Chatter of the Squirrel – Threats and Answers Redux, or, Why Goldfishing Is Actively Terrible

Read Zac Hill every week... at StarCityGames.com!
Wednesday, April 29th – The reason I hate the term ‘answer’ for cards like Plumeveil is threefold. One, it makes it seem like once you’ve answered a ‘question’ you’re somehow at parity, when in reality you’re pretty far ahead. Three, the term ‘answer’ makes it sound like every card in your deck is supposed to resolve a specific question, when in reality the whole value of cards like Counterspell or Plumeveil is their inherent versatility; that is, they don’t care what question you’re asking.

Before I start, I urge everyone to read both my article on Threats and Answers and that article’s forums, as today’s column will expand on what I said last week and I want to make sure we’re all on the same page. Particularly, I’m aiming to address a few specific comments, namely:

“How do you decide what you want a slot, or your deck as a whole, to do? How do you approach these design questions? Would you rather have cards that are slightly threatening against everyone or very threatening against a few decks? When does It Depend and why? When did somebody get it right? Who is currently Doing It Wrong?”

“You’ve advanced the argument that some threats are wrong, particularly in the context of a specific format at a specific time. I think you’ve done this quite well.

I think now you want to model how the ecosystem of a format make threats right or wrong. Or whatever other factors you believe are relevant. Basically develop “how to identify good threats.”

“A threat, aside from semantics, is obviously relative, but in most cases worth discussing on scg.premium.com a threat is FoD, Doran, Mutavault, or even a Rod of Ruin. Without any other spells being cast it wins. It is the basis on which we meaningfully discuss card advantage.”

“The old use of “threat” as (roughly) “something that sits on the board and will actually kill you if left unchecked” has much more value than this article’s use of that term.”

“I agree with using the definition of threat as, broadly, a thing that will win you the game if you are allowed to goldfish against your opponent (or, to extend that slightly, that will materially contribute to that win against said goldfish). Although reactive control cards are clearly potential backbreakers, you literally cannot goldfish a win against a random opponent with a deck full of Wrath of Gods.”

The first two comments I hope to resolve in this article, and the rest I am going to attempt to refute.

I am going to start, though, by articulating the raison d’être of the article to begin with, which I had in my head the entire time but never actually articulated, and which is important to understand so that it doesn’t seem like I’m just haggling over terms. The reason I hate the term ‘answer’ for cards like Plumeveil is threefold. One, it makes it seem like once you’ve answered a ‘question’ you’re somehow at parity, when in reality (because the deck playing cards that function this way very likely possesses inevitability, or because a specific card or card-within-a-deck adds some value beyond ‘answering’ the ‘threat’) you’re pretty far ahead. Three, the term ‘answer’ makes it sound like every card in your deck is supposed to resolve a specific question, when in reality the whole value of cards like Counterspell or Plumeveil is their inherent versatility; that is, they don’t care what question you’re asking.

The issue, therefore, reaches far beyond mere terminology. The existing dialogue of ‘threats’ and ‘answers’ makes the debate seem like it’s about individual cards, when in reality the nature of a threat is determined by the specific interactions of cards-within-decks. Understanding how these interactions work is how you start to answer the questions posed at the beginning of the article.

I am kind of hating already how dense this prose is sounding, so let’s go ahead and just tackle one issue at a time.

Why It’s Useless To Discuss Threats As “Cards Which Win The Game Assuming Nothing Else Happens.”

Because things are always going to happen.

That was easy.

But okay, okay. There’s a little more to it.

“[Things] that win the game assuming nothing else happens.” This is the definition people want to use to define a threat, with Things in this case being Individual Cards. But doesn’t that definition of ‘threat’ sound eerily close to something else that’s entirely different? What if we replace ‘Cards’ with ‘Decks’?

We’ve arrived, it seems, at a pretty excellent definition of Inevitability.

We all know, though, that the way to gain Inevitability is not to be casting Rod of Ruin. In fact, Magic players (particularly the Gerry Thompson, Patrick Chapin, and Manuel Bucher of the world) tend to be pretty good at building decks that possess inevitability. The reason this definition of ‘threat’ is terrible is that if what you’re trying to do is win the game assuming nothing else happens, you have to go about it in a completely different way than using the traditional definition of ‘threat’. Frequently, decks that possess inevitability are composed of ‘answers’. And yet if a deck is composed entirely of cards, then a deck that wins the game assuming nothing else happens must also be composed of some combination of cards that win the game assuming nothing else happens. But I can build one of these decks that kills with Mistveil Plains (which does not ‘win the game assuming nothing else happens’, and indeed requires one very specific thing to happen in order to function at all) and does not contain a single traditionally-defined ‘threat’.

We realize, then, that the attribute of ‘ends the game assuming nothing else happens’ more appropriately applies to cards-within-decks and not cards themselves. That is, that the interplay between cards is what creates inevitability in actuality, is what ‘threatens’. So we arrive at the pithy conclusion I stated at the beginning of this section: that, by definition in a game of Magic, things happen.

If you’re building an aggressive deck, the challenge is to fill it with as many proactive cards as possible with which the opponent cannot constructively interact while maintaining a reasonable expectation against the remainder of the field. If you’re building an inevitability-possessing deck, you want to maximize your ability to interact with as broad an array of the opponent’s proactive cards as possible, and for as many of those interactions as possible to be value-added rather than value-neutral (if they aren’t, then you probably don’t possess inevitability).

That should answer the question of ‘how to identify good threats’. Let’s take two of the best ‘threats’ in recent memory, Wooly Thoctar and Plumeveil. A lot of people misunderstand the Thoctar. They see it as some big animal that Saito played because of its power-to-cost ratio, the ease with which it can be cast, etc. It is all of those things, and the reason that it will always be passable in any matchup is because it’s an ‘inherently’ (w/r/t the average value of a three-mana permanent, the relative size of commonly-played creatures in the Extended format, the frequency with which threats that answer it with value are played, etc) solid creature. But the reason it was such an excellent threat in Saito’s GP Singapore deck specifically was that Saito realized he’d have to beat Faerie players at the top tables. On either turn two or turn three, the Faeries deck has exactly four ways with which it can neutralize Thoctar—Mana Leak. Once the Thoctar resolves, the only ways for the Faeries player to remove it before it deals substantial amounts of damage are Venser, Cryptic Command, Tarmogoyf, and Sower of Temptation. If Zoo is on the play, the non-Tarmogoyf spells don’t come into account until Thoctar gets a hit in, and both Tarmogoyf and Sower are easily neutralized by the Zoo player’s four copies of Path to Exile, in addition to the requisite burn spells. Furthermore, Tarmogoyf is not necessarily large enough to kill the Thoctar.

Choosing a card like Vortex over Thoctar allows Faeries to race Zoo with the creatures it plans to play already. Other two-mana spells may come out sooner, but it’s much easier for Faeries to interact with those spells because of Spell Snare and Spellstutter Sprite. Wooly Thoctar was actually a tremendously-effective niche card that also happened to be a burly enough beast to add value to other matchups. Saito didn’t just play it because it was a big man in the Naya Shard.

See also, Plumeveil. Plumeveil always became insane because (until recently) no one played around it, it was bigger than everything else in the three slot, and it stuck around to buy time until the opposing player could muster an answer. When you’re playing the Cruel Ultimatum deck, all you need is time. Because most of the aggressive cards in the format actually required the aggressive player to attack—e.g. there wasn’t a Cabal Interrogator, Grim Lavamancer, or Nezumi Shortfang that would demand less conditional removal—Plumeveil was able to interact with almost the pure functionality of a traditional removal spell, with a Flametongue Kavu attached for value. Until people started wising up, the Veil was the perfect control-deck threat. It could singlehandedly dominate the Red Zone until the aggressive deck ran out of time.

Other good examples of ‘Ur Doin’ It Wrong’ include electing to play the card Rafiq of the Many basically ever. Rafiq is not a ‘bad card’. It is an excellent card, a card that, when it works, ends games quickly. But it’s a card that is very easy to neutralize if you’re at all interested in neutralizing it, a card with which it’s exceedingly easy to trade with value (Path, Terror, Broken Ambitions, Cryptic Command, Wrath of God, or even cards like Bitterblossom or Spectral Procession or Siege-Gang Commander or Cloudgoat Ranger that neutralize its effectiveness considerably) . Whenever you play that card you’re counting on your opponent to not really have anything. And if your opponent doesn’t have anything, well, aren’t you winning anyway?

Why The Ability to Goldfish Doesn’t Actually Measure Anything, and Will Normally Lead You To Make Very Bad Decisions

I had a conversation with John Treviranus at the end of the season where he was lamenting his lack of success with what was (I believe) a very well-designed Tron deck. The first tournament—at which his Tron deck was most well-positioned—his only losses came to Adrian Sullivan Ponza list, which he actually couldn’t beat but for which there was only one copy in the entire tournament. The second came at the hands of a Zoo player. John had close-to-lethal on the table and had tapped out to play a Triskelion, I think, so that he could attack for the win next turn. His opponent proceeded to topdeck Lightning Angel, and cast Angel, Might, Might, Tribal Flames to kill John from 18 out of absolutely nowhere.

I pointed out to John that tapping out there was awful, as he had a hand full of counterspells and couldn’t actually conceivably lose if he just passed the turn. That sure, Angel (of all cards) was really unlikely, but 3x Flames + Lightning Helix was a realistic possibility and there was no reason to expose himself to that string of plays, either. Trev came around, eventually, but his rationale for the longest time was “I wanted to kill him a turn quicker so he’d have less time to draw anything.”

There is no such thing as ‘time’ in Magic.

Read that again, because it’s important.

There is no such thing as ‘time’ in Magic.

Magic is a game of land drops, untap phases, attack steps, and cards drawn. If you can neutralize any or all of those variables, the quantity of turns taken doesn’t actually matter.

One of the best Magic articles I’ve read recently was Stephen Menendian thought-experiment on what would happen if you unrestricted every card in Vintage. His conclusion was that there was actually a fairly reasonably upper-limit on potential brokenness, because the more finely you tuned your deck to generate the most consistent first turn kill, the more vulnerable you were to Force of Will. The obvious conclusion to draw from this is that ‘speed’, or ‘quickness with which you can kill the opponent,’ only matters inasmuch as it allows you to dodge potential means with which the opponent can interact with you. If quickening your goldfish, at any given point, heightens rather than lessens your opponent’s ability to interact with you, you should not try to quicken your goldfish.

You only have to look as far as Pro Tour Berlin to show the divergent schools of thought regarding this approach. I feel that, despite LSV’s win, our team had the best deck at the tournament, and I believe the subsequent near-universal adoption of Mirror Entity in the PTQ season reinforces this conclusion. ManuB and company decided deliberately to maximize their deck’s ability to goldfish in exchange for the diversity of options that a Hivemaster / Mirror Entity kill could provide. The problem is that the LSV-style build of Elves cannot actually beat Wizards post-board, and is much colder to Rule of Law-style effects than a build that can simply Chord up an Entity and beat for zillions. In fairness, it’s not like ManuB etc. didn’t recognize this; they simply made a value-bet that said answers wouldn’t be as prevalent. My point is not to say that anyone is or is not correct, but rather to show that sometimes even a full turn’s worth of expanding your goldfish can be counterproductive if you’re more vulnerable to the things that your opponents can do in the time you give them.

Richard Feldman proves this theory true virtually every time he chooses to pilot a Dredge deck. And yet people continue to cling to their Flame-kin Zealots.

So why does it matter how good you are against a goldfish, if that means losing more games to real opponents?

What A Threat Really Is

A ‘threat’ should not be defined as literally anything that can kill you, because if you perceive literally anything that can kill you as ‘threatening’ you will start making terrible in-game decisions. You will Resounding Thunder a Rip-Clan Crasher when you are at 16 and have six lands on the table, only to die to the Cavern Thoctar the opponent plays a few turns later. You will fail to compose your deck with Volcanic Fallout in mind against Burn post-board when you are Mono-U Wizards, because Fallout isn’t a ‘threat’ unless you’re at a very low life total. You will play Careful Consideration and be afraid to pitch your Oona. You will focus on what doesn’t matter. You will fail to put the opponent’s potential interactions on a scale of relative importance. You will, ultimately, lose a lot of games.

“Threatening” is a scale along which you measure the cards in a given matchup that cause you to win or lose the greatest number of games. A card against which either you or your opponent doesn’t have to exert active effort to manage (e.g. Mogg Fanatic against TEPS) can not usefully be considered a threat. And, ultimately, I’m concerned about establishing a set of terms that people can actually use.

Ultimately, you hope for every card in your deck to either represent some significant threat to a particular matchup, or be such a versatile threat-neutralizer or parity-maintaining machine that it allows the rest of your deck to capitalize on its quick start or rock-solid finish.

As for what cards do that in a given format?

Well, that’s why we test, now, isn’t it?

See you next week!