There and Back Again: Living End in Overextended

I’m Chris Jones (Voidmaw on MTGO), a Living End pilot in the last two evening Overextended tournaments.  I picked up the deck in part because I enjoy playing nifty combo decks, and in part because it’s one of the cheapest competitive decks that can be built in Overextended. I adopted largebrandon’s 4th place deck from Week 1 in week two for the low cost of 17 tickets.  A few tweaks and another 15 tickets developed the incarnation I played to a winning record in week three.

This article is part primer, part statistical analysis.  For those who might still be unfamiliar with the deck, I go over the basics from my own experience.  I then detail several relevant considerations that come up during play.  After that, I discuss some of the math I applied in modifying my deck, with figures and a table you can apply to your own deck design.

My hope is that you’ll find this a useful read on how Living End works, whether you want to adopt it yourself or see how you would attack it, as well as having a useful resource for future deck design.  Enjoy, and remember:  sometimes the best creature in a game is the one that’s just waiting to come back from the dead.

Living End, as suggested by Chris Jones

Lands: (20)
Blackcleave Cliffs
Copperline Gorge
Verdant Catacombs
Forgotten Cave
Dryad Arbor

Creatures: (29)
Fulminator Mage
Street Wraith
Monstrous Carabid
Deadshot Minotaur
Igneous Pouncer
Jungle Weaver
Ingot Chewer

Spells: (11)
Living End
Violent Outburst
Demonic Dread
Sideboard: (15)
Faerie Macabre
Krosan Grip
Kitchen Finks
Engineered Plague
Ingot Chewer


The goal of a Living End deck is to develop a board state where you control more creatures than your opponent, and to do so within the first few turns of the game.  From here, it should be easy to win by attacking, as even five mana 3/4 creatures can win the game when you control four of them to your opponent’s none.  The means of doing this, of course, is casting the namesake Living End for free via the cascade mechanic, simultaneously wiping the opponent’s side of creatures and giving you an army.

The deck is composed of three main pieces:  combo elements, creatures that can be placed in your graveyard easily and for value, and the mana base.  The combo elements consist of Living End and the cascade spells able to trigger it. Historically, these have been Violent Outburst and Demonic Dread due to their low casting cost and effective coloring.  The creature base is dominated by Alara Reborn creatures that can cycle cheaply.

The main requirements for a creature to be used is that is must cost at least 3 mana (so as not to interfere with the cascade), and that it can be dumped into your graveyard within the first three turns.  The mana base is designed to quickly supply black, red, and green mana, meaning a cheap and effective mana base can be constructed from Scars of Mirrodin duals and basic lands.

The deck aims to be very non-interactive, as it cycles creatures in the early turns before casting a cascade spell to cast Living End for free, returning all the cycled creatures and eliminating those the opponent has already cast.  In some matchups, this is sufficient to win the game, as the opponent doesn’t have a way to stop your army from killing him in one or two turns.  In others, the first attempt at casting Living End doesn’t work out, requiring a second cascade spell or a natural Living End to once again set the board in your favor.

Of note is that since you have no way to recycle cards from your graveyard into your deck, you can only Living End four times at most in a long game.  Also, any drawn Living Ends reduce your opportunities to cascade. If you find yourself holding a Violent Outburst and three Living Ends with one in the graveyard (as I did during one game I lost), you’re in a tricky position.

Basic Lines of Play

A very good hand has 2 or 3 mana sources (covering different colors), 2 or 3 cyclers, and a Violent Outburst.  With this hand, you spend the first two turns cycling (generally on the opponent’s end step), then cast Violent Outburst on the opponent’s end step after you hit three mana.  Thus, you don’t expose your newly arrived creatures to sorcery speed removal before getting at least one attack, and can swing for at least 10 damage. Occasionally, it’ll be lethal right off the bat.  At this point, your opponent likely has no more than a single turn to deal with your creatures before you kill him.

A decent hand has 2 to 4 mana sources and a couple of cyclers, but no way to immediately cast a cascade spell on turn 3.  In this case, you will need your cyclers to dig for a spell, while allowing your opponent to develop whatever board they like.  Against a midrange or control opponent, this can still work out, but a Red Deck Wins opponent may not give you time to cast Demonic Dread on turn 5.  Thus, the decision of whether to mulligan such an opener rides on the opponent and whether you can afford to delay going off for a few turns. (Of course, you might cycle into the missing spell in the first two turns and remain on schedule.)

A weak hand contains a Living End, 1 or 2 mana sources, only 1 good cycler (more on this below), and several cascade spells.  Here, you’re both not assured of having the mana you need to go off, and you might not be able to get a sufficient number of creatures into your graveyard to set you up for a good Living End.  At this point, you should strongly consider mulliganing; while it is possible to draw and cycle your way into a good situation, the deck mulligans well enough that a fresh six can still set you up for a good hand.


Cycling and Beating Hate

Not all cycling creatures are created equal.  There are four types of cyclers in my list, which I refer to as Ones, Twos, Landers, and Others.  Ones consist of creatures whose cycling cost is one colored mana.  Twos are the creatures with a cycling cost of two colorless mana.  Landers are those that cycle for one of two basic land types (rather than drawing a card).  Others are cards that don’t cycle, but serve some other purpose while still getting into the graveyard quickly (Fulminator Mage, Ingot Chewer).  Street Wraith, being (almost) always free to cycle, is separate from these categories.

Because the deck has more mana sources than it does cascade spells, the Twos are more valuable than the Landers, as you’re more likely to be digging for a spell than a land.  On the first turn, you can cycle a One, while on the second turn you can either cycle two Ones or a Two.

In general, the preference is to draw and cycle as many Ones as possible, as this both digs you deeper, and gives you more creatures for the Living End.  However, if after your draw step on the second turn you are holding a One and a pair of Twos, you may want to set yourself up to cycle the Two now, and the One and the other Two on the next turn, for a total of three cards and three creatures when Cascading on turn four.  If you’re good to go on both lands and a cascade spell, then you prefer to cycle the Landers over the Twos, as they generally have more power.

Unlike Jund, where Bloodbraid Elf into Blightning leaves a player holding one counterspell to decide which is more threatening, a Living End deck is actually vulnerable to countermagic despite the cascade mechanic.  After all, the Living End is still cast, whether cascaded into or suspended, and the other spell (Violent Outburst or Demonic Dread) is almost always irrelevant.  Therefore, it becomes important to play around untapped blue mana.

There are two ideal opportunities to sneak under countermagic, both requiring Violent Outburst: at the end of your third turn if your opponent taps out for an instant effect of his own (Thirst for Knowledge, activating Stoneforge Mystic), and at the end of the opponent’s third turn while holding a second cascade spell to play during your subsequent turn.  If neither of these opportunities presents itself, it becomes more difficult to execute your plan, and you will eventually need multiple cascade spells to overwhelm the opponent’s countermagic (since resolving Living End is all that matters to you).  The backup plan of hard casting these creatures and beating is challenging in the face of a control opponent who has counterspells and removal.

Even your fastest hands aren’t fast enough to beat a great draw from Red Deck Wins.  They can easily burn you out before you get the second swing with your army you raised on turn 3.  Fortunately, if they stumble, you can be fast enough.  However, one of the points of my recent sideboarding work has been to add a useful card for fast aggro opponents; I feel Kitchen Finks, as it often has been before, is excellent for this role.  They are great both hard cast, and coming back from the dead in gaining enough life to buy time to win.

Against decks that have significant graveyard hate, you have two lines of play:  go for the combo and hope they don’t have it, or delay cycling by several turns until you can first use an Ingot Chewer or Krosan Grip to remove the relevant piece.  Neither approach is ideal, but such is the drawback of playing a deck so dependent on the graveyard.  However, in a diverse format such as early Overextended, with graveyard decks making up only a small part of the metagame so far, opponents can only dedicate so much sideboard room to beating you, and sometimes, they just don’t have it in time.


Answering questions such as “What is the probability I will have exactly two lands in my opening hand” depend on calculations using the hypergeometric distribution.  In brief, such a calculation takes the sample size and number of sample successes (for the question above, these are 7 and 2 respectively), and the population size and number of successes in the population (here, 60 and the number of lands in the deck, respectively), and computes the probability the event will occur.  The calculation is slightly more complicated when asking for the probability of having at least two lands rather than exactly two lands, but is still relatively easy to compute.

In working with the deck from the starting point of largebrandon’s list from the first Overextended online tournament, I wanted to revisit the mana base and decide how many lands to include.  Thus, I decided to compute the probabilities of having at least X lands in my starting hand, depending on the number of lands in my deck.  The results are shown in Figure 1.  I then performed a similar calculation for having at least X lands in the first ten cards (opening hand plus three draws), to see what I could expect by the time the key turn for casting a cascade spell came up. Those results are shown in Figure 2.

Figure 1:  Probabilities of having at least X lands in an opening hand, depending on the number of lands in the deck.

Figure 2:  Probabilities of having at least X lands in the first ten cards drawn (opening hand and three draws), depending on the number of lands in the deck.

So, to interpret this chart, consider the question of having at least two lands in an opening hand.  Look at the green line with triangles in Figure 1, then consider how many lands you have in your deck.  You can now see the probability of having at least two lands in your opening hand for the number of lands in your deck, as well as if you were to add or subtract a land.

Largebrandon’s original list had 19 lands, giving it a 72% chance of having at least two lands in the opening hand, and a 68% chance of having at least three lands in the first ten cards drawn.  Adding a land (as I did) raises those probabilities to 75% for having two lands in the opener, and 72% for having three lands around the time I’m ready to go off.

All in all, pretty marginal changes, but since the land I added can cycle (Forgotten Cave), I’m also adding another chance at drawing a card.  In my current list, I’ve also added three fetchlands in the place of three other lands, which can slightly alter the probabilities if you plan on cracking in those early draws.  However, the trends above remain the same.

Note that these charts are not just for the Living End deck, but apply to any 60 card deck with a number of lands between 18 and 28 (that should cover most non-Affinity, non-Lands Players).  You can use this chart as a handy reference in deciding how many lands you want to include in your own deck; just think about how often you need at least a certain number of lands in your opener, and decide on the number of lands to include that gives you that probability.  For reference, Figure 3 below shows the probability of drawing exactly a certain number of lands in your opener depending on the number of lands in the deck; the probabilities above simple come from the summation of the relevant curves below.


Figure 3:  Probabilities of having exactly X lands in your opening hand, depending on the number of lands in your deck.

What about some other mathematical questions we can address about Living End?  Well, we’d like to have a Violent Outburst in the first ten cards we draw, since it’s our best cascade spell.  That probability is 40%, given that we have four in the deck and are trying to draw exactly one in the first ten cards.  If we instead want to draw either one or two in the first ten cards, the probability rises to 51%.  So, our chances of having a Violent Outburst by the time we’re ready to go off are better than half.

By comparison, our chances of having a Demonic Dread are 36% for exactly one in our first ten draws, and 42% for having either one or two, since in my list I play only three.  The general probability tables for the chances of drawing a card in your opening seven or first ten cards are shown in Table 1; note that these are the exact probabilities, rather than the at least probabilities, with the “1 or 2” events representing the chances of drawing either one or two, but ignoring the chances of drawing three or more.


Number in Deck
Event 1 2 3 4
Draw 1 in 7 12% 21% 28% 34%
Draw 1 in 10 17% 28% 36% 40%
Draw 1 or 2 in 7 12% 22% 31% 40%
Draw 1 or 2 in 10 17% 31% 42% 51%

Table 1:  Probabilities of drawing a particular card in the opening hand or first ten cards, depending on the number of that card in the deck.

To determine the probability of drawing either cascade spell, we cannot simply add the individual probabilities.  Instead, we rerun the calculation from the perspective of having seven cascade spells in the deck.  Our probabilities rise to 41% for having exactly one cascade spell in our first ten cards, and 66% to have either one or two by the time we’re ready to combo.  Thus, we can expect that in two-thirds of our games, we’ll have the cascade spell on turn 3. (Hopefully there’s someone to target with Demonic Dread!)


Although the pool of options available to Living End is rather small (due to the requirements of costing at least three and being dumpable in the graveyard quickly), there is still some play in how to assemble the deck.  Adding more Ones can improve the ability to cycle, at the risk of being unable to hardcast a black or white creature.  Bloodbraid Elf is both a cascade spell and an efficient threat on its own, but requires an additional mana and means everything else must move up to being a four instead of a three.  Replacing the Scars duals or basic lands with other duals may improve the mana consistency at the price of making you vulnerable to non-basic land hate or an aggressive opponent.

Living End is a fun, inexpensive combo deck that has a good chance to develop a strong hand early in the game.  As the Overextended metagame continues to evolve, the role of Living End (either as a major combo deck or a fringe competitive option) will likely wax and wane depending on previous tournaments.  For those who are looking to get into Overextended and don’t have the cards or the background to play a more complicated and expensive deck, Living End represents an excellent starting point for having fun in the tournament while executing a combo that doesn’t require significant planning.

If you have any feedback or questions, go ahead and post them below. Until next time, have fun with Living End!

Chris Jones

Voidmaw on Magic Online

About Chris Jones

Chris started playing Magic during Tempest Block and loves Voracious Cobra a little too much. These days, he plays casually while working on a PhD in aerospace engineering.