In the last week hanging out with Zvi Mowshowitz, he said something that struck me: approximately, â€˜I don’t build decks the way most people build them. First I look at the mana and what I can do with it, and then build everything else on top of that.’ This didn’t come as too much of a surprise to me since land selection is the first step of every deck design in my process, but I had been overlooking the fact that most people do not do this. He said something else which has been percolating through my brain: â€˜Sometimes I like to think of what the equivalent mana looks like in a Constructed deck,’ during a conversation about Limited manabases. I had never really taken the time to multiply by one and a half to find equivalencies, but once I did a lot of things fell into place. This article will discuss what makes for good mana in both Constructed and Limited, common failings in manabases, and some deck building exercises.
There are some very important numbers in Constructed Magic: 14, 18, 21, 24, 27, and 30. They are roughly the numbers of a kind of card or effect you have to have if you want to reliably see it in the quantity of 1, 2, 3, 4, 5, and 6. Though it can be applied to anything, but the only things we usually want to see in such large and consistent numbers are lands. Let’s say you want 4 lands by turn 4, which means that if you are on the draw you will have seen 10 cards (without further deck manipulation). That means you need 4 of 10 cards to be land, or roughly 40% of your deck. For a 60 card deck that means 24 lands in the deck, and voila, you’re done. This is a relatively mathematically loose model, but pretty accurate. For the lower counts of land, people have to run quite a few more than the “expected” number because of much greater variance. (The number would be 9 to be favorable to draw one land in your opener, and 15 for two. However, we want more than just favorable; we need to be strongly favored.) Once you know to look for these numbers, you will begin to see them everywhere in successful decks.
These Magic numbers are helpful both in telling us how many lands we need to play to cast spells at the top of our curve, but also at telling us how many sources of a particular color we should be looking to acquire. For example, if we want to play Gideon Jura in our Constructed deck, we should be looking at a deck with 27 sources of mana total, of which at least 18 should be White. If we also want to play Jace, the Mind Sculptor, the 27 will have us covered, but we also need 18 of them to be Blue. Quickly adding up our total, we see that we need 36 colored mana symbols out of 27 lands, so 9 of them (36-27) will have to perform double duty for us. In the current format, there are lots of dual lands to choose from, so we can pretty easily come up with 4 Celestial Colonnades, 4 Glacial Fortress, and 1 Terramorphic Expanse, along with 9 each of sources that produce only White or only Blue.
There are many examples of decks which play fewer colored sources but still work, and there are a few reasons why this happens. The first is when decks play card draw spells, because they functionally look at more of their deck. Let’s imagine a world in which we always cast Divination on turn 3, so instead of looking at 10 cards at the beginning of turn 4, we have looked at 12. Where 4 out of 10 are 40%, 4 out of 12 are only 33%, so we could have as few as 20 lands in the deck and behave like a deck with 24 if we lived in this fantasy universe. I am not advocating cutting lands or colored sources when you are drawing more cards, but just pointing out the basis for why some decks look like they are cheating and actually work out alright. (One common pitfall is to count card draw effects that are above the curve of your deck. For example, if we were wanting Divination to help us get to our third source, rather than our fourth, or Tidings to find the fifth source.) The second case is where a spell is not expected to be cast “on curve.” For example, the Red deck that won the Grand Prix in Kuala Lumpur had only 8 Black sources to cast Blightning or his sideboarded Doom Blades (versus Kor Firewalker). Since Blightning is more powerful when it is cast later on, since opponents generally have more expensive/valuable cards left in hand and have played more of their lands, this spell does not have to be cast on turn 3. Even if the deck isn’t likely to see its first Black source until turn 5, it probably doesn’t have any real issues. By turn 5, a deck has seen 11 cards, and has a reasonable chance of seeing one of these eight. Likewise, the decks with the Kor Firewalkers were not creating a lot of pressure, so the player had a fair amount of time to wait to find the land he needed. Even though eight sources are probably on the loose side, the deck only had 4 Blightnings, so even if they were stranded in hand he was only down one card. (Unpleasant, but livable.)
This brings us to another discussion point I wish someone had had with me last year. When I re-joined the Magic world during Time Spiral, I thought I would first play aggressive decks, then combo decks, then control decks, in the order that I perceived their complexity of play and required knowledge to make good decisions. All the aggressive decks worked best on shoestring mana, just barely playing recommended number, or sometimes shaving one off. The White Weenie deck that I made the top 16 with in GP: San Jose played only 23 lands despite Calciderm being at the top of the curve. The reason this tends to work is that aggressive decks have very few of their spells at the top of their mana curve, and it’s not so bad to have to play your low cost spells until you rip a land to play that last spell hanging out in your hand for the final push. If you’ve read my other articles, you know I’ve applied this same reasoning to my Zoo decks this season with good effect. Unfortunately for me, I brought this logic to the Control decks I played last year, and did very poorly. It wasn’t until near the end of the year that I realized I was mostly losing because I was missing land drop(s), and that it was because I was trying to skimp on land to play more spells. Because it is usually rushing to contain an aggressive or combo plan, control decks simply cannot afford to wait another turn to play a card, since often they are looking to play their first or second spell rather than their last spell. Thus, it is often a good idea to either pad the land count or run cantrips or cheap card draw to be sure that the deck will be able to play all of its spells on curve. Essentially, an aggro deck can do things like 14, 18, 20, 23, 25, and 28 while control decks need to run counts like 14, 18, 22, 25, 28, and 30.
Especially in control decks with lots of lands, but in all decks wherever possible, you want to cram lands that do things. I know my dealer friend was surprised when Worldwake was released and the manlands were being priced around $5. After the Pro Tour, everyone realized they gave established archetypes more power, but not even all the entrants to the tour grasped this beforehand and embraced the newly available resources. Any astute manabase builder should have been all over these from the get-go. Relevant to every deck is how many threats and useful spells you have available to you. Generally the price control decks pay for having more expensive, powerful spells is that they get fewer total spells because so many of their cards must be lands. This is why cards like Windbrisk Heights, Pendelhaven, and Raging Ravine see so much play. Not only do they generate mana, but you also get very valuable effects out of them. For example, let’s say you were building a R/G aggro deck that currently topped out at 3 drops and played 20 lands. Well, there you have a deck with 40 threats and 20 land resources. However, what happens when you cut 4 spells for 4 Raging Ravines, and 4 spells for 4 Bloodbraid Elves? Well, now you have 24 lands, 4 of which are threats, and 36 spells that behave like 44 spells (unless you get smacked by a Double Negative or a discard effect). However, in general you have increased your threat density even though you’ve moved your curve upwards, a win in most books. This must be weighed against the relative speed of these threats. In the example the Bloodbraid Elf is about as fast as a three drop, but the Raging Ravine is slower. Currently this is a trade-off most players seem satisfied with. (One innovation I made with the White Weenie deck mentioned above was to replace some lands with Horizon Canopies, so that whenever I drew past the 4 lands I needed for Calciderm, I could convert them back into cards.)
Where Pendelhaven is a freebie, most effect lands come with a price, usually entering the battlefield tapped. How many of these lands you can afford to play depends entirely on your deck. In the abstract, you can look at the Magic numbers to help figure things out. Let’s say you need to cast an on curve spell on turn 3 and on turn 4, but don’t care about the first two turns. That means you will need at least two lands that come to play untapped, which tells us we need 18 lands that come into play untapped. If we’re playing 24 lands, that means we can afford up to 6 tapped lands without likely hurting our deck. On the other hand, if we plan to play spells on turn 2, 3, and 4 (all on curve) then we want 20 or 21 untapped lands, and can only afford 3 untapped lands. If we want to play lots of spells that aren’t on curve, like 2-drop, 2-drop, 4-drop, then we can afford to play a tapped land on turn 3 and change our count. As our mana count goes up, we can essentially get away with murder. Some decks can afford to play as many as a dozen comes into play tapped lands and mostly shrug them off. These things don’t just go by raw numbers of lands, color is important too. For example, if you are playing a deck that plans to accelerate with a Noble Hierarch or Birds of Paradise, you are going to need 14 untapped green sources available in your deck so you can plant them turn 1. Essentially, each turn you have to ask yourself what your mana requirements might be, and whether your lands are meeting those requirements in quantity, color, and tappedness.
Alright, what about Limited? I’ve always felt that I got mana screwed and mana flooded more often in Limited games than Constructed games, and after thinking about it I now understand why. Most Limited decks are two colors with a 9/8 split with the 9 in their more prevalent color. Well, if we multiply by 1.5, we get to 13.5 and 12 as the mana from each color. This means that the number of lands which is considered the basic sane threshold in Constructed we are falling somewhat below for our second color in Limited. Though it makes it no less frustrating when I lack my second color, I can begin to see why it happens. The next point is where one has trouble casting a spell with double or triple color symbols. If we translate from Constructed, we would want 12 (two-thirds of 18) sources to make that sort of mana happen, and 14 for triple color! All I have to say is that after crunching those numbers I will never look at a Terramorphic Expanse or other dual land in quite the same way ever again. I always thought they were helpful in splashing off-color cards, but did not appreciate how much they could smooth out your colored mana sources in even two-color decks. 8/8/1 looks much, much better to me than 9/8 does, especially if there are any double color spells in the deck. I also will more seriously regard the real cost of double and triple colored spells, especially when they cost 3 mana or less.
The only reason Limited decks work at all is because it is generally much slower than Constructed. Even very, very fast Limited decks are unlikely to kill you before turn 5, even with no resistance whatsoever. This means a lot more draw steps to find the color of mana you are looking for without imminent death. This is how people get away with splashing a card or two from another color off of only 3 or 4 mana sources. (In Constructed, this would be the equivalent of running say 3 Ajani Vengeant off of only 7 Red sources.) In Zendikar, almost nobody could afford to splash because the format was lightning fast and any stumble ended in death. In Rise of the Eldrazi, splashes are fairly common since games go longer and people get a chance to assemble the “combo” of their splash spell and their off-color mana source. In addition to the speed of the format, the slower and more controlling your deck is, the better it will be at making these sort of off-mana plays happen.
In Limited, we are essentially forced to make the best compromises possible under a constrained situation. If our curve goes up to 4, we decide we only want 16 land (equivalent to 24 Constructed), or if it goes up to high we might want more sources. However, we don’t play extra sources just because we want colors, so no matter the number we run we just have to do the best we can with what we have. For example, on 17 land we have 10 White spells, some wanting WW, and 13 Red spells, none needing more than R. There are Red spells as low as 2 on the curve. Well, in a perfect world we’d have 12 White sources and 9 Red sources, but the world is far from it so we make some compromise. We decide which spells we need to cast most, and decide that perhaps 10 Plains and 7 Mountains gives us the best odds of being able to cast most of our spells mostly on time.
Another thing to think about in all of these equations is how to count mana fixing spells and creatures. How often is Birds of Paradise worth a mana? What about Rampant Growth? Ondu Giant? Like card draw effects, we can only begin to count these spells if our curve gets up high enough to cast them naturally. Then we need to consider the likelihood of their resolution or survival. Killing a Birds of Paradise is almost always correct, so I have a hard time counting them as a mana source in a deck. However, that is contextual, since if your opponent has no removal, it will always be a mana source. One needs to plan for whether or not it’s likely to work. Knowing your format is key for knowing whether these spells will prove to be worth mana. In Limited, Ondu Giant is pretty much always getting you your fifth land, but in Constructed he might be enough of a threat/accelerant that he gets countered. The rule of thumb I’ve heard is to count such things as half a mana source, which sounds good until two halves add up to zero. (But works out often enough.)
An interesting story I heard recently is about a Pro Tour player (and possibly winner?) who ran a deck with 19 lands and 41 spells. He would mulligan aggressively for 3 lands, and then his deck just provided a constant stream of gas. I’ll try to ask Kowal and Zvi to see which one of them told the story, but I think it goes to an interesting point (much as the Red deck with 8 sources for Blightning or Planeswalker control with only 7 sources for Ajani) that you don’t always have to live and die by these numbers. In all Magic decks, you can design to maximize for different things, and consistency comes at the cost of power. During Alara block it was a joke that people would always kill you with a Woolly Thoctar they cast off naturally drawing Plains, Mountain, Forest despite the low odds that that would occur. The point is that sometimes it can occur. Sometimes you push the envelope of reasonable probability and it pays off. It should collapse just as often, but when you’re talking about a consistent Top 8, or occasionally washing out and occasionally blowing people away, maybe the high risk variance approach has its merits. I certainly lost the finals of a PTQ playing the Faeries mirror against an opponent who had miscopied his deck so that it only had 23 lands instead of the generally agreed on 25-26, but in the match he had to mulligan a bit but his deck top decked much better than mine due to the spell density. Maybe it wasn’t right, but the variance worked in his favor on that occasion.
A good manabase makes a world of difference for your deck. I couldn’t count the number of times during Lorwyn block I saw some poor PTQer sitting with a winning Cryptic Command in their hand and only UU in play. I would always ask them afterwards how many Blue sources they had, and the number was invariably somewhere between 16 and 18. They felt like they’d had bad luck, but really they’d shot themselves in the foot before they even shuffled up. I hope that these things I have learned and put down here are of help to you. As always, I’m happy to answer questions in the forums and will discuss answers to the exercises the day after the article is up.
Exercise 1: You want to build a Red-Green aggressive deck with Colossal Might, Plated Geopede, Colossal Might, Searing Blaze, and Bloodbraid Elf. What should the manabase look like? Is this a good deck for Llanowar Elf or Goblin Guide? What else do you think goes well with the manabase you have built and is in-strategy?