Measuring Magic is a tricky business. None of us want to be playing “Bad Magic.” If we didn’t care about such things, we’d probably never go to a tournament. While not every tournament player is hungry to win and prove their skill (or improve it), most are. Even for those that are in the room for the pure love of the game want to be “Good” or to get “Better.”
In some games, these things are easily measurable. We can look at hard and fast statistics, and have a sense of it. While it is often still the subject of argument (particularly about players in games from different eras), the statistics in professional sports give a very solid measuring stick. We can say that Michael Jordan is the best, or Tiger Woods. There are still a lot of spaces of stickiness, but very strong cases can be made to support a point.
In Magic, the measuring stick is a lot more difficult. Which player was the best at their peak? Probably Finkel or Budde, right? From a statistical standpoint, it looks as though Budde has it, but others will argue that Finkel simply can’t be touched. The statistics would give it to Budde, but for me, I know that the game of Magic is so complex that it seems as though those statistics can only barely contain the beginning of answering the question. Other questions are even harder. Who is the best deckbuilder ever? Who can make the best sideboard? If someone is Better at something, then something must be Worse. And once we get to that point, it stands to reason, both logically and by just looking about at the world of Magic, that something must be Bad. What deck is Good and what deck is Bad? What play is Good and what play is Bad? (And should I put a trademark on Good and Bad?)
When we hear terms like Good and Bad being bandied about in articles, they are often such mushy affairs. Sure, we want to be Good, but is there a monolithic sense of what Good is? Can Good be distilled into something that will be usable for everyone?
We can all agree that Jon Finkel is Good, and we could probably all agree that Christopher Eucken is Bad (sorry, Chris, but a 984 Limited rating is incredibly bad). So, we should try to be more like Finkel and less like Eucken, right?
What the hell does that mean?
The Problem with Success
We would probably all play a lot better if we had our own electronic Jon Finkel sitting over our shoulder, giving us advice. Even a rusty Finkel is a good Finkel. And a Finkel that has been rehoned is very good at reminding us just how good he is. While I might not have bet on a Kuala Lumpur victory, I know that I was expecting good things, and his finish is absolutely no surprise.
So, our electronic Finkel is over our shoulder, talking to us, and correcting our play to conform to his own. “Attack with these two creatures, hold that one back, tap these two land.” At a certain point, it is absolutely possible that we might have our electronic Finkel giving us less and less advice, as we become ever better Finkel emulators. Maybe we get so solid at it that we qualify for the Tour and loan our precious Finkel emulator to someone else.
Without it, we do pretty well for a while, until we come across a new situation. Gatherer tells me that Extended currently has 4,888 cards, so the possibility of a new situation coming up seems incredibly reasonable. Without our Finkel-guide, we stumble. Even being better than we were before, our crutch made developing certain skills completely unnecessary.
If the measure of Goodness is just success, we are playing Good Magic when we are playing with our electronic Finkel. But we might as well just have had an accurate Magic Eight-Ball guiding our play. What are we without it? And is success even a good measure?
A few examples would make it seem clear that it shouldn’t be. Take the classic example of the Bad play resulting in a win. In the very early game, they search up the wrong land with an Odyssey sack land, and can’t cast their turn 2, 3, and turn 4 Tarmogoyf. Still, they are doing good work with a pair of Kird Ape and a Dark Confidant. You Deed, and clear the board, and lay out an impressive Ravenous Baloth on turn 5. They finally draw the land that they need, and play out three Goyfs over the next two turns. In the back and forth, you manage to kill two of them, but the third one goes all the way. If only the Deed had killed just one you would have been able to manage the other two. Or, simpler…
Your opponent forgets to dredge back a Life from the Loam, despite the many cards they could have drawn with it, and has no means to bring it back until the next turn. This did mean that they happened to topdeck a Devastating Dreams, and cast it for 6, locking them out of Loaming for a while, but destroying you as well. Or, from another game…
You raise in an early position with a shooting for a low gut-shot straight. You hit it, and no one beats your hand, and you win big…
Success isn’t even a particularly good measure of a deck. In two tournaments, I was the only player in the room playing a particular deck, and I won the whole thing. Does that make the deck I played the best in the room, or moreover, in the format? No. It very well might be the best deck in the format, but there is no way to be able to measure it on that. Similarly, if I give a deck to twelve people, is how they do a measure of the quality of the deck? Is another deck given to twelve different people in the same room measurably better or worse, based on the average results of the deck? No.
The factors are complex. Who received the deck? What did they play against? How well did they play? There are certain things that can definitely be called a resounding success. Many years ago, I gave two people my copy of PT Junk, and out of the two-hundred or so people in the PTQ, they met in the finals. I’m going to claim that that is evidence for success. But it isn’t proof. Looking at the most recent Regionals, one can see another measure of success. Both Patrick Chapin and Richard Feldman wrote about new decks, Korlash dot dec, and OSS, respectively. While personally, I think that Patrick’s deck is the better deck, and I prefer it for other reasons as well, both decks succeeded in the hands of players that Patrick and Richard had no way of knowing personally, and both decks qualified players. Again, this is evidence for success. But it isn’t proof.
Even as I would argue it is much stronger evidence than the evidence from my two decks, we have to remember that these things are two sides of the same coin. Success can be misleading. Is a successful deck Better because it has had many wins? (Well, how many people were playing it?) Is a successful deck Better because it has a high average finish? (Well, how good were the few players playing it?) Much like Good play, we have to be cautious about trying to reduce the analysis of a deck to simply, “it won.” Many, many things can cause a win. That doesn’t mean that they are right or Good.
On Empiricism
If you look at many of the most successful Magic players out there, you’ll find that they don’t necessarily know much theory. They know the game from an entirely different place. Whether it is their gut or their head, they don’t necessarily have the kind of background in certain skills that are often sold to you, here on this website and others, in theory. This doesn’t mean that they aren’t without weapons of some kind.
Some people have an incredible natural talent. I’ve been privileged enough to live in Madison, Wisconsin, and I have to tell you that this town has seen an awful lot of natural talent. Mike Hron is the “newest” of the people here to make Top 8 of a Pro Tour, and he is hardly alone. All told, we’ve had four people from Madison in PT Top 8s (bonus to anyone who can name all four in the forums), and I’m not going to even cheat and count Mark Justice, occasional Madison resident. I know all four guys to greater and lesser degrees, and let me tell you, I wouldn’t go to any of them for thoughts on theory.
For a player like Jon Finkel, he might not need to put in the same kind of work as another player to achieve success, but he’s still going to be well served by it. If any of you listened to the fantastic coverage of Kuala Lumpur by Rich Hagon and Bill Stark, you’ll have learned that Guillaume Wafo-Tapa spends something like fifty-plus hours a week on Magic.
I can’t speak to what level they are informed by theoretical concerns, but it seems clear that a huge portion of Wafo-Tapa’s success can be put down to his empirical knowledge of game situations. Empirical knowledge comes from seeing it happen. It is the kind of knowledge that comes from not merely believing that a matchup is 70%, or that a sideboard card is effective, but from actually testing it out. In some ways, we could think about empirical knowledge in Magic as coming at the problem of being Good from an inductive reasoning side. Inductive reasoning draws conclusions about the nature of reality by taking many examples from reality to draw a larger conclusion. This can be incredibly accurate (“Every fire I’ve gotten near is hot, therefore fire is hot”). It does, however, require either a lot of trials to really be useful as a good predictor, and the less of those that one has, it requires a larger and larger creative/intuitive/logical leap on the part of a person to make predictive conclusions.
The problem is, for a lot of us, we can’t translate this process into wins.
The Limitations of Empiricism
Jon Finkel famously said, “Focus only on what matters.” FOOWM.
This is not useful to most of us. He might be able to use this mantra to turn it into wins that many of us cannot. But that doesn’t mean that I can or you can. What does matter? I’m no Jon Finkel. I can’t simply make the leaps that he can in nearly as many situations. For Finkel, innate talent has to be a huge portion of what he is using when thinking about what matters. While it is possible that he is also drawing upon some reserve of theory, it isn’t useful to you or me simply because we don’t have access to it.
Wafo-Tapa’s empirical data can be translated a lot more readily; we could simply play any deck and sideboard combination that he told us to, respecting his copious data. But, that is still largely an exercise in imagination for most of us. Most of us won’t have access to his data, and even if we did, it is still possible that his data is wrong, because something about his testing for it might not have been Good.
One player that I often think of with the limitations of empiricism is a player who employs it more than most that I know: Mike Hron. Hron is a kind of robot for Magic. At the start of any format I’ve seen him play, he tends to be pretty average. Collating, collating. But, eventually, he gets to a point where he’s explored a format so much that he just knows it. You can look to formats like that in Geneva and Kuala Lumpur for the evidence of it.
I’m still reminded of the bad points of it. I don’t know which succession of players must have garnered massive game swings on Hron during Invasion Block draft with Exclude. Well, it must have made an impression. Hron would famously not cast creatures into a potential Exclude (U2) unless he was trying to bait it out, was planning on double-threating, or had some otherwise decided it was okay. (You can still see this kind of play in Hron — note the finals of PT Geneva.) Unless he had to, he would let the board state sit, casting the occasional Instant or Sorcery rather than let an Exclude hit. And this would work out for him often enough, not hurting him too often, even though it still commonly would.
But, perhaps with a more reasoned understanding of the interactions of the tempo loss for both players in waiting on the Exclude/playing around it, he might have been able to know when to give up the ghost more often. I know I definitely saw Hron lose games playing around an Exclude that wasn’t there, and more importantly, shouldn’t have even been played around in the first place, because the cost to tempo simply wasn’t worth it. Hron just didn’t really have the theoretical background to help him make the call, so he banked on his empirical data, and it didn’t serve him well.
This is where the value of theory begins to have more and more merit.
On Theory
Theory in the game of Magic attempts to explain the way that the world works, so that we can deduce correct play without having to go through an infinite number of trials. (“Yep, chump blocking is bad again. Let’s try it another time and see if it still is bad.”) Where empirical knowledge comes from induction, theory proposes a view of the world where there are rules in place that will allow us to use deduction to come to the right decision.
At least that’s how it works in theory (har, har).
As in science, theories gain more and more acceptance the longer that they are around and withstand scrutiny. If someone says that something is “just” a theory, I’d like to introduce them to the Theory of Gravity. However, we non-scientist types tend to mess up what we mean by theory. People call their conjectures, their hypotheses, their speculation, and their opinions “theories.” Magic players are as guilty of this as anyone else.
Most Magic theory is much more akin to hypothesis — a reasoned proposal to describe why things happen — than theory. And that’s fine. We can still call it theory if we want. But this does mean that if we want to be winning in the game of Magic, we should make sure that our theories about how the game works are useful.
Disagreement about the utility (or even the veracity) of something is a key to Magic theory. Getting better and better models for how the game works can do nothing but help us improve. Where empirical knowledge has its limits in only being useful to what has already been experienced, theory, if correct, can help us understand new situations far more quickly. When Richard Feldman disagreed with Mike Flores recent article, it isn’t a pissing contest, as some have claimed, but rather a disagreement about how the game works. Whether one side or another “wins” the disagreement is less important than the recognition that a poor theory can actual lead people to Bad magic. When theory succeeds, people should be able to use it to help play Good magic, right?
A good example of really, really bad Magic theory was Geordie Tait well-intentioned Virtual Card Advantage Theory. In total, Geordie’s three-article series set off twelve pages of forum conversation, and really left people angry. Geordie has attempted to simplify the concept of Card Advantage, and in so doing, helped new players better understand the concept, while simultaneously putting forth a theory that wasn’t just less accurate, but was less useful.
For theory, this is very bad indeed.
Richard Feldman feelings about theory mirror my own: well-wrought theory should make people better at the game, and should be successfully applicable whenever the game comes under its sphere of influence. In other words, it should be a useful tool at whatever task it is designed for. Furthermore, a theory should never hurt you in the long run. It should not give you a somewhat useful tool, but plateau your skill at some level of serviceable mediocrity. This is not easy
It’s been a while since Eric “edt” Taylor has crashed on my couch before a Magic tournament, but he once said something that I’ve taken to heart. “Inventing a term,” he said, “is one of the very worst things you can do in Magic writing. A term should only get invented because it has to be.”
Magic has very few concepts that approach theory in a way that requires new terms. Some of these are fairly canonized. Most people who think about Magic would agree that tempo and card advantage are both theories that have use (though getting people to agree on a theory of tempo is hard indeed). Beyond that, the jury is out.
Theory is no substitution for empirical data. But theory can do something that empirical data could never do for you: it can give you a guideline in an unknown situation or in a situation where you just don’t have enough experience to really rely on the empirical.
The Limitations of Theory
One of the problems that I see in beginning “theorists” is the way that they conceive of how the theories work. They seem to think that there is “a” theory for card advantage and “a” theory for tempo, and that you use one at one point, and another at another point. Some people seem to think that theories of card advantage don’t apply when they are playing a particular deck versus another, say RDW versus TEPS. “Oh,” they say, “I don’t use that theory now.” Perhaps, this is where Finkel’s Maxim: FOOWM can come in and remind us otherwise.
The rules of Magic that we try to describe with theory isn’t something that we turn on or off. They are there. When Finkel says FOOWM, what we should remember is that the various theories that inform our games might matter more or less in a particular moment, but they are still there. Conceiving of them as things that aren’t interconnected is a bad call. Beginning theorists will say things like “I am using the theory of card advantage because I am a control deck” or “I am using the theory of tempo because I am a beatdown deck,” when in reality those elements to the game are always with us. If we are to win, we need to determine how much a particular element matters at any particular time.
This reminds me of a classic moment between Brian Kibler and Neil Reeves. Reeves told me about a draft at a larger event where he beat Kibler’s vastly superior Blue/White control deck with a poor Red beatdown deck with various cards using the so-called “Punisher” mechanic (opponent may pay life/take damage to stop the effect). Reeves’s deck had at least 6 Blazing Salvo, while Kibler’s deck was jam-packed with all the goodness a Blue/White deck could muster. But Kibler didn’t focus on what mattered.
“You know Kibler,” Reeves said. “He just ‘knew’ that he wanted to get card advantage, so every time I cast a Blazing Salvo, he would pay the life. I 2-0’ed him. He was so mad that he wanted to play again. I can’t even remember how many games I beat him. Finally, I just said to him, ‘Kibler, let the Salvo resolve.’ I couldn’t win a game after that.”
Kibler put too much weight on one aspect of the physics of Magic, more than all of the massively intertwining rules of Magic. In this sense, he failed to FOOWM, choosing instead to believe that a theory — a theory he knew to be true — was what mattered more than anything. Once he corrected this, his superior deck could win.
Understanding a theory can help us win, but we can’t let any particular theory somehow make all of our decisions for us. We’re not playing Checkers. We’re playing Magic, and Magic doesn’t care if we have a theory or not. The innate way the game works in a particular moment in time will still be the innate way the game works at that particular moment in time.
On Good Magic
First of all, what is Good Magic? What is there about it that we can learn, that isn’t simply innate?
It’s not just succeeding. Winning can happen from happenstance. Winning can happen from pure natural aptitude. But, if we’re not lucky or (naturally) good, how can we get the game wins we want, on our merits, if not through Good Magic?
Good Magic is this: play (or thought or preparation) that increases the chances of our victory. Good Magic has nothing to do with guaranteeing victory — we call that cheating. Good Magic has everything to do with making that chance as big as it possibly can be.
The complexities of Magic are vast. It might very well be that Magic is the most complex popular strategic game on the planet. Maybe it isn’t, but I wouldn’t be surprised if it is. As such, many of our estimates about the importance of certain aspects of the game are only best guesses. Even if the formula for success were something like w*CA + x*T + y*PS + z*L (Card Advantage, Tempo, Play Skill, and Luck, respectively), we can only guess what the values of those constants, w, x, y, and z are.
Thus, we need to have both induction and deduction, in some combination, if we hope to increase our chances of victory without the kind of colossal Magical skill that a Finkel has. And even if we had that kind of skill, let me tell you, I bet that Finkel still practices from time-to-time, so you should too.
Theory is, for some, the harder of the two beasts to understand. And more poignantly, the more contested. In the contesting of theory, we can get a stronger picture of the game. I know, for example, that disagreeing about card advantage with Chapin has only furthered my theoretical understanding of the concept, as I defended my position.
Mixing the two is important. Evidence from the real world can butt up against one’s theoretical assumptions of how it works. Maybe card advantage’s hypothetical importance isn’t worth what you think it is, and a little bit of empirical evidence will bear that out. Similarly, just because you’ve never seen a card work, empirically, doesn’t mean that it can’t be made to work, if the card is theoretically sound. Coming at the issue from both sides can be the best way to achieve Good decks or play Good Magic.
I know I come at Magic largely from a theoretical perspective, because I’m good at it and I like it. I come at it from this perspective more and more, because I lack the time to put in all of the work that I might otherwise like to. Other people are in the exact opposite boat. Thankfully, if you get people together that work the other way, there are a lot of ways that you can share knowledge to your mutual benefit.
There are elements of my Magic that are definitely Good Magic. There are times when I play when I am definitely playing Bad Magic. Honing our skills with empirical knowledge and a solid understanding of theory are the way to move more deeply into Good territory. If becoming Good is important to us, than Good testing and Good theory should be important to us as well, even if we aren’t the ones testing or theorizing.
See you next week!