Searching For the Best Deck: Do We Even Know What We're Talking About?
Whenever a format is shaken up, or a major tournament happens, there are always people ready to talk about the best deck. One faction will think it's Deck A, another will think it's Deck B, and one group of loonies will even decide (can you imagine?) that it's Deck C. But what does it mean, to be the best deck? It seems like something which should be obvious, but I have recently discovered that while everybody knows what the best deck is, very few people know what the best deck is. I have decided, therefore, to put my meager skills towards laying down a definition of the best deck and discussing its implications.
Building the Definition
So what is the best deck? What does it mean for a deck to be "best"? Well obviously it has to be best in relation to something, otherwise saying a deck is "best" doesn't mean anything. A selection of other decks, perhaps. But what decks? How about a metagame? A metagame, of course, is a collection of decks you expect to see at a tournament, and since our ultimate interest is in winning tournaments, that seems like a good thing to define the "best" deck with relation to.
So the best deck is "best' with relation to a metagame, a metagame being a specific set of decks. You aren't just playing against a deck though, you're also playing against people, and people aren't all at the same skill level. As a result, many of the decks in the tournament may be improperly piloted, which could influence how a deck would want to work in that matchup. This means that defining the best deck just with relation to other decks isn't enough. We also have to define it with relation to the people behind those decks.
Okay. The best deck is "best" with relation to a specific set of decks, or metagame, and the people piloting each of those decks, with their varying levels of skill. Where does this leave us?
Well unfortunately we still don't know what "best" means, but at least we have a framework now in which to consider the term.
As I said before, our goal is ultimately to win tournaments. We would therefore like to have the "best" deck be one that stands a good chance of accomplishing this. In fact, we would like the "best" deck to be the deck that has, well, the best chance of winning a tournament.
And there you have it. The best deck is the deck that has the best chance of winning a tournament consisting of a known set of decks and a known set of players piloting those decks.
Why only the best chance? Why can't we define the best deck as the one which will win that tournament? The reason is a simple matter of statistics. The fact is that Magic is not a deterministic game. If two people shuffle up the same decks and play a game two different times, the game will proceed in a different fashion each time and the outcome will be different in some way, potentially in which deck won. As a result, it is impossible to guarantee results for a deck; one can only talk about a deck's potential for winning in terms of probabilities. Thus we cannot guarantee that the best deck will win, only that it has the highest probability of doing so. So now we're done, aren't we?
"Wait!" you say, "What about the player! Surely you cannot ignore that the person piloting the best deck could affect what deck it is!" That is a good point. In reality, you can't do this. In reality, if you want to try and determine the best deck for you to take to a tournament, you have to consider your own limits as a player.
Why, then, do we not include the person playing the best deck in the best deck's definition? Again, the reason is simple. We are trying to define the best deck, not the best deck for a person. Obviously, despite all Flores' contentions that Mono-U is the best Standard deck, if I give it to a six year-old who's been playing Magic for a week, he's going to have a much poorer showing than he would with the relatively simple White Weenie. If we allow the best deck to be different depending on the person playing it, then we could never come to a consensus as to what the best deck actually is. Obviously we would like to have some idea as to an objective best deck when discussing a format, so we need to be able to come to such a consensus.
We must therefore abstract out of the definition of the best deck all factors that are not inherent to the deck itself, and this means that we must ignore the playskill of its pilot. Instead, for purposes of the definition, we assume that the pilot of the best deck is some theoretical player of sufficient skill to properly play the best deck. Doing so, however, opens up the possibility that taking the best deck to a real-world tournament will not guarantee you a win, or even the highest probability of one, as you may not be competent enough to play it correctly.
The Definition of the Best Deck
So let's put it all together:
Given a tournament consisting of a known set of decks and their pilots, and given a pilot of sufficient skill to properly play the best deck, the best deck is the deck which has the highest probability of winning that tournament.
The Implications of the Best Deck
There are a number of things which follow from the definition of the best deck. The most obvious is, as I have mentioned, that the best deck will not necessarily win the tournament. This is a result of the non-deterministic aspects of Magic, and is no surprise. It means, however, that the deck which wins the tournament is conversely not necessarily the best deck. This is why we don't merely look to tournament-winning decks for guidance, but also to the decks that were in the Top 8.
It also turns out that the best deck for a real-world tournament may not even be there. There is a huge (infinite, in fact) variance in the potential decks one could use. There is no guarantee that for a given tournament in real life, any one of the decks could not be improved with respect to the rest by the changing of some cards, thus creating a deck which had a better chance of winning the tournament. This is an important distinction between theory and the real-world. This fact is what lets us say things like "In hindsight, I would have been better off with cards X, Y, and Z than cards A, B, and C." Although you obviously can't go back in time and change the tournament, you can still create a "best deck" for it that wasn't actually there.
Metagaming and the Best Deck
At this point you're probably wondering how we can use this definition to help us find the best deck. I have some bad news for you. You can't. At least, not directly. As I already implied, the best deck could be any one of an infinite number of permutations of cards. Furthermore, in order to find the best deck for a tournament, you would have to have prior knowledge of the decks being played in that tournament and the exact skillset of the people playing them. You would then have to spend a long time testing large numbers of decks often varying by only a few cards in order to get those last little incremental increases in your probability of winning. It is because of these two things, the huge time commitment and the need to be able to see the future, that directly finding the best deck based on the definition is impossible in practice.
If that is the case, then of what use is this definition, indeed this very concept? If you have no hope of finding the best deck, why would anyone bother discussing it? Well it turns out that for any given metagame, a best deck is guaranteed to exist*, and knowing this allows us to do some things which are very valuable in preparing for a tournament or discussing a format in general.
You see, knowing that a best deck exists allows us to try and find it. Trying to find the best deck is actually a quite common thing. You already do it all the time. It's called metagaming. When you metagame, you are trying to guess as to the deck composition of an upcoming tournament, and choose a deck you think will work best there. You obviously can't be certain that you are correct, and so you approximate the best deck by making educated guesses about the tournament composition and cutting corners in your testing based on your past experience.
The goal of metagaming, therefore, is to come as close as possible to finding the best deck given the limited information at your disposal.
Talking About the Best Deck
So the concept is not entirely useless, but this does not explain why we constantly hear theorists talking about what the best deck in a format is, asserting that it is some archetype or another. What does it mean to talk about the best deck in a format? Furthermore, if the actual best deck cannot be found without both incredible time resources and prescience, how are they able to talk about it so casually? Are they superbeings, with powers beyond those of mere mortal men?
While some writers may in fact be brilliant, I assure you there is nothing supernatural in their apparent powers of deduction. It's merely a case of being lazy with a definition, and happens all the time in other areas as well (for instance, the common misconception of a theory as not requiring rigorous testing). It just so happens that in this instance we haven't until now consciously known the definition for what we were talking about. This laziness isn't necessarily a bad thing, as conversation becomes unnecessarily encumbered if we are to qualify everything we say in terms of a formal definition, but it is important to clear up what exactly we mean at some point.
When people talk about the best deck in a format, they actually have a specific metagame in mind. What they are thinking of as the format is a certain collection of decks with various archetypes in certain proportions. We can therefore characterize a format in the following way:
A format can be described as having a characteristic tournament in which the major archetypes are all present in a certain proportion.
For example, if we had a Ravnica Block format consisting of BDW, Dredge.dec, and Mill.dec, with BDW being twice as popular as either of the others, then you could think of the format as a tournament of 100 people, 50 of which are playing BDW, 25 of which are playing Dredge.dec, and 25 of which are playing Mill.dec. If you were then to describe a best deck with respect to this hypothetical format, it would be the deck that had the greatest probability of winning that 100-man tournament (You can see why enforcing proper use of the definition becomes cumbersome).
The "best deck" people refer to in this context is not, strictly speaking, a specific decklist either. Instead, people discuss a format's "best deck" in terms of an archetype. To pick on Flores again because it's an obvious example, at least up until Worlds he asserted that Mono-U was the best deck in Standard, despite the fact that the optimal list for that archetype changed over the course of a few months to the point where it wasn't even mono-colored anymore.
Thus the best deck for a format is the general archetype which, given a format, has the highest probability of winning that format's characteristic tournament.
It is in this way that people are using the concept of the best deck when discussing formats and metagames at large, whether they are doing so consciously or unconsciously.
In Conclusion...
I have not really discussed anything new in this article, nor anything particularly complicated. Certainly the idea of a "best deck" has long been in the general consciousness of the average Magic player. Nonetheless, if we want to make serious inroads into the vast field of Magic theory, it is necessary to have formal definitions for even the most basic of concepts, in order to both better understand how we use those concepts and to build further theory on top of them. I hope that in this article I have been able to clearly lay out both a definition for the best deck and a strong justification for it. I also hope I have showed how this concept is used by people in a way that is, if not exactly correct, at least in line with my definition.
-David Kleppinger
dr.klep@gmail.com
*It is important to note that the best deck may not be unique. It is conceivable, though improbable, that there could be two decks that have an equal probability of winning a tournament which is higher than the probability of any other deck winning a tournament. Thus, even if you were to have the amazingly unlikely metagame in which every conceivable deck had an equal chance of winning if piloted properly, you would be able to choose a best deck as being any one of those. Incidentally, that would be a really terrible format, and fortunately will almost certainly never exist in Magic.
| 
