one if you’re behind in the match already and need to gamble.
🚨 Snap Mulligans Versus Hands That Only Look Scary
Some hands mulligan themselves. Zero or one land, six or seven lands, or a hand entirely in one colour when your deck needs two colours by turn three are all snap mulligans regardless of archetype. So is a hand with lands but literally nothing castable before turn four in an aggro shell.
Plenty of hands that feel risky are actually fine. A three-land hand with a slightly awkward curve but on-colour mana and a card advantage spell is usually a keep, because you have time to draw into the gaps. An aggro hand with only two lands but a turn-one and turn-two play that curves out is often better than a “safer” four-land hand that stumbles into turn three with nothing to do.
⚠️ The Two Mistakes That Cost The Most Games
Mulligan-averse players keep hands out of habit, treating any hand with two or more lands as automatically fine regardless of colour fit or curve. This is the single most common leak in ranked ladder play, and it quietly loses games that never should have been close. If you’re still building out your deck and wondering whether your mana base is the actual culprit, our guide on upgrading an MTG Arena deck without wasting wildcards covers consistency fixes that reduce how often you face these borderline hands in the first place.
The opposite mistake shows up mostly in Bo1: players over-mulligan because there’s no sideboard safety net, so any hint of risk triggers a send-back. Every mulligan costs a card, and going to five or six repeatedly because a hand was merely unglamorous, rather than genuinely unkeepable, is how a fine starting position turns into a real disadvantage.
Building The Instinct
None of this replaces reps. The framework above gives you a checklist for the moments when your gut and your actual hand disagree, which happens more often than most players admit. Run the same four checks often enough and the process stops feeling like a checklist and starts feeling like instinct, which is exactly the point.