In Development – Using Abstractions in Deck Design

It’s hard to resist the impulse to tinker.

It’s the weekend after a GP, and a fascinating new design made the Top Eight. You’re thinking of playing it for next week’s PTQ, but first, there are just a few cards you want to swap

That’s probably a bad idea. More than one Magic writer has given us the advice to “Just play the deck” a couple times before we start swapping things around. It’s useful to do this, as you’re likely to learn something about how the deck works, how it flows, and from there you can think more about what changes, if any, the deck would need to let you compete successfully at that PTQ, on Magic Online, or in your local FNM.

I find that just playing around with the deck for a while isn’t always enough – whether my goal is to play the deck as customized to my tastes, or to simply understand how the deck as it stands is meant to operate. When I want to take a better look at the “spine” of a deck – how it works, how it flows – I sometimes like to pull the camera back a little and take a more abstract view of the deck.

That’s what I’m going to talk about today, but first I’m going to do a little side trip to the backstory behind this approach.

Adventures in science – bar-coding a protein

I like to write a lot of “sciencey” introductions to Magic columns, largely because there’s a lot of cross-pollination between ideas I encounter in my profession and ideas I apply in Magic. Typically these relationships are by analogy and metaphor, but today’s piece is about a method that I quite literally ported directly over from some experimental work I did, into my deck comprehension and design methods. With that in mind, I wanted to spend a little bit of time talking about where it came from.

A bit shy of a decade ago, I needed to modify a protein I was studying so it would contain an epitope tag – basically a molecular “bar code,” that would let me easily track the protein’s abundance, which would let me run experiments that asked questions about just that. Normally, this kind of tagging isn’t a big deal, but the specific protein was super-twitchy, and had a tendency to fail and get mulched by the cell’s garbage disposal system if you made even one little mutation in it.

Bit of a pain, really.

There was one tiny little spot on the protein that seemed like it might be okay, but I still wanted to make a “conservative” change – like swapping in cards that don’t screw up the fundamental mana curve of a deck (more on that below).

Proteins are made up of chains of amino acids. This is nice, since it lets us treat them like words much of the time. For example, one of those “tags” is the chain tyrosine-proline-tyrosine-aspartate-valine-proline-aspartate-tyrosine-alanine (each one of those words is an amino acid) which we can abbreviate to “YPYDVPDYA,” much as we abbreviate Magic’s colors to WUBRG. The section of protein I wanted to swap a tag into looked like this:


So how did I figure out which tag to use, and where to put it? This feels a lot like figuring out which cards I can swap in, either in deck design or in sideboarding. As it turns out, amino acids, like Magic cards, have important differentiating features. For cards, it’s going to be mana cost, effect, type, and so forth. For amino acids, it’s chemical features”¦so I used that to abstract, rewriting the sequence in terms of its chemical features, so that sequence up there becomes:


(The letters and plusses and stuff are chemical traits of each amino acid. The details are unimportant for today’s column.) The tag YPYDVPDYA becomes BNB-NN-BN. Staring at the chemical traits, it was clear there were no especially good places to line the two up. Hm. But another tag, EQKLISSEEDL, could be abstracted to these features: -P+NNPP—N. Hm. A plus and then some minuses and stuff, that could work – and it did, letting me swap that tag in for another part of the protein without screwing the protein up too much.

Folding this back into deck design – how we can abstract a deck

So, some years later I was trying to figure out some changes I wanted to make in a Lorwyn Block Constructed deck ahead of a PTQ, and I found myself asking the questions that we ask when we’re trying to tinker with a deck. What do I want to put in? Take out? More to the point, what cards can I swap in and out without fundamentally screwing the deck up? Which cards need to come out to let it work?

That’s when it occurred to me that I could probably do with a bit of abstraction, much like what I’d used to handle the protein modification problem. To illustrate what I mean by abstraction, let’s take the winning deck from a recent MTGO PTQ:

Mythic (chrism315)

One possible abstraction of the main deck of this Mythic build might look like this:

7 –
8 – TT

Before I go on to add any explanation about what all those letters mean, I’d like to take a moment to say that I’m really writing about the concept of using this kind of abstraction, and the specific way you abstract a deck will entirely depend on what works for you. Much like with note-taking, I don’t think there’s any “one true way,” but it’s good think about the concept and then adapt it to your needs.

With that disclaimer, let’s return to what I’m doing in that mishmash of letters above. You’ve probably already sussed out that the numbers there are the mana curve. Basically, when I want to try to understand the spine, or basic structure, of a deck, I go through it card-by-card, abstract that card out to its role(s), and then plug it into the appropriate part of the mana curve.

So, starting through chrisM’s Mythic list in the order it was written in above, we have [card]Baneslayer Angel[/card], which I’ve abstracted as a “creature,” which gives us three “Cs” on the 5-mana spot on the curve. After that, we have Birds of Paradise, which I’ve abstracted as both a “creature” and as “acceleration,” which gives us four “As” and four “Cs” on the 1-mana part of the curve. It continues on like that through the list. Here’s the abstraction shorthand I was using this time around:

A – Acceleration (mana acceleration)
C – Creature (creatures, pretty much)
D – Disruption (countermagic, land destruction, discard, and other elements that screw up your opponent’s progression)
F – Fixing (mana fixing that does not accelerate)
M – Mass removal (stuff that kills more than one creature)
R – Point removal (stuff that kills one creature)
S – Solutions to problems otherwise not covered (e.g. Pithing Needle)
T – Noncreature threats (like a Cruel Ultimatum)
V – Card advantage generators (draw spells, either Jace)

Again, I want to emphasize that this is just how I happen to be abstracting things lately. You could pretty much go with any abstraction based on your needs, and I think you can benefit from the insight it gives you into a deck’s structure and flow. Just skimming the list of abstractions above, I’m sure you can come up with some possible different directions to go. For example, I use “Creature” as a sort of shorthand that means “something that gums up the board and whose status as a threat correlates pretty much with its mana cost.” You could also easily decide that this “gum up the board” feature is its own thing (maybe you can call it “B”) and abstract creatures into both “Ts” and “Bs” – this could be a useful in a world with Wall of Omens.

As I hope is clear from the example above, a card can be abstracted as multiple things. A Noble Hierarch is both a Creature and Acceleration. Jace, the Mind Sculptor is both a Threat and a Card Advantage Generator. Sometimes I like to note “special case” differences by using upper- or lower-case letters or some other differentiation. For example, during Lorwyn Block, [card]Shriekmaw[/card] was both Removal and a Creature, but had the special case of evoke. To note this, I put a small “r” at 2 on the mana curve, and a big “CR” at 5 on the mana curve.

So what’s the payoff for abstracting a deck design? There are probably more benefits than I know of, but the big ones that draw me to this approach are how it aids in understanding the flow of the deck, understanding sideboarding, and facilitating effective deck modification.

Abstracting to understand flow

So, say someone handed you a Boss Naya deck in the week after PT San Diego:

Boss Naya (as played by LSV)

So how is this deck meant to flow?

The answer may be intuitively obvious to you, or you may have figured it out since then, but it was not clear to a lot of the people I talked to at the event or in the next week or two after that.

So let’s see what happens if we abstract it out:

2 – CCVV

Hm. Well, that’s pretty clear, isn’t it? Do you think the deck’s flow might not be meant to be incremental, “curving out,” so much as it’s meant to immediately establish a board presence and then jump ahead to its cavalcade of super-powered four drops?

Again, this may have been obvious to you the moment you glanced through the list – and if so, good for you – but I know from talking with a lot of players that they weren’t really clear on what the deck was meant to do even after looking through that list and watching a couple games or trying it out themselves. This naturally impacts things like which hands you keep, how you play your early turns, and even whether you think you want to play the deck in the first place.

This method of examining a deck’s structure and flow is obviously no substitute for playtesting, but I find that it’s a useful base to build your playtesting on. It’s also super helpful if you, like me, like to scour the Internet for deck lists. When you find that fascinating rogue deck list that Top Eighted a PTQ in Slovenia, you probably don’t get to ask its creator how it’s meant to be played.

Abstracting to guide sideboarding

So far, we’ve just looked at some abstracting on the main deck, but there’s no reason to keep the sideboard clear of that treatment as well.

Although we should probably go into our matches with sideboarding notes, these notes are always guidelines, and we’re going to have to think about our sideboarding on the fly. Understanding the basic structure of the deck you’re using helps both in preparing sideboarding guidelines and in those on-the-fly decisions.

Consider the following full deck list from that same MTGO PTQ:

Jund (fingers1991)

This deck abstracts out like so:

Main deck

6 – C


1 –
3 – DDRR
5 – CC

Once again, the idea of being very conscious of your mana curve and how you might be changing it during sideboarding may be second nature to you – if so, good for you. However, one of the big flaws that can trip up a lot of us as we try to develop our skills is accidentally shifting the deck’s mana curve in the wrong direction during sideboarding.

Note that this isn’t just about shifting the overall curve. Usually, we’re pretty conscious of how that works, and if you’re on MTGO, you can always sort your cards by mana cost to make sure that you’re not egregiously harming the overall flow of your deck as you change things around. However, you may still be messing up what happens at each point along your curve if you aren’t careful.

In other words, it’s not just the overall curve of the deck, but the purpose of each part of the curve. In the Jund deck, above, we see that the early part of the curve is largely removal and some creatures gumming up the board. So if you’re playing against an opponent who is so impolite as to keep casting Spreading Seas on your lands, what do you swap out for those two Prophetic Prisms you’re bring in?

The autopilot move might be to just swap out two of our two mana spells to “preserve your curve.” This is a good time to refresh yourself on what each point in your curve is trying to do, and how this applies to the current match, right now. In this case, if we say that you’re playing against some form of control deck, then yeah, you probably don’t need to have so much removal at the 1- , 2- and 3-mana spots on your curve.

In fact, given what you’re playing against, the 1-, 2-, and 3- positions are essentially the same, which is our cue to (1) say that yes, we can swap out some of our removal for the Prisms and (2) exit the abstraction and choose which cards from that pool to remove. As we return to the real world of actual, non-abstract cards, we see that this suite of removal is:


Well, against our imagined control opponent, I’m probably happiest clipping two Lightning Bolts, so we can imagine doing that.

As before, this may be intuitively obvious to you, but I promise you that just as many players out there follow the incredibly strong, very tempting path of “I’m swapping in a two-mana card, therefore I must swap out a two-mana card.” A dose of abstraction can help pull us away from this precipice and let us make more thought-out choices in our sideboarding.

Abstracting to guide deck design and development

Much as abstraction can help prevent poor decisions in sideboarding, it can keep you on the right track during deck design or development.

One of the subtle traps in tweaking a deck is that short series of incremental changes that leave you with a deck that simply does not work. This the situation where you’re kind of disappointed with your deck’s acceleration, so you add in some Rampant Growths, and then you decide that you’re not fond of Rampant Growth, so you add in Nest Invaders, and then when you’re thinking about how Nest Invader isn’t doing quite the job you want in taking down opposing creatures, you decide to run Putrid Leeches after all, and all of a sudden, you’ve replaced your mana acceleration with a 2/2 creature.

It’s especially disappointing to realize you’ve done this once you’re already under way in the tournament.

Having a sort of “running abstraction” of your deck lets you check to make sure you haven’t fundamentally altered its flow away from what you wanted it to be”¦or lets you troubleshoot the deck’s structure to understand why it isn’t working the way you intuitively expect it to.

Consider the following deck:

URW Planeswalkers (TPABHuK)

I abstracted this main deck like so:

1 – RRR

So this deck’s abstracted curve very clearly cries out “Not in the face! Not in the face!” Which is to say that it’s largely removal and disruption – trying to stall the opponent’s game plan.

Even so, we can see that there’s a consistent spine of removal running all the way up the curve – there’s no point lacking a significant contribution of removal. The upshot of all this is that if we make the decision to start swapping cards around, we’ll want to consider whether that spine of removal is critical to how a planeswalker deck performs – and if so, how we interpret potential card swaps as viewed through this lens.

For example, what does this view of the deck tell us about the choice between [card]Jace Beleren[/card] and [card]Divination[/card]?

As I’ve abstracted it, that swap is basically a wash. Obviously, there are more detailed concerns here – Divination gives us two cards right away, and Jace offers more long-term benefits as well as the “take a bullet for me” effect all planeswalkers have. But given our basic structure of packing some extra card advantage into the 2- and 3-mana slots, Divination is just fine.

What if we wanted to instead clip two copies of [card]Gideon Jura[/card] and go up to a full four copies of [card elspeth, knight-errant]Elspeth[/card]? Well, that changes the 4- and 5-mana positions from this:

5 – RRTT

To this:

5 – RT

The upshot here is that our “creature” count is a little richer, our “threat” count is the same, and we’ve lost two cards from our “removal” total in that range. If we’ve decided and playtesting has shown that this aspect of removal is key to the deck’s structure, we want to be very away aware of, and perhaps wary of, this kind of change.

This kind of abstracting can also help diagnose potential problems when the deck isn’t quite working. Especially if you have a tendency to throw a bunch of good cards together and then find yourself disappointed when the deck stumbles, generating an abstraction that works for you can highlight potential problems, whether it’s a lack of sufficient early removal, an uneven distribution of threats along your curve, or simply a lack of unified purpose in the deck’s card suite.

All tools are just tools

I hope I’ve emphasized sufficiently that this is a handy visualization tool that helps us focus on a deck’s essential structure. It’s certainly not an end-all in deck comprehension or design. These are both very holistic areas, where we’ll benefit from a lot of different point of view.

I’ll also reiterate that the specific abstractions I used in this article were not the ones I used last year and may not be the ones I’ll use next week. There are obvious places to think about things differently. For example, I just skipped out on discussing how we might think in the abstract about lands, or mana bases. Should you include creature lands on the curve? Perhaps they’re a key part of almost every contemporary Standard deck, so it might be worth remembering that your deck has a threat “at” five mana in the form of Raging Ravine, perhaps with an asterisk to note that it’s going to cost you that effective five mana every turn.

That said, I do think some form of abstraction is tremendously powerful in deck comprehension and design. It keeps us focused on what matters, whether we’re breaking down a new list, designing our own, or even sideboarding during a game.

Do you have an abstraction that you’ve found particularly useful? Let us know in the comments.


Scroll to Top