One of the first things that every Magic player learns is that you must play a minimum of 60 cards in Constructed and a minimum of 40 cards in Limited. One of the next things you learn is that it’s best to stick to that minimum. The reasoning for that is simple: adding more cards will decrease the consistency of your deck, in particular the probability of drawing your best cards.
But does this wisdom always apply? Or are there exceptions where playing more than the minimum number of cards is better?
Before I begin, I need a huge disclaimer: I will consider bizarre, niche examples that I had to rack the internet to find. Please do not think this gives you a justifiable reason to play more than the minimum number of cards—you need a specific set of reasons for that.
Reason 1: You Are Playing with Cards that Devour Your Library
A great illustration of this is the following 88-card deck that Noah Boeken and Trevor Blackwell played at Pro Tour Osaka 2002:
[deck]3 Darkwater Catacombs
1 Tainted Isle
1 Faceless Butcher
4 Mesmeric Fiend
4 Shadowmage Infiltrator
4 Aether Burst
2 Buried Alive
4 Chainer's Edict
2 Circular Logic 1 Concentrate
2 Darkwater Egg
1 Diabolic Tutor
4 Insidious Dreams
2 Morgue Theft
By playing more than 60 cards, they could [card]Traumatize[/card] themselves and get enough cards into the graveyard to win in one blow with [card]Psychatog[/card]. I'm not sure why they decided to play exactly 88 cards, as it still would have been possible to attack for 20 on turn 5 with an 80-card deck. But the underlying idea makes sense: playing more than 60 cards makes Traumatize much better.
However, playing so many cards comes with drawbacks. Most importantly, it substantially reduces the probability of drawing [card]Psychatog[/card] in the first place. The upside of milling 40 cards doesn’t help if you didn't actually draw Psychatog.
As it turned out, both Noah and Trevor left the tournament with zero wins, so their deck innovation was not a success story. But there are success stories with decks that go over 60 cards. Hall of Famer William "Huey” Jensen made the Top 8 at Grand Prix Milwaukee 2002 with this 244-card deck:
Battle of Wits
[deck]4 Adakar Wastes
4 Battlefield Forge
4 Caves of Koilos
4 City of Brass
4 Coastal Tower
4 Darkwater Catacombs
4 Salt Marsh
4 Shadowblood Ridge
4 Shivan Reef
4 Skycloud Expanse
4 Sulfurous Springs
4 Underground River
4 Urborg Volcano
4 Flametongue Kavu
4 Battle of Wits
4 Chainer's Edict
4 Charcoal Diamond
4 Chromatic Sphere
4 Diabolic Tutor
4 Dromar's Charm
4 Fact or Fiction
4 Fire and Ice
4 Innocent Blood
4 Insidious Dreams
4 Marble Diamond
4 Memory Lapse
4 Prophetic Bolt
4 Sky Diamond
4 Sleight of Mind
4 Spite and Malice
4 Tainted Pact
4 Urza's Rage
4 Wild Research
4 Yawgmoth's Agenda[/deck]
This deck is basically a lot of cantrips and tutors, all to get [card]Battle of Wits[/card]. You want it to win the game even when you draw it on turn 20 after drawing a bunch of extra cards, so that's how you end up with 244 cards. Another bonus of playing so many different cards is that [card]Wild Research[/card] and [card]Insidious Dreams[/card] get better. All in all, Battle of Wits is a perfectly acceptable reason to play more than 60. A much better one than Traumatize, I would say.
Traumatize and Battle of Wits are not the only examples of cards that need a big library. Here are a few more:
Back when Alliances was legal, Erik Lauer played a 150-card deck with [card]Thought Lash[/card] and [card]Browse[/card]. And back when [card]Arc-Slogger[/card] was still in development and an activation dealt 3 damage, Patrick Chapin tested an 84-card deck in the Future Future League because he wanted to be able to Slog 7 times by turn 7. These are cool ideas, but in my opinion, these examples are still a bit suspect because Thought Lash, Browse, and Arc-Slogger don't immediately win you the game (in contrast to Battle of Wits). And there's still the huge downside of substantially reducing your chances of drawing your key cards. So, I wouldn't recommend playing a huge deck for those cards.
Reason 2: You Want to Deck Your Opponent or Don’t Want to Lose to Decking
The best example of this is the 4000-land deck (with 4 [card]Sword of the Ages[/card] for style points) that was played during the combo winter of 1998/1999. Back then, everyone was playing combo decks based around [card]Tolarian Academy[/card] whose sole win condition was a huge [card]Stroke of Genius[/card]. These decks could generate a ton of mana, but not infinite, and they were unable to cast a Stroke of Genius for 4000. So, eventually, the Academy deck would just get decked itself. Even though you could never beat a creature deck, playing a 4000-land deck in this combo-heavy metagame was a stroke of genius.
A more recent example is the deck that Jeremy Neeman was sideboarding into at Pro Tour Dark Ascension:
Lost in the Woods
1 [card]Lost in the Woods[/card]
Jeremy's plan was to mulligan in search of his [card]Lost in the Woods[/card], all the way down to one card if necessary. Once he found it, he could thwart every attack from his opponent, as he would reveal Forest after Forest, until he eventually would deck his opponents with his 45-card deck. Neat.
Other stories in which people played more than the minimum amount of cards revolved around these cards:
And Circle of Protection: Red.
I've seen a mono-red player boarding up to 42 cards when facing Circle of Protection: Red, and I've boarded up to 60 or even 80 cards when facing mill cards like [card]Dampen Thought[/card] or [card]Jace's Erasure[/card] in draft. I've also seen Constructed decks with a [card]Stasis[/card]/[card]Kismet[/card] lock or an infinite life combo that run 61 or 62 cards, planning to deck the opponent in case their primary win condition fails.
Finally, various people have played more than 40 cards because they drafted all removal and card draw spells, with no actual win conditions. For such a deck, the only reasonable win condition is to deck the opponent. Craig Wescoe, for example, played a 50-card deck once because he lacked creatures, and he did win the draft by decking most of his opponents.
In my opinion, when decking is a viable path to victory, then it is acceptable to play more than the minimum number of cards.
Reason 3: You’re Superstitious and Need Your Deck Size to Be a Prime Number
This is not a good reason.
Reason 4: You Have a Toolbox Deck with Cards You'd Usually Prefer Not to Draw
Gerry Thompson and Patrick Chapin were experimenting with 65- and 66-card [card]Mystical Teachings[/card] decks back in 2007. The idea was that additional one-ofs improved the versatility of Teachings. Moreover, running more than 60 cards reduced the chance of drawing silver bullets in matchups where you don't need them. As an example, here's the deck that Patrick Chapin once played:
[deck]4 Snow-Covered Island
4 Dreadship Reef
4 River of Tears
1 Sulfurous Springs
4 Tolaria West
2 Urza's Factory
1 Vivid Creek
1 Academy Ruins
2 Urborg, Tomb of Yawgmoth
2 Frost Marsh
1 Mouth of Ronom
1 Bottle Gnomes
1 Teferi, Mage of Zhalfir
1 Venser, Shaper Savant
2 Coalition Relic
1 Mind Stone
4 Prismatic Lens
2 Careful Consideration
4 Cryptic Command
1 Grim Harvest
4 Mystical Teachings
1 Pact of Negation
4 Rune Snag
1 Slaughter Pact
1 Tendrils of Corruption
1 Nameless Inversion[/deck]
It is undeniable that 66 cards improves [card]Mystical Teachings[/card] and reduces the probability of drawing a worthless [card]Extirpate[/card], but it also hurts the consistency of the deck.
Was it worth it? Well, to quote Patrick Chapin's article on this deck:
"This deck needs to be cut down to 60 cards. All this rationalizing is just wrong. I am not sure what all cards to cut, but I know that I want to draw my good cards more, as well as get better mana draws. ... There may come a day when it is right to run more [than 60 cards], but that day has not yet come.”
A couple of years later, Patrick Chapin tried again:
The Number of the Beast
[deck]444 Shadowborn Apostle
1 Reaper from the Abyss
1 Sire of Insanity
4 Cavern of Souls
The reason for running 666 cards is that you don't actually want to draw any of the Demons; you just want to draw Swamps and [card]Shadowborn Apostle[/card]s. While this is a cool idea, [card]Griselbrand[/card] and [card]Shadowborn Apostle[/card] did not post impressive finishes in any major Standard tournament. The power was simply not there.
Yet, the day may have come in 2011 when Makihito Mihara made it to the finals of Grand Prix Kobe with this 64-card deck:
3 Exotic Orchard
1 Fire-Lit Thicket
4 Misty Rainforest
1 Murmuring Bosk
4 Seaside Citadel
4 Valakut, the Molten Pinnacle
4 Oracle of Mul Daya
4 Cryptic Command
4 Prismatic Omen
4 Rampant Growth
The reason for playing more than 60 cards is that you want to play at least 7 Mountains for [card]Scapeshift[/card], but also need enough blue sources for [card]Cryptic Command[/card]. The only way to obtain a high enough ratio of blue-producing sources, under the Mountain requirement, is to go over 60 cards. This does reduce the probability of drawing Scapeshift, which is a huge downside, but not having triple-blue for Cryptic Command may be even worse.
Makahito Mihara is not the only player to find success with this approach: James Zornes made it to the Day 2 of Grand Prix Atlanta 2011 with a very similar 66-card Scapeshift deck. You can find his list and reasoning here.
In my opinion, if you want to play Scapeshift and Cryptic Command in a format without [card]Steam Vents[/card] or [card]Stomping Ground[/card], then going slightly over 60 cards is acceptable.
Reason 5: You Want to Optimize the Mana Ratios
In a 40-card deck, you deck can only contain 37.5%, 40.0%, 42.5%, 45.0%, or 47.5% lands (unless you want to play less than 15 or more than 19 lands). In a 41-card deck, you can reach certain numbers in between: 39.0%, 41.5%, 43.9%, and 46.3%. On more than one occasion, I have played a 41-card deck with 18 lands myself, though a rare and specific combination of two circumstances needed to occur for that: (i) I would have a very strong feeling that 42.5% was too few, 45.0% was too much, and 43.9% was just right, and (ii) my deck did not contain any bombs, i.e., the 41st card was just as good as the 1st card. Yet, for those situations, I had always felt that playing 18/41 was an effective way of getting an improved mana ratio.
Moving to Constructed, there have been multiple 61-card decks in the finals of Pro Tours. Billy Moreno, for example, finished second at Pro Tour Los Angeles 2005 with his 61-card Madness Tog deck. But the more famous example is Gabriel Nassif’s 61-card deck that he used to win Pro Tour Kyoto 2009.
3 Broodmate Dragon
3 Wall of Reverence
4 Broken Ambitions
1 Celestial Purge
2 Cruel Ultimatum
4 Cryptic Command
4 Esper Charm
1 Pithing Needle
4 Volcanic Fallout
2 Cascade Bluffs
2 Exotic Orchard
1 Mystic Gate
4 Reflecting Pool
4 Sunken Ruins
2 Vivid Crag
4 Vivid Creek
3 Vivid Marsh
2 Vivid Meadow[/deck]
Gabriel defended his 61st card by stating that he was getting mana screwed with 26 lands and 60 cards and getting flooded with 27 lands and 60 cards in playtesting. To avoid that, he went for 27 lands and 61 cards.
But is the improved land ratio really worth the inconsistency of a larger deck?
I set out to shed some light on that topic by applying the methodology I described in my article “Finding the Optimal Aggro Deck via Computer Simulation.” I used basically the same code as in that article, except that I considered formats with only two cards (e.g., Grizzly Bears and Forests, or Mons's Goblin Raiders and Mountains) and allowed 61-card and 62-card decks in addition to 60-card decks. I also studied the 40-card equivalents.
My algorithm enumerates all possible deck configurations, evaluates for each of them the average goldfish kill on the play under an optimal mulligan strategy, and determines the best one.
Here are the raw results:
Only 1/1s for 1 and lands, 60-card minimum: 60 cards (42 creatures, 18 lands) is optimal
Only 2/2s for 1 and lands, 60-card minimum: 60 cards (40 creatures, 20 lands) is optimal
Only 3/3s for 1 and lands, 60-card minimum: 60 cards (35 creatures, 25 lands) is optimal
Only 2/2s for 2 and lands, 60-card minimum: 60 cards (36 creatures, 24 lands) is optimal
Only 3/3s for 2 and lands, 60-card minimum: 60 cards (30 creatures, 30 lands) is optimal
Only 4/4s for 2 and lands, 60-card minimum: 60 cards (30 creatures, 30 lands) is optimal
Only 1/1s for 1 and lands, 40-card minimum: 40 cards (28 creatures, 12 lands) is optimal
Only 2/2s for 1 and lands, 40-card minimum: 40 cards (27 creatures, 13 lands) is optimal
Only 3/3s for 1 and lands, 40-card minimum: 40 cards (24 creatures, 16 lands) is optimal
Only 2/2s for 2 and lands, 40-card minimum: 40 cards (24 creatures, 16 lands) is optimal
Only 3/3s for 2 and lands, 40-card minimum: 40 cards (20 creatures, 20 lands) is optimal
Only 4/4s for 2 and lands, 40-card minimum: 40 cards (20 creatures, 20 lands) is optimal
So, in all of these formats, the optimal 40-card deck had a faster goldfish kill than the best 41-card or 42-card deck, and similarly for the 60-card variants as well. This was a big surprise to me. I had honestly expected to find at least one example format in which a 41-card deck would be best. If there ever would be a format where that would happen, then it would be here. After all, the 41st card would be just as good as the 1st, and the improved land ratio could matter. But as it turned out, the minimum number of cards was still better.
An intuitive explanation for this is that a smaller deck reduces variance. Suppose that you’re playing a 40-card deck with 24 [card]Grizzly Bear[/card]s and 16 Forest and that you’re drawing your opening hand. If your first card is a land, then this reduces the ratio of Grizzly Bears in your deck from 24/40 to 24/39—a small reduction, but it does make it more likely to draw a Grizzly Bear as your next card, which eventually improves the likelihood of drawing a nice mix between creatures and lands. Now suppose that you’re playing a 41-card deck with 25 Grizzly Bear and 16 Forest. In the same situation, the ratio reduces from 24/41 to 24/40. This reduction is slightly less than with 40 cards, which eventually leads to a slightly higher chance of mana flood. An alternative way to grasp this is by thinking about extremes: If you’re playing a ten-card deck, then it is effectively impossible to get mana screwed or mana flooded, but if you’re playing a million-card deck, then a card drawn will not noticeably affect the ratios in your remaining decks, and there is no smoothing or deck thinning effect.
As a result, running a 61st card to add “half a land” to your deck is almost always wrong. If you really need to add half a land, then cut an expensive spell and add a cheap cantrip.
In this article, I discussed various reasons for running big decks. The ones that I found acceptable were based on cards like [card]Battle of Wits[/card], [card]Dampen Thought[/card], and [card]Scapeshift[/card]. I also investigated via a simulation experiment whether or not it is worth running 41 or 61 cards to improve the land/spell ratio. The results indicate that the optimal 40-card or 60-card deck is still better. All in all, even though it’s fun to think about possible exceptions to the rule, it’s usually best to stay disciplined and register the minimum amount of cards.
I’ll be back in two weeks with my Pro Tour Born of the Gods report. I’m not bitter about the wild unbannings, as brewing has been fun so far, and I expect that the Pro Tour will be interesting to watch. I’m not at liberty to discuss my thoughts on the Modern format yet, but I do have one piece of advice for my fellow Pro Tour competitors: If you’re thinking about playing a 61-card [card]Birthing Pod[/card] to squeeze in an extra utility creature, or a 61-card Affinity deck to effectively obtain 16.7 lands—don’t. Just don’t.