The Epistemology of Deckbuilding

[TheAmericanSpirit]

I don't think we can really talk about optimizing so much And I think that is kinda what rubs me the wrong way about this whole idea of deck building templates (which is how this ultimately sniffs to me). What makes a deck "good" is a question that needs to be answered with "good at what?"

Yeah! There's the philosophical uncertainty I was looking for! What makes a deck "good"? Is statistical analysis the best perspective to reach the desired deck state. When a deck comes into its own, does the player determine the ethos of the deck, or does the deck determine its own ethos and we can only choose to accept or reject it? Let's get epistemological! From where does truth come?

For a while, this comment by tstorm823 has stuck in my brain like overcooked chicken sticks between your teeth. What is the epistemology of deckbuilding? Each question posed is certainly worthy of discussion, so I figured we should discuss them! So:

1. What makes a deck "good"?
Imho "goodness" is a quality 75% determined by expectations being met by the development and successful performance of a given deck. I like my decks to be able to consistently enact Plan A and be able to rebound after some roughousing. A "good" deck does both of these things. The other 25% is determined by the Rule of Cool. A "good" deck will let me play a larger number of cards I arbitrarily consider to be %$#% awesome than a mediocre or bad deck.
2. Is statistical analysis the best perspective to reach the desired deck state?
IMO it can be useful but play experience comes foremost in why and how I tune a "good" deck to be better.
3. When a deck comes into its own, does the player determine the ethos of the deck, or does the deck determine its own ethos and we can only choose to accept or reject it?
The third question is what truly fascinates me. I honestly think the deck may actually determine its own ethos. Certain decks and archetypes just operate certain ways within the context of magic, thus you can only either enjoy that aspect of the game or not. Is that not mechanical self-determination?!

Anyway, most of you are smarter than me. What do you folks think?

[Mookie]

What is 'good'? Going off the Oxford definition:

adjective
adjective: good; comparative adjective: better; superlative adjective: best

1.
to be desired or approved of.
"a good quality of life"
pleasing and welcome.
"she was pleased to hear good news about him"
expressing approval.
"the play had good reviews"
2.
having the qualities required for a particular role.
"the schools here are good"

As a result, I would say that a good deck is one that is capable of filling a particular role. In its most common usage, I would assume the role that most people aim to fill with their decks is 'a deck capable of winning an above average percentage of games'. As a result, decks that win often are generally good, while decks that win rarely are not good. However, that isn't the only type of deck. Maybe you want to build a group hug deck that aims to make everyone at the table laugh, or a theme deck that aims to emphasize a particular mechanic or theme. If I were to make an 'exalted deck' that only had Rafiq of the Many as its commander and no other exalted cards, I wouldn't consider it a 'good exalted deck' even if it had a high win percentage - it has failed at filling the role it was given.

I'll also note that it's possible for decks to be designed to fill multiple roles. Maybe you want to build a deck that both wins a lot and plays a bunch of exalted cards in it. The more roles you want to fill, the more difficult it will be to fill all of them. At that point in time, I think you need to reevaluate your priorities and consider whether you're willing to compromise one of your deck's roles in favor of another.

...for my own decks, I would say I generally want them to 'win a reasonable percentage of games', 'have a low number of non-games', 'emphasize a particular theme, mechanic, or emotion', 'fun to play' and 'be relatively affordable'. Winning is one of the goals, but it isn't the only one unless you're playing cEDH.

---

As for statistical analysis vs experimentation... I think both are useful tools. There are a lot of variables that analysis can't help with, particularly with respect to how it interacts with other decks. No amount of analysis will prepare a graveyard deck for a meta where everyone is packing Rest in Peace. On the flip side, if you rely entirely on experimentation, then you'll be prone to overfitting your meta (7 pieces of enchantment removal in a mono-black reanimator deck is probably a bad idea, even if everyone in your meta is playing RIP).

---

Point 3 gets into authorial intent // death of the author territory and is more of a philosophical question. That said, the fact that decks can be altered suggests 'accept/reject' isn't the only option. If you have expectations for a deck and it doesn't meet those expectations, you can alter the deck to perform differently... or you can alter your expectations / goals. These aren't mutually exclusive.

[tstorm823]

2. Is statistical analysis the best perspective to reach the desired deck state?
IMO it can be useful but play experience comes foremost in why and how I tune a "good" deck to be better.

Anyway, most of you are smarter than me. What do you folks think?

Your answer to question number two frames up a classic debate of epistemology: rationalism vs empiricism. Epistemology is the branch of philosophy that considers questions of knowledge itself, and within that is the question of where knowledge actually comes from. Rationalism is the theory that human reason is the chief source of knowledge, that we can logically deduce facts using our own mental powers. Empiricism is the theory that experience is the chief source of knowledge. Though some people historically have chosen one of these theories to hold exclusively, I think the majority of people would say both reason and experience play a role toward validating information. That's basically how the scientific method plays out: a logical hypothesis is formulated by reason and then tested through experimentation.

Your answer to reaching a desired deck state is pretty much that: reasoned analysis is useful, but experience is ultimately how you validate the quality of a deck. An empirical look at the results having tested the deck repeatedly is undoubtedly worth more than what logic may dictate should happen. But another question occurs to me: is that what we want? Brewing an untested deck or competing in an unsolved meta are some of the most exciting things in the game, and maybe those things are so exciting because we want to beat that answer. We want to find the truth through pure reason, we want our minds to be the source of truth before even having that direct experience. A deckbuilding game like this allows us the opportunity to compose something using our own logic and reason and then find out after through experience whether its nature really is what we intended it to be.

[Treamayne]

1. What makes a deck "good"?
Imho "goodness" is a quality 75% determined by expectations being met by the development and successful performance of a given deck. I like my decks to be able to consistently enact Plan A and be able to rebound after some roughousing. A "good" deck does both of these things. The other 25% is determined by the Rule of Cool. A "good" deck will let me play a larger number of cards I arbitrarily consider to be %$#% awesome than a mediocre or bad deck.
2. Is statistical analysis the best perspective to reach the desired deck state?
IMO it can be useful but play experience comes foremost in why and how I tune a "good" deck to be better.
3. When a deck comes into its own, does the player determine the ethos of the deck, or does the deck determine its own ethos and we can only choose to accept or reject it?
The third question is what truly fascinates me. I honestly think the deck may actually determine its own ethos. Certain decks and archetypes just operate certain ways within the context of magic, thus you can only either enjoy that aspect of the game or not. Is that not mechanical self-determination?!

1. What makes a deck "good"?
A deck is "good" if it is both fun to pilot and fun to play against.
2. Is statistical analysis the best perspective to reach the desired deck state?
I think statistical analysis can be useful as a starting point, or "troubleshooting" a deck; but playing is required to iterate and fine-tune
3. When a deck comes into its own, does the player determine the ethos of the deck, or does the deck determine its own ethos and we can only choose to accept or reject it?
I think a player's ethos determines the types of decks that player will build/modify. Sure, a specific deck's version of that will grow and evolve as it is played and iterated (and possibly rejected when something that was expected to be "fun" turns out not to be the case); but the overall ethos has to be a symbiosis of both what the deck "does" and how the player pilots it to "do that."

YMMV

[materpillar]

1. What makes a deck "good"?

2. Is statistical analysis the best perspective to reach the desired deck state?

3. When a deck comes into its own, does the player determine the ethos of the deck, or does the deck determine its own ethos and we can only choose to accept or reject it?

1. The spikey definition is that the deck wins a statistically relevant % higher than average. I consider my decks to good when it has a statistically solid mana base (or very very good reason for not that), a clear goal and clear plan to reach that goal. The goal doesn't necessarily have to be good and the plan may be a huge stretch. My Changeling deck is built around saying as many creature types as possible while casting a Changeling and being as cutthroat a deck as possible. My Tivadar deck is built around using Tivadar killing goblins. The decks are functional with clear goals and card choices to work towards those goals.

2. Kinda? All of our personal statistical sample sizes are minuscule. Using stats for results based analysis is pretty poor for analyzing how well the deck is built based on its winrate. Statistical analysis is great for figuring out mana-base ratios and similar.

3. I'm not George RR Martin. I maintain control of my builds.

[Legend]

What makes a deck "good"?

"There is none good. No, not one." - Jesus

Clearly, He was talking about Commander decks. 😅

[Jemolk]

Well, for my own two cents...

1. What makes a deck "good"?

2. Is statistical analysis the best perspective to reach the desired deck state?

3. When a deck comes into its own, does the player determine the ethos of the deck, or does the deck determine its own ethos and we can only choose to accept or reject it?

1. When I think what my deck does is cool, and my opponents look at what I'm doing and think, "hey, that's cool," that makes a deck good IMHO. This condition is, in my view, both necessary and sufficient -- meaning, a deck cannot be good without meeting it, and a deck that does meet it is guaranteed to be good. This, naturally, relies on some level of agreement between you and your opponents, and everyone's state of mind. Thus, the goodness of EDH decks is entirely subjective.

2. Nope! Largely irrelevant to what I think makes a deck good, in fact! Sure, the deck has to work to at least minimum specs, but that doesn't require that degree of detailed analysis at all, and in fact, relying on such analysis would probably get in the way of the creativity that I think is essential to good commander decks.

3. A little of both, really. As you play, you learn about what a deck needs/wants, and this guides the changes you make. You still make the card selections, though, and we can also decide what the destination point is that we will allow our deck to guide us to more closely approximating. I have quite a few decks that feel like cohesive wholes because of that iterative design process where I "listen" to the deck, but I also decided what I was iterating towards very consciously before starting, and that has also been essential to liking the end result.

[Avacyn Believer]

1. I suppose there is a scientific answer along the lines of 'good deck functions, doesn't stall often (if at all), and wins above average'. But I agree with the sentiment that 'good deck' depends on what your intentions are. Personally, what makes a deck good for me, rather than winning, is engaging with the game. I don't care if I win as long as I felt like I had a chance at winning, and had fun playing the deck. I don't remember wins, but I do remember cool interactions or funny moments. To me Commander is like a drinking game, you don't play to win, you play to have fun.

But I will admit that this is entirely subjective and some people play to win, or have other criteria that make a deck good to the.

2. Others said it already, it is useful but not the only tool for tuning decks. Recently I've been away from home for couple of months and couldn't play Magic, so I spent the time fine tuning my pet deck and gold-fishing it a lot. Which was useful to give me ideas on lines of play... but it couldn't prepare me for actual games with politics and people doing random stuff to disrupt my plays.

3. I like the idea that the deck tells you what it needs, but I think you can just ignore it and force the deck into whatever you want it to be. It might not work as well, but you can do it. Self-imposed limits I'd say fit into 'ignoring the deck'. I also think that by experimenting with breaking the 'context of Magic' you can discover some unique ways to play.