The Long and Winding Road – Mythbusting Complexity

Read Matt Elias every week... at StarCityGames.com!
Monday, March 16th – Before beginning this week, I want to make it clear that this isn’t meant to be an attack on Jeff Phillips or even a full rebuttal of his thesis, but rather a discussion and clarification of what I consider to be an oversimplification of what it means to be, or to strive to become, a superior Magic player…

Before beginning this week, I want to make it clear that this isn’t meant to be an attack on Jeff Phillips or even a full rebuttal of his thesis, but rather a discussion and clarification of what I consider to be an oversimplification of what it means to be, or to strive to become, a superior Magic player. I know I previously wrote that this week would be a review of the Philadelphia PTQ, but I’ve had this piece nagging at me for a few weeks now, so I thought it best to just get it out of the way.

I’d like to start by taking a look at a quote from Zac Hill that Jeff referred to in his article:

“The thesis is this: the value of a given action can be measured by the number of favorable interactions it makes fungible relative to a theoretical maximum number of interactions of which your deck is capable, or the number of an opponent’s favorable interactions it correspondingly negates. That’s quite a mouthful, so I’ll summarize: You want to maximize the impact of what you’re doing and minimize the impact of what your opponent is doing, and you achieve that by interfacing with the opponent [in ways favorable to you].”
Zac Hill, Interaction Advantage

Here is Jeff’s response to Zac’s thesis, and his own thesis from State of Perfection:

My Thesis reads as such: The superior player overall gains advantage as the number of complex choices within the game increases. Thus, it is the goal of the superior player to create as many complex choices as possible. Conversely, the goal of the inferior player is to minimize complexity, so as to minimize the advantage of the superior player.
~Jeff Phillips, State of Perfection

I want to discuss the idea of complexity in Magic, specifically Jeff’s idea that the goal of a superior player is to create complexity. The theory here is that given a high level of complexity in the game, a weaker or inferior player is more likely to make an incorrect or suboptimal decision ultimately resulting in defeat. Therefore, it behooves a superior player to create as much in-game complexity as possible, as this additional complexity increases the likelihood of victory for the better player.

When I first read this theory, it rang true with me because I recalled a specific example in my past where I found this statement to be accurate. I have always enjoyed playing Chess, but I don’t get to do so very often. While not completely rubbish, I am not particularly skilled at the game. Because I am a “gamer” I have always felt capable of beating the “average Joe” at Chess, but anyone with play-skill specific to Chess itself is probably going to be a better player than I am.

In college, one of the guys in my Fraternity was moderately skilled at Chess, and we used to play semi-regularly. When we first started playing, he would beat me around 75% of the time because he would see lines of play that I completely missed. Over time, I realized that a pattern was emerging in the games that I won — I was throwing out my evaluations of the pieces outside of King and Queen, and ranking everything else as either Pawn or Other (other being comprised of Rook, Bishop, and Knight), and trading one-for-one whenever possible. In the early game, whenever any piece that was in a position to make a one-for-one trade, I would make that trade. As I started to do so in a very aggressive fashion, my win percentage went up and began to approach 50%. Interestingly, this approach really irritated my friend in that it wasn’t “proper” Chess strategy at all.

What I was unknowingly doing was accomplishing two things. First, I was dramatically minimizing the complexity of the game state by eliminating the number of pieces that could cross the board with any type of speed or aggressiveness, and limiting the chance that I would accidentally walk into a trap. Secondly, I was throwing my opponent off of his game plan by, in his view, haphazardly trading pieces in a suboptimal fashion. Because I was willing to make trades that most people wouldn’t, and was in fact doing so aggressively and purposefully, his traditional lines of attack were much less effective because I wasn’t reacting in the “correct” way.

It seems logical that I might see a relationship between my experience with Chess and with Magic. Just as I learned to be the aggressor and trade one-for-one, similarly many Magic players will choose to play an Aggro deck that features minimal complexity and maximum aggression to end the game before the control player is ready to set up his defenses and his traps (his complexity, if you will), winning before he can effectively utilize what might be his superior play-skill.

There is undoubtedly some validity to this idea, and I think much of the bias against playing Aggro decks comes from this line of thought (control = complexity = sign of a superior player, and conversely aggro = simplicity = sign of an inferior player). If I am playing Faeries in Extended, it is quite possible I will achieve a board position that involves a Vedalken Shackles, Glen Endra Archmage, Mutavault, and Riptide Laboratory in play, along with cards in my hand that might include Mana Leak, Cryptic Command, Stifle, Vendilion Clique, and Spell Snare. The number of possible options and permutations such a situation creates is exceedingly difficult to navigate and provides my theoretical Aggro opponent many possibilities for error and misplay — opportunities for them to make the “loser choice” as Jeff might say.

There was quite a bit of discussion during last year’s Extended season regarding Dredge, and why it was a “poor” choice for a superior player. Specifically, the thinking was that Dredge basically “plays itself” — there is little to no room for a superior player to “out-play” his opponent. Using Jeff’s thesis, we could suggest that this is because Dredge does not create complex situations for its opponents. The deck requires specific plays to be made each and every time with little to no variance outside of what is actually “Dredged” into the Graveyard, and as a player you have no control over this. Essentially piloting Dredge requires the following skills:

• Making correct Mulligan decisions
• Sideboarding correctly, including knowledge about how to beat specific hate cards and how to sideboard “blind” after winning Game 1
• Being able to count (so that, you know, you Dredge the card with the highest number printed after the word “Dredge” first, and can count to three to notice when you have enough creatures to Flashback Dread Return)
• Correct understanding of the triggers the deck creates, specifically those of Narcomoeba and Bridge from Below
• Knowledge of key cards in the format as well as potential hate cards and tutors for those cards, to maximize use of Cabal Therapy

I don’t mean to minimize the amount of practice that goes into mastering these skills, as outside of being able to count, they all require practice with the deck, especially learning to sideboard and to make Mulligan decisions. Interestingly, many articles and forum posts discussing Dredge (across all formats) will reference the fact that most Dredge pilots are “unskilled” and make many mistakes, which seems to prove that practice with the deck does create better results. As you can see, though, the majority of these skills don’t relate specifically to the game state itself, but rather to the mechanics of the deck and how it functions. Dredge allows a practiced pilot to gain percentage, and the deck itself is so high on the power axis that it can often just “get there,” but there is little ability for a pilot to actually out-play his opponent or create a complex situation resulting in a win rather than a loss.

We’ve established that Dredge doesn’t create complexity — once you start vomiting cards into your Graveyard, most of your lines of play are as obvious to your opponent as they are to you – but does that make Dredge the intentional choice of the inferior player? If we apply Jeff’s thesis, aren’t we suggesting that Dredge players are, by default, either a) inferior players (Sorry, Tomoharu Saito! Apologies, Richard Feldman!), or b) are making an incorrect deck choice that fails to reward their superior play-skill? I don’t believe this to be the case — it seems more likely that a thesis suggesting that superior players should always seek complexity is either flawed or incomplete.

Let’s look at a deck like Belcher in Legacy. The skill level required to play Belcher is similar to that of Dredge, although in some ways the deck is even less skill-intensive – the games tend to actually be less interactive and take even less time, and Mulligan decisions are far easier. During Legacy Champs last summer, I had to play against a B/U/G Faerie / Fish deck twice, in round five and then again in the Quarterfinals. This deck ran Daze, Force of Will, Standstill, Spellstutter Sprite, and Bitterblossom, to name a few powerful and interactive cards. This type of deck creates significant complexity in the game state as the game goes on: Do you break Standstill, and when? Does my opponent have Force of Will, Daze, or both? Do I need to bait out Spellstutter Sprite and if so, what cards am I willing to throw away?

The problem is that Belcher doesn’t care about ANY of that nonsense. Standstill? Irrelevant. Generally speaking, it is too slow to have any impact on the game until the game is already “over” for one player or the other. Bitterblossom? Completely irrelevant. Daze? Can be relevant in games where Belcher is on the draw, often totally irrelevant otherwise. In fact, against this deck, the Belcher pilot has only one question to ask: Do you have Force of Will, and another Blue card, AND will you play your Force of Will correctly — because if you assume the Belcher deck is trying for a Belcher and they are on the Empty the Warrens plan instead, you’ll often still lose. Furthermore, after sideboarding, what tools will you worry about from the Belcher deck? Is it bringing in Duress? Thoughtseize? Red Elemental Blast? Xantid Swarm?

I chose this example because Belcher does the exact opposite of adding complexity to the game state: it tries to bypass the game state completely in order to win before complexity ever becomes an issue. I don’t consider this the same thing as “minimizing complexity” in the way that the Extended Burn deck minimizes complexity, for example, but rather, it is choosing a completely different axis on which to function. Stated another way, it is actively dictating the field of battle.

This strategic deck choice wasn’t a concern about my “inferior” play skill but an intentional choice that I believed would maximize my chance to perform well at the event given inexperience with the specific format and an over-representation of Aggro decks. Why give my opponent’s deck a chance to function at all? If many players at this event were likely to choose Aggro decks to prey on Threshold, it made strategic sense to play a deck that had a decent match-up against Threshold AND an almost auto-win against most Aggro decks. Even if you view a deck like Belcher as one that minimizes complexity, doesn’t this suggest that in some instances, simplicity is the correct choice?

Compare these examples to the thesis from Zac Hill Interaction Theory, because I think one statement there in particular is very telling. Zac states that “the value of a given action can be measured by… the number of an opponent’s favorable interactions it negates.” What Belcher is doing, in effect, is bypassing almost all of the favorable interactions that its opponent can use. In this specific example, it takes away Counterbalance/Top, Standstill plus man-lands, Spellstutter Sprite and Bitterblossom, and so on. The only interaction that matters in this match-up is Force of Will plus a Blue card, and even that can be trumped after sideboarding. Despite what many people think, Belcher has a decent match-up against these types of decks, generally much better than you’ll find in play-testing. Why is this true?

If you play-test against Belcher, you generally have an idea what hands you can and cannot keep. You can keep a hand with Daze and Brainstorm but no Force of Will on the play, but would probably Mulligan to 6 for a shot at Force of Will on the draw (or, alternately, keep a hand that can beat a turn 1 Empty the Warrens or turn 1 Belcher without activation, such as one with Engineered Explosives and/or Pithing Needle, or one with mana acceleration, Moat, and Runed Halo). The problem is that you’re not recreating an actual, game one, tournament-authentic situation — “game spots, game shots” as they say in Basketball.

Let’s say you’re playing in a Legacy tournament with a Counterbalance / Top deck, and draw this hand. Do you keep, or Mulligan?

Daze, Brainstorm, Counterbalance, Sensei’s Divining Top, Standstill, Flooded Strand, Underground Sea

What about this one, which approximates what the B/U/G deck I mentioned above might see?

Standstill, Mutavault, Bitterblossom, Daze, Brainstorm, Underground Sea, Spellstutter Sprite

In these specific cases, the complexity of the Blue Control deck is actively working AGAINST the pilot, because it is drawing him to powerful and complex lines of play that may, in fact, be completely irrelevant in this specific match-up. Dredge is able to function in much the same way, in that it is extraordinarily favored to win game one, and regardless of the preparedness of the opponent’s sideboard, often simple variance will carry Dredge to victory in the match no matter how skilled the opponent.

Given this information, I believe that all things being equal, Jeff’s thesis is accurate but incomplete. Allow me to clarify. I believe that as the available card pool is minimized and the average length (in mean turns, as opposed to clock time) of any particular game increases, superior players will seek complexity in the game state to add to their likelihood of victory by maximizing the chance that an opponent makes a mistake. I think this statement is specifically most accurate in environments like Sealed Deck, Draft, and Block Constructed. In these situations, a skilled player is highly likely to know the entire card pool as well as how all of those cards interact (or at least, the interactions between the cards that “matter”), while the less-skilled player may have only passing knowledge of some cards and probably hasn’t seen many of the interactions possible within the population of cards legal for play.

To some extent, I think this line of thinking is often valid in Standard, as Standard tends to have only a few main Linears active at any one time. As tournaments occur and the Metagame shifts, new interactions come up (such as Fledgling Mawcor or Voice of All) and then fade once they become a known quantity, as players learn to interact with these cards and the advantage in running them decreases.

More than anything, I think this theory hinges on how quickly decks can move from Phase 1 to Phase 3. In formats like Vintage and Legacy, decks can reach Phase 3 as quickly as the first turn of the game. Consider Vintage… An Oath deck can resolve a first turn Oath of Druids using a Mox and Forbidden Orchard with Force of Will back-up, moving immediately to Phase 3. A Workshop Aggro or Stax deck can drop a turn-one Trinisphere and immediately lock its opponent out of the game, preventing them from ever exiting Phase 1. Legacy offers similar situations, such as Dragon Stompy locking out an opponent’s ability to cast spells, or Belcher, Ad Nauseam, and Dredge winning on the first turn of the game, or Counterbalance decks setting up a CB/Top soft lock on turn 2. In these instances, the choice of the superior player has little to do with complexity, and everything to do with angle of attack, with limiting meaningful interaction and dictating the field of battle.

Here is another specific example to illustrate this idea. At Grand Prix: Chicago, my 2nd round opponent was playing Goblins. This is a great example of a deck that seeks to minimize complexity, and does so by cheating on mana costs (Goblin Lackey, Goblin Warchief, Aether Vial) while simultaneously stunting the opponent’s mana development to keep them in Phase 1 (via Wasteland and Rishidan Port). I was playing my Royal Painter deck, a Painter’s Servant toolbox deck. I was on the play game one, and led off with a Polluted Delta. My opponent played Goblin Lackey, and in response I cracked my Delta, fetched a Volcanic Island, and played Brainstorm. I untapped and played Ancient Tomb into Trinket Mage, for Grindstone (I already had Painter’s Servant in my hand). I now had a blocker for his Goblin Lackey. On his turn, he played Goblin Piledriver and passed the turn back to me. I untapped, played an Island, and cast Painter’s Servant off the Ancient Tomb and Grindstone off the Island. The game was now over, as I had both Stifle and Force of Will in hand. Although I was functioning as a Combo-Control deck, the fact that I was able to play a turn 2 blocker essentially won me the game by keeping my opponent in Phase 1. In addition, once my combo is in play, I can win during my Upkeep, trumping the use of Rishidan Port.

Before game 2, my opponent was furiously sideboarding what looked like 8-10 cards. I was expecting some combination of Red Elemental Blasts, Pyroblasts, and possibly Pithing Needle or Artifact destruction for my combo. Despite the fact that Stifle is relatively weak against Goblins, I kept in the full play-set and both Phyrexian Dreadnoughts, as my opponent had no idea that this combination of cards was in my deck. Sure enough, in game 2, I played out turn 2 Dreadnought plus Stifle, and won on my fourth turn. My deck had completely changed the field of battle — I moved immediately to the endgame on my second turn in a method of attack my opponent didn’t anticipate. I had become the Beatdown, and in effect won the game by minimizing complexity.

Even if he was a Legacy “regular,” my opponent probably would not have expected this line of play, because Painter’s Servant decks aren’t “supposed” to run Stifle and Dreadnought. Similarly, I can’t even count how many games I won during the GP weekend by trumping my opponent’s Force of Will with a Daze that my deck “shouldn’t” have anywhere in its 75. In this specific instance, my game 2 plan was actually to minimize complexity by making my opponent’s sideboarding less relevant, and never reaching a game state where his deck’s plan of attack would matter. This choice of simplification has nothing to do with me being an inferior player; rather, it has to do with my use of strategic advantage and initiative in deciding actively the way in which the game would play out.

Here is my clarified version of Jeff’s thesis: “Provided that both players are competing on a similar field of battle, it is in the best interest of the superior player to create complexity as a strategic advantage, while it is in best interest of the inferior player to eliminate complexity.” Although incomplete by definition, I consider this statement to be considerably more accurate. It accounts for the fact that as you move backward through Magic history and increase the available card pool, there are a number of strategies that completely bypass complexity, or actively punish it — in other words, field of battle trumps the theory of complexity. I suppose the follow-up thesis would be that the truly superior player strategically bypasses complexity through positioning and initiative; the superior player seeks not complexity as a specific goal, but rather may use complexity as a tool (among many) to transition from Phase 2 to Phase 3.

Thus ends my attempt at theory — next week we’ll return to our regularly scheduled look at Extended.

Matt Elias
[email protected]
Voltron00X on Xbox Live & SCG Forums