This entry is part 14 of 14 in the series Decksplanations

Grandpa (Eric)

By Eric, AKA Grandpa Growth


Welcome back! This is a direct continuation of last week’s article. If you missed it, click the links below to get up to speed before continuing.


Last week we were walking through a hypothetical deck design: a Grixis Control deck built to resolve Cruel Ultimatum with protection. The process so far is to picture how we want to end the game and then plan out a script of what needs to happen on each turn in order to get us there. Once we have the script in place, we’ll know what different groups of cards will be necessary for our deck to function. In this example, we need some lands, counterspells, removal, disruption, mana acceleration, a backup win condition, mana acceleration, and then we fill up the rest of the slots with tutors and card draw to improve consistency. The major design questions left to answer are:


  1. How many copies of each kind of card do we need for the deck to function smoothly?

  2. How many cards do we need at each mana cost to lower the likelihood of clunky hands that don’t follow the script.

  3. How to decide which cards/groups of cards to include at each mana cost?


Marginal Redundant Cards and Consistency

I already talked last week about some substitutes for Thoughtseize, so rather than circle over the same ground let’s instead look at alternatives for mana acceleration and see if we can meet our mana needs. We aren’t playing Green, so we need to lean on artifacts to supply our mana acceleration.


Here’s a list of some good two-cost ramp spells: three signets, three diamonds, Coldsteel Heart, two Talismans, Sphere of the Suns.


The signets are obviously the best, but those cards are all pretty interchangeable for our purposes and they’ll meet our needs just fine. However, with just ten cards in the group, we don’t have enough to reliably draw one in our opening hand. I could stretch my definition of interchangeable a bit to allow Star Compass and Fellwar Stone, but we lose out on card quality because these might not produce any useable colors unless certain conditions are met. Even still, we only have twelve cards there. We came up short here in this example; in this hypothetical deck would need to mulligan more often to produce a hand that stays on script.


Why did I make the assumption that ten ramp spells isn’t enough? This digs into the underlying mathematics of combinations. I won’t go into detail here, but the formula we use to determine the odds of something like, say, drawing one out of ten ramp spells from your ninety-nine card deck in your first seven cards is calculated using a hypergeometric distribution. I don’t normally perform this type of analysis on every deck I build, but over time I’ve come to rely on the technique to help me gauge the odds of my deck producing a hand that has certain key cards that are important to the deck’s script. This is a very easy heuristic to use when designing decks and it gives me useful information about how to evaluate my mulligan decisions when doing sample hands or Goldfish testing. Take a look at this chart:


Number of Cards in Set Chance to see it in your opening hand
1 0.07
2 0.131515
3 0.185195
4 0.231653
5 0.271468
6 0.305187
7 0.333325
8 0.356366
9 0.374765
10 0.38895


Several assumptions have been made to get to these numbers. For reasons I won’t explain here, these numbers aren’t perfectly accurate, but they are enough for you to get dangerously analytical about deck design. If you are really interested in learning more about applying this type of analysis take a look at this article.


This chart assumes you can’t mulligan and you always start with seven cards. You can find the chances of ‘missing’ by just subtracting the probability numbers from one. This is important if your play group allows for one free mulligan. With ten cards in the set, you have roughly a 60% chance of whiffing in your opener. If you’re allowed to throw that hand back and draw a new seven, then we can find the combined chances of whiffing twice in a row by just squaring that 60% – (0.6)*(0.6)=0.36. So you see,  if you have ten ramps spells in your deck and you draw a seven card opening hand, there is about a 36% chance that you will not have a ramp spell in your first seven OR after your free mulligan. A little more than one out of every three games, you won’t be accelerating your mana according to the script. Is that good enough to call your deck consistent? You be the judge. To me it isn’t good enough, so I’m looking to put more mana accelerators into my decks. Taking the example from above, we can include Sol Ring, Mana Crypt, Mana Vault, Grim Monolith, etc. They might not cast Cruel Ultimatum, but they will get us off the ground to cast early card draw spells and decoy threats.


This data has allowed me to piece together the basic framework for decks that I haven’t even built yet. I have a general skeleton list that I can just fill in with specific cards and I know ahead of time that the distribution of key cards will work out alright. It’s all there in the math. You can engineer more harmonious draws through the use of the hypergeometric distribution. Knowledge is power, my friends. I want about forty lands, twelve or more accelerators, probably ten or so tutors, fifteen cards split between counters/disruption/removal, three-to-five win conditions (including my Commander), and ten-to-twelve ways to get card advantage. Whatever space is left over is where I experiment with new cards.


Because the best cards in each of these categories usually cluster around certain converted mana costs, my curve is pretty much locked-in already. According to the script, I know I need a counter on turn three, so my counterspells must cost three or less. You can easily see how this process essentially builds a workable mana curve for you. I simply adjust the number slightly, depending on what a given deck needs to see more often, but these small tweaks come much later in the testing process.


And so, we’ve arrived back to where I first started: iteration through testing. We still need to grind games to get the deck to work perfectly, but through the use of some careful planning and analytics, we’ve shaved literally thousands of games off of the testing process. Unless you have infinite time to grind games on MTGO, using this method can save you a world of hurt when assembling new decks.


Sadly, this might be a disturbing realization for some Commander players. This format is supposed to be about creativity and telling good stories. The thought of standardizing your decklists may make you feel a bit icky. I’m not telling you to make your decks boring and always build from a stock list, but I am saying that if you want to take a trip to Magical Christmas Land, do yourself the favor of building a deck that can actually get it done. Understanding the mathematics of combinations and how they influence necessary card densities is key to developing consistent decks.


Odd and Even Curves

There’s a fascinating consequence of designing decks using this process (script first, then match card density to script needs): because the same basic card types are used in all decks, you see patterns emerge after a just few iterations.


Also, because of the way R&D designs cards, certain types of effects continually pop up at the same mana costs. When designing cards for an entire format, you can potentially imbalance the metagame if you change the cost of even one key card. Imagine how much worse Modern Burn decks would be if Goblin Guide cost two…


R&D tinkers with the specifics of each card and each type of card and then adjusts the dials on the design machine to change things up from set to set. This results in limited formats that always have familiar elements, but the specifics change to keep things fresh. Well, in Commander we can only play one copy of each card, so we have to collect different cards from across several sets that all work in fundamentally the same way if we want to include any interchangeable elements in our deck. Because we want our hands to fit the script we created in deck design, all of our effects in a certain group should have the same converted mana cost. This naturally leads to a choice in deck building: what type of card do we want to be playing at each spot on the curve? This has massive implications for card selection when building Commander decks. An example:


Should you play Birds of Paradise or Rampant Growth? They both accelerate your deck’s mana by one, the equivalent of a single turn’s worth of land drops, so the effect they have on your mana development is the same. There are metagame considerations– like whether your Bird might get targeted by removal or your lands could be caught in an Armageddon–but before we examine other decks, we should start by taking an informed look at our own deck first.


We play mana ramp because we want to cast something more expensive than normal on a given turn. Are you paying attention? This is the important part:

Which ramp card you play depends on the cost of the thing you are ramping into!


If you want to ramp into a three-drop on turn two, Rampant Growth doesn’t help you do that. You will untap on your third turn, make your land drop, and then you will have four total mana. If you instead played Birds on turn one, you will untap and lay your second land for a total of three mana on your second turn, which lets you cast your three-drop ahead of schedule. This full turn of effective acceleration makes Birds a better strategic choice if you want to ramp into a three.


Let’s say you wanted to ramp into a four instead. They would both get you there on turn three, but then the Bird is exposed to removal for an extra turn. At that point it might be safer not to leave a creature to target. This naturally creates a very elegant structure for your decks. You can best streamline your draws by playing either ramp spells with even costs or odd costs.


One’s jump directly into three’s, two’s jump directly into four’s, and odd-cost ramp spells work well with other odd-cost spells. The same is true of even-cost spells. However, if you draw a hand that has one of each, you minimize the effectiveness of both, so you want to avoid that. That would be the clunk I’m trying to avoid.


Using this knowledge, I’ve assembled two basic skeletons for my lists. One that runs on either side of this odd/even divide. Odd decks have mana curves where the plays are concentrated at one, three, and five mana; Ezuri, Renegade Leader is the perfect example here. You cast a dork on one, Ezuri on turn two, deploy as many elves as you can on turn three, and then use the Overrun ability for five mana to smash. This script has key plays on one, three, and five mana. Two and four are not designed to be emphasized, because we expect to move right past those spots on the curve.


Let’s discuss two more sample scripts.


Sample 1: Strength Through Unity


  1. Untapped Green source, cast Llanowar Elves.

  2. Untapped Blue source, cast Edric, Spymaster of Trest

  3. Play lands, attack, hold up countermagic.


From turn three, we’re attacking for three damage a turn, drawing three cards a turn, and with a high density of counterspells in this Edric deck, we’re probably winning very easily. This script illustrates why one-drop mana dorks are the smartest choice for Edric decks. Accelerants that cost two delay the script. Furthermore, any card that costs more than two and doesn’t counter a spell just doesn’t fit into the deck’s strategy. Those misfits will screw up your mana curve and create clunky hands that don’t execute the script well. The vast majority of the deck should be creatures and counterspells (with fewer than forty lands), because of our low curve.


According to our numbers from the table above, we want more than ten one-drop mana sources to give us the best chances of having one in our opener. Take note of what happens on turn three in the script, though: We draw three cards and we want to counter a spell from the opponent. To sustain this soft lock, we need to draw a new counterspell every turn. With three draws to find one, we need more than thirty counterspells in our deck to average drawing one per turn until the game ends.


So we have thirty-something counterspells, fifteen or more one-drop creatures (preferably that make mana), and around forty lands. That doesn’t leave much room to explore the finer points of the archetype. This decklist might not seem realistic, but the mana curve is par excellence. This deck can execute its script two-thirds of the time and because that script is so tough to beat, you can expect to win a high percentage of those games when your plan comes together. The fact that you can tutor for Forbid saves us from the horror of having to play dozens of extra low-quality counters. This is prime example of where I would employ extra tutors and card draw to cut down on the redundancy in the deck. High levels of velocity and access will keep card quality high and ultimately win more games.


Sample 2: Curve Overlap

Being disciplined with your card selection is important, but you don’t need to be a strict puritan when designing decks this way. Looking at the extremes can help guide you in the right direction, but we’ll inevitably end up with a deck that has some cards that don’t perfectly fit into this rigid script. A mix of odd/even ramp spells and some amount of lands that enter the battlefield tapped is pretty normal in most decks and that isn’t a bad thing.


Because of the raw power of certain lands like the Theros-block Temples, for example, we want to play them in our deck even if they could potentially screw up our early mana development. Ramp spells that cost slightly more, but offer the opportunity for card advantage (like Kodama’s Reach), are very powerful. These cards can be worth playing even if they aren’t using an odd 1->3 architecture for our deck. This creates a certain amount of curve overlap where we could play either a three-drop or four-drop at a given point in the script. Take a look at this second script:



  1. ETB tapped land.

  2. Untapped land, cast Rampant Growth

  3. ETB tapped land, cast Kodama’s Reach


This script allows you to take full advantage of both odd and even ramp spells, as well as lands that come into play tapped. This probably resembles your favorite EDH deck much more closely than the Edric example. Most “good stuff” decks in Commander want to play some amount of acceleration and cast a large threat that costs six or more mana. This is a typical goal for an even-cost ramp deck. We build our mana curve with sufficient two-drop ramp to let us regularly accelerate, but it’s okay if we draw more than one of these cards, because we’ll just cast both and play our big spells even faster. The threats in our deck should be concentrated at four mana and six mana to streamline our draws and ensure that we can actually develop our board on the turns where our script calls for us to do so. We can sequence our lands so that we play a tapped land on a turn where we can still take advantage of some of our mana by playing an odd-cost spell.


In these decks, I always want to play Kodama’s Reach, Cultivate, and the new Nissa’s Pilgrimage, because staying ahead on cards is so important to winning in Commander. However, if my draw includes no two-drop ramp spells, or it has multiple three-drop ramp spells, I have a pretty clunky hand that will not develop well.


To rectify this, (when designing an even-cost deck) I don’t group these odd-cost ramp spells with my other ramp for scripting purposes. I’d label them as generic card advantage spells and group them with Fact or Fiction or Compulsive Research. Once you get through the early game smoothly and make it up to four or five mana, you have more flexibility about what you play. You can cast cards that have odd mana costs and use that turn to play one of your all-important tapped lands.


To plan for this you need to keep the number of the spells in this group much lower. As opposed to the fifteen ramp spells, I might only include a total of 4-8 things that have off-script mana costs in my deck. With a sufficient density of tap lands to use in combination, you can minimize the disruptive impact of taking a turn to cast a potentially clunky spell off-script.


I want to leave you with a caveat: this technique is only valid for very generic deck lists. Decks with strong themes need to play the proper cards to support those themes. That’s potentially more important than having an air-tight curve.


Look at Jhoira of the Ghitu, for example. This deck is going to be full of gigantic spells that cost 10+ mana. It’ll have maybe a dozen ways to destroy all lands. The ideal script for this deck is so unique, and the card requirements so specific, that you can’t just pick up a pre-built skeleton list from a Keranos, God of Storms deck and change a few cards. When you don’t even depend on mana to cast your spells, whether your acceleration has an odd-cost or an even-cost is of little consequence.


Special thanks for this article go out to John Prywes of and our very own Uncle Landdrops. Go forth and experiment with the ideas I presented here and let me know what you think in the comments.



“Decksplanations” is all about sharing my deckbuilding philosophy. If you want to change how you behave, you must change how you think. The same is true of your decks. My goal is you give you the foundation to analyze and improve your decks. In each article I’ll share one idea that shapes how I approach creating a new deck.


Series Navigation<< Strategy: “Decksplanations” – Durability (how to ramp smarter AND harder)