Missed you guys last week!
When I wrote Interaction Advantage, I knew I’d want to take some time after my initial excitement had dissipated to sit down and test the theory, to make sure it really works, to explore some of its intricacies and examine some of its quirks. To see if it actually accomplishes something. And so I knew that my follow-up article would go beyond some of the simpler examples contained within the piece itself and apply the theory to deck construction, to in-game decision-making, to the construction of a framework that helps you understand a format (e.g. the defining principles outlined in Understanding Shards of Alara Sealed). The question, though, was: How am I best going to do that?
My initial thought, for which many of you asked in the forums, was to construct a The Play’s The Thing-style situation-by-situation analysis. I’d lay out the framework in this week’s Chatter, let some discussion simmer in the forums, maybe form some suppositions about how existing theoretical frameworks would handle the issue. Then, next week, I’d approach these problems from the perspective of Interaction Advantage and see what light that perspective shed, see if it revealed anything new.
I was excited about the possibilities this approach would enable, but ultimately it presented some irresolvable obstacles. First of all, much of the value of interaction advantage comes from its effect on your strategy, not your tactics, so breaking down a game to its fundamental turn didn’t really accomplish anything. Related to this is the influence Interaction Advantage has on deckbuilding; you can’t really make any headway in that area when all you’re doing is focusing on what happens within the game. More importantly, though, was the realization I had when I was about halfway through the second scenario – it’s just not efficient. Zvi was and is a master, and yet his column was fraught with errors that ultimately led to its demise. I was spending so much time fact-checking and unintended-escape-covering that the point became muddled in the details.
Unfortunately, all of this happened a couple of hours before my due date, and as a result I missed you.
A collective tear was shed, I’m sure, et cetera.
What I’m going to do instead is take some more sophisticated play-or-deck-construction possibilities, things about which there isn’t necessarily consensus, and simply articulate Interaction Advantage’s perspective on the issue. Before I do this, though, there are a couple misconceptions I want to clear up.
First of all: Interaction advantage is not not NOT some kind of “option advantage.” The point is not “well, you can do more things, so you must be getting ahead,” or that your cards simply interface with one another in a whole lot of different ways. That is why the part about your deck’s maximum capability and your opponent’s deck’s maximum capability is written into the definition of Interaction Advantage. If you are playing against Reanimator, and you have ten different ways of making them discard a single card on the first turn, you may have a lot of “options” and you may even possess a lot of feasible “interactions.” But you are not gaining interaction advantage. Similarly, one can conceptualize some kind of deck that cycles cards between zones like a true champion; a poster in the forums mentioned a hypothetical Promised Kannushi deck, for example. But insofar as raising dead until actual cows return to the actual barn doesn’t really interface with the opponent, or allow you to build to something that interfaces with the opponent, it’s not interaction. It’s just Action. Another way of thinking about this that, when you measure your net gain or loss of potential interactions, they have to really be “potential” events – that is, possess the real capability to manifest themselves within a game. For all practical purposes, a Bog Hoodlums, though technically a ‘card,’ is really a set of null interactions in the Type 1 format. No conceivable game you are going to play in that format is going to provide that card the opportunity to interact meaningfully.
Another trap people fall into is the tendency to start evaluating things on a quality matrix, which is exactly the type of thing that theories like Card Advantage and Interaction Advantage aim to avoid. The problem is not just that ‘quality’ is subjective. It’s that you need, by definition, some kind of theory to be in place in order for you to articulate and measure ‘quality.’ Some people are comfortable with ‘quality’ being something that ‘you just know,’ and that’s fine when we’re using examples of 5/5s and 3/3s and no one talks in polysyllables. We all think we know a good card when we see it. But I invite you to go back and read initial set-reviews containing evaluations of cards like Skullclamp, Arcbound Ravager, Tarmogoyf, Academy Rector, Dream Halls, Yawgmoth’s Bargain, Goblin Lackey, Aether Vial, and Glimpse of Nature. There’s a reason all of us authors have started to shy away from doing those: frequently, they are downright embarrassing. Remember when Street Wraith was supposed to change everything? Magic players are notoriously bad at evaluating card quality until the community itself, through a rigorous naturally-selective process, determines what is and isn’t good. Nobody had good Ravager Affinity decks when Darksteel was spoiled, and it took months for Vial Affinity to catch on. Nobody played Psychatog for months except as proxies for their Shadowmage Infiltrators. So when you act all coolly sardonic and ironic and cast jaded, critical gazes at authors who try to discover ways of evaluating cards objectively, when you roll your eyes and say it’s self-explanatory, when you criticize people for ‘wasting time’ and for aggrandizing their respective places within the community, I’m sorry, I call.
But, to segue ungracefully, Interaction Advantage does allow you to construct fundamentally useful frameworks of relative card quality, at least as measured against a hypothetical norm. You can use Interaction Advantage to understand, for example, why it’s so good to trade your twenty-third pick for their third-pick even if it doesn’t gain you card advantage. In a field full of competent players, you can assume that basically every deck is going to have N sum total of possible interactions. The ‘worst’ card in a deck can be defined by the limited number of interactions it creates or eliminates. This is one of the reasons removal spells are so ‘good’: they both interact broadly and negate broad interactions, and they do it simply. Rather than engineer some intricate, board-spanning attack phase, rather than wait a turn and cross your fingers (unless Everything Has Haste), you just point and click. So when you trade your 23rd card for their first-pick – when you Coma Veil their Woolly Thoctar – you’ve upped significantly your remaining total of available interactions. Furthermore, this perspective does actually alter the way you construct your deck. I made it to my last PTQ finals against Cynic Kim largely on the back of Hindering Light, which I maindecked because I noticed an incredibly high number of removal spells circulating around the table. I didn’t even have to waste an active pick on the card, and yet it dealt with everything from Soul’s Fire to Cruel Ultimatum. It’s very limited in application – possesses limited interactive potential – which is why you don’t windmill slam it. But when it does work, it negates severely an attempted interaction on the part of your opponent, throwing their gameplan off entirely, and draws you another card (set of potential interactions) in the process.
Interaction Advantage also allows you to see why the Viashino Skeleton scenario was so poorly handled on the part of Brian Kowal’s opponent. Basically, for those of you without premium, BK’s opponent was reluctant to trade a high-quality card (Akrasan Squire) for a low-quality one (Viashino Skeleton) and lost the game because of it. This is the problem with a card-quality framework. Furthermore, from a card-advantage framework, you’re trading one-for-one, which may or may not be good, so you don’t really gain much in the way of insight there. But with IA, you can realize that if you don’t get in your damage before that Skeleton’s ability goes online, every single one of the potential-interaction-advantages that the Exalted ability creates – essentially, replacing the entire ‘blocking’ axis of the combat step, those blockers’ power, toughness, abilities, etc., with “Sacrifice [this]: Gain life equal to the power of target attacking creature” – can be mitigated by the simple act of discarding a card, and that you’re losing value every single time you cast another Exalted creature because it doesn’t affect the board or change the way the opponent plays. He’s going to block and discard a card whether you’re attacking with a 6/6 or 7/7, in other words, and so you need to interact positively with that ‘terrible’ Skeleton while you still can. Now, it may still be good for you if he discards a card to prevent damage, but it’s unarguably bad for you if he blanks the text of your creatures’ most relevant ability. You want to interact with that Skeleton on the turn that is best for you – the turn where its interactions are the most limited.
In fact, this notion of limited potential interactions allows you to understand why the AIR deck in Extended even works in the first place. All that deck tries to do is work within the principle, inherent to how modern Magic works, that you’re not going to be able to do all that much, comparatively, on the first turn of the game. It aims, through its speed, to circumvent entirely the way your deck wants to interact. This can happen directly via Blood Moon – your lands are all of the sudden unable to interact with your spells, and you’re left sitting there doing nothing – but the principle is still the same with something like Demigod. Faeries, for example, is a deck that exists and is successful mostly because it can interact favorably with virtually everything an opponent can do in the early game. One-drop? Deal with it, get a creature. Two-drop? Deal with it, pay half the mana. But against AIR, all of the sudden the sixty cards’ worth of potential interaction is reduced to zero if, for example, they cast a turn 1 Deus on the play. It’s not that they won because “wow it’s hard to deal with a 6/6 when you have no lands.” They won because literally every single one of your cards is blank*; your deck contains no means of interacting with the opponent at all, at that point, because every single one of your cards neither constrains your opponent’s interactive potential nor expands your own. The game is land, take six, sack a land, go until it ends, and nothing at all can possibly change that. What is happening when you side in Chain of Vapor and/or Echoing Truth to combat the deck, then, is that you’re expanding the number of cards in your deck that interact with AIR. Their deck operates on the principle that you won’t be able to interact with them, and you respond by making that premise false. The more cards in your deck that nullify the axis upon which they intend to operate, the more likely it is that you are going to win. By the same token, as I talked about in the article on Next Level Gifts, if you are playing a Phase-3 control deck, it’s absolutely essential that an opposing deck doesn’t possess an axis of interaction upon which you cannot operate. I needed to sideboard in a Tormod’s Crypt versus AIR because if all they wanted to do was cast Demigod after Demigod, I had no way of stopping it. I had no way of interacting with that particular plan.
Finally, Interaction Advantage reveals why Manuel Bucher revelation about not Esper Charming Bitterblossom even when you have the chance turns out to be valid. To understand this, we have to understand the means through which Five-Color Control can conceivably interact favorably with the Faerie deck (if none of these prove to be true, you have a matchup that can’t be won strategically, like Jushi Blue versus Kuroda-style Red in Mirrodin-Champions Standard, or Extended Life versus Next-Level Gifts today). Fae, obviously, can counter Five-Color’s big spells with Cryptic Commands and sometimes Remove Souls or Broken Ambitions. It can strip them from the hand with Thoughtseize and occasionally mise a midrange spell with a Spellstutter Sprite. Occasionally, it can Time Walk with Mistbind Clique or present a threat on the end step with the same. It possesses initiative with Mutavault, ensuring that Five-Color Control must cast the first spell, if that turns out to be relevant. Jace Beleren is a source of card advantage with which Five-Color Control cannot easily interact. Finally, Bitterblossom ensures that Fae always has initiative after a non-Cloudthresher mass-removal spell, makes certain that Mistbind Clique is always capable of interacting with the opponent, and provides a means of killing the opponent relatively quickly without a major main-phase investment in mana.
Fae’s interactions are limited primarily by mana and ‘cards.’ Spellstutter Sprite, Mistbind Clique, and Scion of Oona all require other cards to be in play in order to do anything in this matchup, meaning that their number of potential interactions decreases exponentially the more pieces of the puzzle that you deal with. This problem is magnified by the number of potential cards they have whose interactive potential is limited in the first place: redundant Bitterblossoms, Agony Warps, Sowers of Temptation. This is magnified by the fact that Faeries can’t resolve a single spell and expect it to win them the game, while you can, and you are capable of dealing with most of their threats for one or two mana, while they are not capable of doing the same with yours. It’s clear, then: you have the advantage in terms of mana and cards, meaning that you have inevitability, meaning that you are the control deck, meaning you should win the long game assuming your deck is capable of interacting with their threats.
With the list that I posted, you’re very capable of dealing with a Mistbind Clique; you have 4 Broken Ambitions, 2 Terrors, and a Bant Charm to deal with it in an ideal world, and 2 Plumeveils and 4 Cryptic Commands to get you by. You use similar cards to deal with Mutavault, which is their other primary threat. They are left, then, to killing you with Bitterblossoms and Scions, for which you have 2 Pyroclasms, 2 Cloudthreshers, and 2 Wraths of God to combat. How on earth, barring an early Jace, is this matchup possibly good for them, then?
Exactly.
Your deck is favored in almost every single inter-card interaction you can conceive, the notable exceptions being Thoughtseize and Jace. As the game progresses, their interactive potential dwindles and dwindles, due to blank cards and redundant copies. Yours, meanwhile, blossoms, as more cards yield more mana to create more and more relevant interactions. But you have to get to that point.
When you deal with an early Bitterblossomn, you are ceding your IA. All of the sudden, their midgame Bitterblossoms become interaction-producing cards again – probably the best interaction they can throw at you! Meanwhile, your destruction of a Bitterblossom has cost you two cards and has tapped you out, meaning that Faeries is free to deploy a threat, or, crucially, to cast Cryptic Command on your next attempt to broaden your interactions. Once Faeries hits four mana, until you cast a card-drawing spell, you are a worse control deck. They can present threats on the end step and you have no means of guaranteeing that you hit enough land drops to out-land them in the midgame, even though you play a couple more. Variance happens, and one or two lands is not the difference. In other words, by choosing to Charm the Bitterblossom, you open up their ability to interact with you along an axis you’re unprepared to combat. It’s as if you’ve gone into their deck and shuffled up more interaction-enabling cards.
Your deck is built with the interactions that Bitterblossom enables in mind already. Of course, you Broken Ambitions it on the second turn, because Broken Ambitions serves a threat-negating function in your deck. The function of Esper Charm, however, is to enable favorable interactions – not to constrain them. Basically every single other theoretical framework fails in its ability to articulate this. Card advantage says that both uses of Esper Charm are fine. Card Quality, or some sort of intuitive threat-assessment-based-on-in-game-synthesis, is going to tell you to kill that Bitterblossom because it’s their best card against you bar none. Only when you look at interaction advantage – that Esper Charm enables velocity, which enables land-drops, which make your Counterspells superior, which forces Fae to cast threats, which allows you the window to resolve more card drawing, which starts the circle all over again – do you see the big picture. The way you win this match is to facilitate more and more favorable interactions – better Broken Ambitions, better Cryptic Commands, fewer and fewer dead cards on your end with more and more dead cards on theirs. And when you play the games out, they go exactly like this. The games you lose are largely because of Jace, and you lose those games precisely because it’s difficult to interact with him as he sits there and gives their deck the same advantage I just articulated your card-drawing spells allow you to possess.
It really is crystal clear.
What I’d like to happen now is for y’all to start presenting situations in the forms, a lot like Craig Wescoe did last week, where IA may break down. Where you might be skeptical. And I want to employ this framework and see where it takes us. I think the examples I’ve used today have been sufficiently sophisticated to silence naysayers who say the model only applies in incredibly simplistic situations. That said, it’s not a Grand Unified Theory, nor does it attempt to be. It’s a tool, a useful and specialized tool to be used in our arsenal along with many, many others. There are places where it may break down.
I’m eager to see what you can dig up.
Until next week…
Zac
*Actually, this is not quite true; Land into Mox into Sprite is typically the only play that can alter the outcome of the game, at that point, but you have to have something to do at three mana, or four mana plus another Mox, even then. The point is still the same.