The highlight of my experience at U.S. Nationals this year, aside from the fact that my long-time BFF won the 2008 Vintage Championship, was my conversation with Magic creator, Dr. Richard Garfield. Even before arriving at U.S. Nationals, I was euphoric about the possibility of picking Dr. Garfield’s brain about Magic, and Time Vault in particular (just ask my travel buddies — I was laughing maniacally for a good deal of the trip). I recounted that part of the conversation in the first part of my Vintage Championship tournament report. But, as I alluded to there, that was only a small part of a wider ranging conversation.
When Richard Garfield indicated that Time Vault and Twiddle were intended to interact, he explained that it was because Alpha was designed with as many â€˜combinatorial possibilities’ as they could muster. In the debate over Time Vault, the question that kept coming up was what the actual “intent” of the card was. Indeed, the original announcement declaring the removal of power-level errata explicitly stated that Magic Rules Manager, Mark Gottlieb, “is a strong, strong believer that cards should have functionality that matches their printed intent as often as possible.” For me, this raised a series of questions: what is “printed” intent? Is that different from “design” intent? If so, how? How do we discern this intent? And why should it matter? What difference does it make?
As a lawyer, the notion of “printed intent” struck me as somewhat incongruous. There is the text and then there is intent. A textualist ignores intent and focuses on the text. An originalist focuses on the intent of a provision and construes the text in a manner that is consistent with the purpose behind it and its original meaning. The notion of “printed intent” suggests a conflation of the two concepts.
In any case, many people speculated that it was clearly the intent of the card that Time Vault should only ever untap if you skip a turn first. When asked to give some reason for this speculation — some evidence in support of it, people pointed to the thematic function of the card: that it served as a “store” of turns, and that as such, one must store a turn on it for future use.
Given that Mana Vault has the same templating and the same thematic concept (a store of “X” where “X” is mana rather than turns), and yet worked with Twiddle, that argument was pretty silly. Mana Vault was clearly a store of mana, yet you could use other cards to untap it. Nonetheless, people still “felt” that this was the design intent of the card, including, the Rules Team. Feelings are not reasons. Sometimes, feelings can point us to reasons. When I tried to pin people down, they seemed to resort to nebulous grounds such as “their feelings” about the card. The Rules Team was no different…
Our current wording makes some assumptions about printed intent, that’s for sure. The printed text is slightly ambiguous about how untapping Time Vault is supposed to work. The key question we asked ourselves was, “When this card was made, was the intent that it be incredibly easy to skirt the drawback?” We went with “No.” Does that make the card feel weak? Yes, but we feel that initial intent is captured, regardless of how people have been playing the card for the past several years.
This just served to underscore my fundamental concerns about this entire manner of inquiry. Whoever was reading the card seemed to impute a design intent based upon their reading of that card. This was backwards. Since a person’s reading of a card is influenced by many different factors, including notions of power level (“was the intent that this be easy to skirt the drawback?”), it was really a circular argument. It wasn’t a search for “intent” so much as it was a conclusion drawn from one’s assumptions regarding that intent based upon a reading of the text.
There are further complications.
In the first year of law school, students are quickly put through a mental wringer when it comes to understanding concepts like “intent.” In both Torts and Criminal Law, two major first year courses, questions regarding the “intent” of the defendant are central to legal analysis. In criminal law, there are a host of mental states (mens rea) that describe various levels of intent, and there are archaic and unusual doctrines that describe subtle differences between them.
Constitutional law is concerned with intent, in a similar way to the design context. When analyzing a Constitutional provision or a piece of legislation, one question that may shade the interpretation of a given provision is the intent of the legislature or the framers. Yet, that enterprise is a risky one. How does one discern the intent of hundreds of different lawmakers voting on a given provision for myriad reasons? In other cases, the law is being tested or applied in a situation under which the authors of the law or framers of the Constitutional provision could never have foreseen. If a situation is unforeseen or unforeseeable, what role does intent play in resolving an ambiguity? Is the internet protected “speech” under the First Amendment? Are automatic weapons “arms” as contemplated by the Second Amendment? These are questions that the Framers of the US Constitution could not possibly have imagined.
These considerations raise further questions, such as whether it even makes sense to ask Richard Garfield his “intent.”
Magic cards may have a designer, in the sense of the person who initially proposes a card concept, but I doubt that many cards see print — if any – as they were initially conceived. The Development Team, the Rules Team, and the Future Future Leagues all have a say in what a card will do and how it will work. The Design Team is merely one part of the overall equation. How are we to discern the design intent of a Magic card, something that people sometimes futilely refer to, when there are possibly dozens of people with different conceptions of the card, motivations, hopes and goals who have a hand in the cards design? It’s an impossible task.
In addition, like law, Magic cards have unforeseen applications. That’s the beauty of Magic. No one could have possibly imagined that Black Lotus could be used to play Tezzeret the Seeker in 1994. The unforeseen interactions are part of what makes Vintage so interesting. It is almost entirely built on unintended interactions. Bazaar of Bagdad plus Dredge. Illusionary Mask plus Phyrexian Dreadnaught. And so on.
While bald speculation as a method of analysis was common in the debate over Time Vault, one question that might have a bearing on the design intent of Time Vault is whether the original developers anticipated or came across the Twiddle and Time Vault combo. The more likely it is that they came across the combo, the more likely it would seem that the original makers of the game were comfortable with it, since they printed the cards as is.
I mused, in my conversation with Dr. Garfield, that the Time Vault and Twiddle interaction was intended, in the sense that it was foreseen, but that the Instill Energy, Animate Artifact, and Time Vault three-card combo that caused the banning of Time Vault was probably unanticipated and unintended. Surprisingly (or not), Dr. Garfield cautioned me from drawing that conclusion. He implied that those interactions were probably foreseen as well. He explained that the original designers were not concerned with power level. Unlike today, players were not expected to have access to all of the printed cards so there was little to worry about in terms of cards being overpowered. In any case, there was a clear demarcation, after the creation of the DCI, between design concerns and tournament rules.
I then told Dr. Garfield that I believe that much of the fun from Magic, and particularly Vintage, comes from the use and discovery of unintended and unforeseen (or foreseeable) interactions. Interactions like Illusions of Grandeur and Donate, Cabal Therapy and Academy Rector, Intuition and Accumulated Knowledge are what make Vintage so interesting. There are over 15 years of accumulated printings interacting in completely unanticipated ways. I told him that I couldn’t wait to use Black Lotus to play cards that would be printed 15 years from now. Indeed, Dr. Garfield remarked that emergent properties are a key part of Magic’s success.
At that point, I became even more excited, almost agitated. Dr. Garfield was employing a systems theory principle, which only served to underscore his understanding of this incredible game that he had created and heighten my appreciation for his marvelous achievement.
Emergence is the systems theory principle that complex systems have features or qualities which are not present in the parts of a system; it’s a rejection of the idea that a system is merely the sum of its parts. Rather, a system is different (either or more or less than) than the sum of its parts. This is a very simple concept, and yet it is not easily grasped. And even when the concept is grasped, it is quickly lost in the thick of analysis.
The concept of emergence is a rejection of reductionism, a characteristic of analytical reasoning, that things are merely the sum of their parts. Most of us accept the idea of synergy — that two things interacting can be more than the sum of their parts. Both Black Lotus and Yawgmoth’s Will (or Intuition and Accumulated Knowledge) are amazing, but they are even better together. They synergize.
But emergent properties don’t have to be more than the sum of the parts of a system. One need only reflect on the behavior of an angry mob to realize that even rational, intelligent people sometimes behave differently as part of a crowd. Similarly, two great cards might interact poorly. Trinisphere and Yawgmoth’s Will do not work well together.
Dr. Garfield followed his remark by saying that this is a critique he has of modern design. The designers of modern magic sets try too hard to make the individual cards complicated. It is the interactions among cards that create complexity in Magic. Even a list of simple cards will be complicated if there are enough interactions.
Consider this format, proposed by Patrick Chapin:
The only legal cards are:
Seal of Fire
Simian Spirit Guide
Heart of Ramos
There is no limit on the number of each card you can include in a deck, but otherwise the usual rules of Magic apply, including 60 card minimum and a 15 card sideboard.
This list of cards is quite clever. Even with this bare minimum number of cards, metagame dynamics emerge with myriad interactions. One can build a deck that is more controlling, a deck that is more aggressive, and a deck somewhere in between, with a rock-paper-scissors metagame quickly emergent. In a sense, this hypothetical format reduces Magic to its essential elements, and helps us see how complex it can be even with a minimal number of interacting parts. Now think about the fact that there are 9000 unique cards in Magic.
The point is that the reason systems are complex is not because of the quantity of parts in a system. As I said, it is the potential for interactions among cards that create complexity in Magic, as the hypothetical format above shows. The assumption is that if all of the individual parts are simple, then the system could be understood as the aggregate of the individual parts, and therefore understandable in those terms as well. This is a reductionist way of thinking, and it’s wrong.
Systems are complex because of the interactions among parts. The interactions create synergies, economies of scale, and totally different behavior. This is why, from a systems perspective, a system is not sum of its parts; it’s different from its parts.
If one does not understand the basic systems theory principle of emergence, then creating overly complicated cards could actually have the paradoxical result of reducing the possibilities (emergent properties) and complexity in Magic rather than expanding them.
At this point, my excitement had subsided and was replaced with a moderate state of enthrallment. Given the ingenious achievement that is Alpha, one would know that Dr. Garfield was nothing less than a minor genius, but now I was officially blown away.
More than anything, systems thinking is a shift in emphasis. It draws our attention to the one thing that tends to get overlooked in reductionist thought: the relationships between parts of a system. From a systems perspective, the relationships — the interactions and connections — between a system’s parts are as important as the parts themselves. The design focus, Dr. Garfield was arguing, should be on maximizing these interactions and connections. It is these interactions which produce the emergent properties in Magic. It is not that designers have to be particularly conscious of all of the potential interactions, just as a Lego designer does not foresee all of the possible shapes and structures that one could engineer using the same set of Lego bricks. It is the combinatorial possibilities that the designers must seek to foster.
But the application of systems thinking can be used to underscore the importance of the interactions among cards at an even deeper level. In fact, within the game of Magic, there are only interactions. As a corollary, there is no such thing as a “Magic card.” This is a difficult ontological reality for many Magic players to accept.
A common response would be: How can that be true? There are, in fact, Magic cards. I’ve seen them. I’ve held them. And I use them to populate my decks and play spells. Of course cards exist in a vacuum. I draw a Mountain, and play it. There — that is a card which I have played.
Yes, there is such a thing as a Magic card. A “Magic card” is something you open from a pack or put in a binder. And yes, Magic cards are described in the rules. But within the game of Magic there is no such thing, at least not in any meaningful sense. In the game itself, a card is merely a part of an interaction with other cards or with the rules of the game itself. Lands do nothing without something to play with them. Spells do nothing without the lands to play them. Equipment does nothing without creatures, and so on. Each card requires a context in order to operate, and that context is the series of interactions, whether it is with the rules, other cards, or an opponent. Therefore, within the game of magic, it is the interactions that are the relevant metric, not the cards.
But doesn’t that make my claim that there is no such thing as a “magic card” irrelevant and best and unnecessary at worst? Let’s assume a card’s worth is ultimately defined not by the effect it generates, but by how that effect interacts with the effects generated by other cards. Even if that is true, why say that there is no such thing as a magic card? Doesn’t that go too far?
The answer is quite simple. How are we to tell what effect a card creates by looking at the card in isolation? It’s impossible. The very notion of an “effect” implies some type of relation or context (which implies a set of relations). Moreover, cards do not generate effects without using other cards. With only one exception, every single card in Magic requires other cards to do anything relevant. The only card I can think of that does anything by itself is Dryad Arbor, and even it requires a context.* Thus, the very assumption that a card “creates” an effect is mistaken. Cards interacting produce effects. (see figure below). The effect is only generated through the interaction. Cards in isolation cannot produce that effect.
In short, there is no such thing as a Magic card in the game of Magic. There are only relationships and interactions. In Magic, everything is an interaction or a combination. Lands interact with spells. You tap lands for mana to play spells. Spells interact with each other. A Lightning Bolt can be used to kill a Savannah Lions, which can be stopped by a Counterspell. (See figure 1) A player can equip a creature. Creatures can block other creatures.
We talk about two- and three-card combos, but in truth, all Magic cards interact with other Magic cards. It is the sheer quantity of interactions that defines magic design possibilities. It is the various design possibilities that therefore define a metagame.
Why does this matter? It matters for many reasons. As an analytical matter, it corrects a major gap in Magic theory. This correction has many practical implications for our thinking and our practice.
In Vintage, from time to time, someone will pose the question: what’s the “most powerful” card in the format? My answer to that question has always been the same: what do you mean by power? How do you define and measure it? By its ability to contribute to game wins? If so, how do you measure that? It’s frankly impossible.
By now we see that there is actually analytic error made in asking the question. If there are no such thing as cards, only interactions, then our reference to a “card” is either mistaken or a sloppy way of describing the sum of the interactions of that card with other cards. One of the reasons that Black Lotus is considered to be such a great card is because it has virtually infinite synergies. It works with almost every spell in Magic (except cards like Bridge From Below).
What we understand as powerful cards are actually synergies. It is the interaction of cards that makes a system more than the sum of their parts. And, as you no doubt understand, decks are complex systems with their many interacting parts. One only need see how the interlocking engines of TPS work to see that this is true.
Sometimes synergies between cards are hard to see. Not all synergies are as obvious as Black Lotus and Yawgmoth’s Will, Tinker and Darksteel Colossus, Brainstorm and Polluted Delta. This is because of the complexity of the systems we are analyzing.
Wheel of Fortune has synergies with cards that are common in Vintage, specifically cards like Yawgmoth’s Will and Tendrils. The burst of card advantage creates lines of play that lead to Tendrils because it generates large amounts of cards advantage which translate easily into storm and mana.
For the last few years, Brainstorm was an automatic four-of in any blue deck. Yet, that was a relatively recent trend. Before 2002, Brainstorm saw much less play. If you go back and look, although some of the Invitationalists ran a couple of Brainstorms, and a few niche decks ran four-Brainstorms, it was a far more limited card in terms of its usage. It wasn’t until Onslaught was released that Brainstorms went from seeing a marginal amount of play to being the most-played Blue card in Vintage. This is the essence of what I mean by relationships rather than parts. The importance of Brainstorm wasn’t the card Brainstorm itself, but the interaction of Brainstorm with other cards. This is logically proven by the fact that Brainstorm was everywhere after the printing of Fetchlands, but saw very little play before the printing of Fetchlands.
The illusion is that a particular card is powerful, when in fact it is the sum of the synergies with that card that gives a card value. This illusion can cause players to overlook important cards. It is easy to overlook cards labeled as “objectively bad” and fail to see how they might create very powerful synergies. Cards like Illusions of Grandeur and Donate are a case in point. When we label cards as “bad” or “good,” what we are actually drawing on is the sum total of synergies, filtered through the lens of our experience and projected onto the card itself. This is an analytic error, and causes us to overlook cards that may not be as generally synergistic, but have a potent application nonetheless.
This illustrates another point: which is that the inclusion of powerful cards is not actually simply the input of powerful cards, but the inclusion of powerful synergies. When we talk about cards, what we actually referring to a series of interactions, often prominent interactions, such as Black Lotus + anything, Brainstorm + Polluted Delta, Tinker + big expensive artifacts, Merchant Scroll + Brainstorm, and Goblin Welder plus cards like Black Lotus and Thirst For Knowledge.
In fact, and this might be true of other formats as well, the structure of a format is really a function of the most powerful and prominent synergies in that format. In Vintage, some synergies are so powerful and therefore so prevalent that they constitute the structure of Vintage. A good example is the Gush-Bond engine.
When we talked about the Gush-bond engine, we weren’t talking about Gush, or even just Gush and Fastbond. We were talking about a host of interactions. Specifically, the Gush-bond engine was an engine that tethered together Merchant Scroll, Brainstorm, Ponder, Gush, and Fastbond together, aided by a host of restricted tutors and draw. This was a bundle of synergies. And once Ponder saw print, it became a very fast engine that could accelerate and combo towards a endgame almost without fail from as early as turn one, so long as you could find Fastbond.
Once it was identified and used, it was ported into other decks. Eventually, what was once an engine used by GAT became an engine used in: Oath, Dark Ritual storm combo, Painter decks, and even Bomberman. It wasn’t that one or two cards were ported, or even a half dozen cards. What was ported was a bundle of synergies, a structure of interactions.
The analytical error, repeated time and again, is to describe these synergies as a product of a card rather than the interaction among cards. Thus, the reason that the question about what the most powerful card in Vintage is flawed is because it is reductionist, when things in a system cannot be reduced to their parts.
The reductionist error emerges in another kind of discussion: causality. From the previous two systems theory principles, emergence and the importance of interactions, we come to a new paradigm for understanding causation. Consider this example:
“I Thoughtseized his Force of Will and won because of it.”
This person is claiming that Thoughtseize was the ultimate cause of game win.
There are a number of different points I wish to underscore here.
The first is that in complex systems causation is always multiple. For any given outcome, there are many causes. The claim that Thoughtseize “won the game” is false.
Consider the example of Thoughtseize in more detail. There are at least a dozen identifiable causes contributing to the outcome of the game, even in the very narrow scenario provided:
1) The fact that Thoughtseize was included in the deck.
2) The fact that Black mana was included to play Thoughtseize
3) The fact that Thoughtseize was drawn in a timely fashion.
4) The fact that Black mana was drawn to play it.
5) The fact that Black mana was played to cast Thoughtseize.
6) The fact that Thoughtseize was played.
7) The fact that the opponent drew an otherwise unstoppable win condition or that you had one which could now resolve.
8) The fact that said win condition was in the deck.
9) The fact that the opponent drew of Force of Will.
10) The fact that the opponent included Force of Will in their deck.
11) The fact that the opponent included other Blue spells to use force without paying mana for it.
12) The fact that the opponent drew at least one of those Blue spells.
More importantly, this example ignores all of the decisions that led up to this position, from deck construction, to mulligans, to sideboarding, to all of the in-game plays that brought us to this point.
To conclude that Thoughtseize was the ultimate cause of the outcome of the game is circular. It assumes that Thoughtseize was the ultimate cause of the game win because everything else is a given. Of course, if you hold everything else constant – ceteris peribus – then there are no other causes. But everything else is not constant.
The reductionist error here is a common one. Magic is a series of nested interconnected, complex systems. You can never isolate one particular element as an ultimate cause for the outcome of a game. Every single game/match outcome is the product of dozens, if not hundreds of causes, often interacting. Mutual and multiple causation is also non-linear and not summative, it cannot be reduced to a linear frame of reference. And being non-linear means that the multiple causes behave different when they interact than if they were alone on account of synergies and other emergent properties.
If this mistake is so common, why does it recur with such alarming frequency? The answer is in our heads. Human beings are notoriously bad at identifying causation in complex systems. In the first place, studies in human cognition demonstrate that people tend to assume that each effect has a single cause and often stop looking for additional explanations when a sufficient cause is found (i.e. Thoughtseize).
Second, the typical cues to causality, such as proximity of cause to effect in time and space, lead to great difficulty in complex systems. For example, when a person living on the penthouse goes to turn on the hot water spigot to draw bath water, there may be a five-minute delay as the water travels up the pipes to the bath. Although we have little difficulty in recognizing the cause of the delay, we have great difficulty seeing it when there is additional complexity involved. In systems, causation is multiple, effects are multiple and nonlinear, there are many interconnections, and delayed and distant consequences. An example that I enjoy talking about is Patrick Chapin claim that Ancestral Recall wins 97% of the games in which it resolves, including a turn 1 Ancestral. The point is that Ancestral Recall — whenever played — creates such potent card advantage that it allows the pilot to turn all of those advantages into further advantages culminating in a game-win. We can quibble about the accuracy of such a statement, but the point that there are often profound time delays between events is important.
The claim that Thoughtseize won the game here is actually an illusion of human cognitive functioning. There is a long literature of research in psychology on how humans put events into narratives. These narratives often have dramatic arcs. It’s simply how the brain organizes information. Thus, we attribute primary causation or ultimate causation to proximate events, such as Thoughtseizing a Force of Will.
The truth is what we are doing is filtering information. What is often going on is that the play that we highlight as “most important” is really just a play that is particularly well made, and therefore memorable, and fits our narrative of the game, or, more likely, is contemporaneous to the end of the game. Either way, it is a trick of the mind, and not an accurate reflection of empirical reality.
The reductionist, linear thinker might reply: “Okay, so, even if we can’t isolate or identify one ultimate cause, we can at least say that some are more important than others.”
My answer to that is, yes, but it depends on what we mean. Important for what? Just because there are multiples causes doesn’t mean that we can’t tell which causes are most relevant for the objectives which we are pursuing. It just means that, as an analytical matter, there is no such thing as an “ultimate cause.” Clearly, from a perspective of trying to understand how to improve a deck by questioning the presence of certain cards or evaluating our play, questioning the presence of Black mana may make less sense than focusing on the play of Thoughtseize. But from a strictly analytical perspective of causation, they are both causes, and neither can be described as “more important.”
Our discussion of the definition of the irreducibility and immeasurability of an effect of a Magic card in isolation should explain why this is true. In terms of producing particular game outcomes from a strictly factual point of view, the presence of mana to play Thoughtseize or the fact that Thoughtseize was drawn cannot logically be thought to be less significant than the play of Thoughtseize itself.
Even though we can separate these various causes, we can’t say which is more valuable to us. Remember, the play of Thoughtseize is actually a synergy between several cards. The very notion of a synergy is that A+B > A and B in isolation. Since synergies are value beyond the individual components, you can’t actually separate out the components and measure them individually. The synergies only emerge once the components are put together.
Here is an example that should drive home my point:
Take hydrogen and oxygen. Both elements are gases. When they combine into water, they change form entirely. How could you say which element is more important? You can’t. It’s the combination which creates the valuable outcome.
You could search hydrogen and oxygen individually forever looking for the property called “wetness” and you would never find it.
The reductionist fallacy is in thinking otherwise. Using systems thinking means that we need to focus more on the interactions and relationships that are overlooked by a reductionist/linear frame of reference. A more accurate analysis of causation using systems theory principles does not preclude me from seeing what’s important. It just leads me to a more accurate model of what’s going on.
Systems theory, above all, is a shift in emphasis. Magic is not about the discrete components. It’s not about the cards themselves; it’s about the space between the cards — the relationships and interactions. Systems theory is a paradigm with a different philosophical lineage and a different set of assumptions about knowing. The contrast is often between Newtonian science and Quantum physics, with the latter having parallel assumptions about the impact of the observer, the dynamic, evolving nature of reality, interconnectedness, and the appropriate unit of analysis being relationships rather than parts. The failure to recognize this leads to all kinds of difficulties. The restrictions of Ponder, Brainstorm, Merchant Scroll, Flash and Gush all at the same time is a case in point.
The DCI restricted each card, and presented its analysis of each card in isolation. Thus, the discussion of Flash focused on its ability to lead to turn 1 and turn 2 victories. And the discussion of Merchant Scroll emphasized its power as a tutor. In fact, the cards restricted on June 20th were highly interconnected. The restriction of any one of them would have had an impact on the use and abuse of the other four. Specifically, in the discussion on Gush and Flash, nothing was said about the impact that the restrictions of Merchant Scroll and Brainstorm would have on the use of those cards, particularly the uses that the DCI cited as troubling.
This is a major gap in reasoning. If the restriction of Brainstorm, Ponder, and Merchant Scroll effectively rendered Flash unplayable, then the restriction of Flash was completely unnecessary. The same is true of Gush.
If they had adopted a systems view, rather than a reductionist, linear perspective, they might have accounted for the ways in which each individual restriction would have lessened the need for the others.
After over 15 years of existence, our understanding of Magic remains remarkably shallow at times as revealed by our statements, our analysis, and our actions. These errors are compounded by the arrogance with which we think we can know and understand the world around us. A person, particularly a smart person, may think their analysis does, but more than likely, they are wrong based upon some incorrect assumption or overlooked factors. Applying particle physics, we might know the position of a particle, but we might not be able to know its momentum very well. Knowing both is actually impossible beyond a certain level of precision. That to our knowledge is a fundamental rule, and it actually is pretty applicable to a lot of things in life – you might know what you want to find out, but it’s actually not knowable under the current paradigm. And a person’s analysis is quite limited by cognitive, psychological, intellectual, and other limitations. This is not simply true of magic players, but politicians, military tacticians, economists, and the Federal Reserve.
A generation ago, a Pulitzer Prize winning journalist wrote a sardonically titled book “The Best and the Brightest” to describe how the hubris of the most intelligent men in America led to the disastrous and bloody course in the Vietnam War. The managers of that war included men like defense secretary Robert McNamara, the youngest and highest paid Harvard Business School assistant professor of his era, and the first non-Ford to lead Ford Motor Company at the age of 44. Yet, it was their arrogance that heralded disaster. The very notion of “domino” theory, the ultimate linear metaphor, belies the linear-reductionist analysis they used, wrongly, to a very complex, non-linear world.
That humility should temper our analysis. In the forum thread following Richard Feldman recent article, “What’s Wrong With Being Results Oriented,” Mark Young excoriated Richard for suggesting that we should be somewhat results oriented. He stated:
You just can’t estimate (whether you made a correct decision) from your results. The two have no dependence upon one another. My yearly salary is paid because there are millions of people who think otherwise.
A number of forum denizens, including myself, pounced. Clearly, there is some dependence between the two. The evaluation of whether a play is right or wrong will ultimately come down to whether it helps us win games or not. There is no such thing as a â€˜right’ or â€˜wrong’ play outside of the context of a game or a match, and therefore outside of the context of the objective within that game/match. At a very fundamental level, the way in which we know whether our decisions are correct or not depends upon results. Maybe not from a particular match result, but from the range of results that constitute our experience of Magic. It is from results that we construct a mental paradigm about the connections between decisions and outcomes with which we draw inferences about causation. Without results, there is no evidence from which to build a mental map of these relationships.
A person’s analysis of results, no matter how smart they are, is likely to be flawed in some way. Human beings simply cannot correctly map complex systems. My article last year at this time makes that quite clear and documented the numerous ways we screw it up. If I can’t trust Alan Greenspan, our President, or Robert McNamara, why should we trust anyone to get anything perfect in a complex system?
Results have the benefit of reflecting the entire outcome of multiple, overlaid complex systems. A person’s interpretation of those results does not. A very good example is one that I am fond of giving in relation to magic: far too often, the analysis of whether a decision is ‘correct’ or not very narrowly defines ‘decision’ as the in-game decision-making. Major decisions such as deck-choice, and deck design often fall outside of the analysis, yet they are possibly more important than in-game decisions. Results cue us into not simply in-game decisions, but reflect all decisions. A person’s analysis will often hold factors constant ceteris peribus, when other factors aren’t constant.
Let me give you an example that illustrates my point.
Let’s say I’m playing dredge, and I play a turn-1 Bazaar of Baghdad. We will assume that my opponent does nothing relevant on his turn, for the purpose of this discussion. On turn 2 I draw a card before using the bazaar. Then, if I am not a moron, I realize that I have absolutely made the wrong decision. As you know the dredge deck can still win from that position (in fact, correct me if I’m wrong, but I think most dredge decks can win that same turn, despite the misplay, if they draw/dredge the nuts).
Now, it won’t always be that simple. But my point is, using results to determine whether or not that play is a wrong is incorrect. You have to be capable of analyzing the play later – either later in the game, or afterward – and determine if it was correct (in terms of offering you the best chance to win).
Here is my response:
Suppose, for whatever reason, that we repeated that error many times throughout a tournament. Suppose further, oddly enough, every game in which we made that ‘mistake’ we won, and every game in which we activate Bazaar in our upkeep, before the draw step, we lost.
One explanation could be your basic (overbroad) suggestion that proper play is ‘independent’ of game/match outcome. However, perhaps the results are suggesting that there is something else going on that is not captured by our analysis.
The basic assumption that underlies your analysis of the value of that play is that plays that help you goldfish faster will lead to more wins, and are therefore, in general, more correct. At the very least, that assumption would have to be scrutinized. It is the results that help us question whether that assumption is true or not. We may discover any number of possible alternative explanations. Perhaps that line of play was more likely to induce errors by the opponent somehow which actually created a net gain which offset the slower goldfish. Perhaps that line of play somehow created internal interactions, such as a more stable goldfish by having a larger hand size at a critical moment, interactions that we were unaware of.
I have no quarrel with the notion that we should be capable of analyzing plays later, but I also think that our analysis should be tempered by the humility that our analysis is likely quite limited and very possibly flawed. Something is invariably going to be overlooked. Our assumptions are contingent and should be constantly scrutinized. Results have the benefit of forcing us to rethink our paradigm. The notion of what play is correct cannot be uncoupled from results.
The ideas I have just presented represent a radical departure from the most common ways of thinking about Magic. I am certain that many of you will have questions or take issue with my claims. Rather than expand or elaborate on those claims here, as a way of preemptively addressing those criticisms, I will instead continue the conversation in the forums and answer your questions and address your points there.
Join me next week, in the New Year, as I turn to Legacy!
Happy New Year!
* Okay, so I guess Pact of the Titan also does something by itself.