Dominion Strategy Forum

Dominion => Dominion General Discussion => Topic started by: Kirian on October 11, 2016, 02:00:50 pm

Title: Meta?
Post by: Kirian on October 11, 2016, 02:00:50 pm
Is there a meta in online Dominion?  Can there be a meta?  If there is one, is it fast for some reason?  Am I dumb for asking these questions, or for other gameplay reasons?

I've played a number of games lately where my opponent started grabbing single Provinces early--theoretically a bad strategy unless there's no +Buy around and an engine is available.  Yet I find myself losing these games; even though I can build to a three-province turn, when the opponent has five Provinces, that becomes a PDDD turn, and if the opponent hits another Province, the next turn is DDDD, and hope the opponent can't spike a Province.  As it happens, 6P beats 2P-7D.

Any thoughts?  When do you cut your losses on your engine and take single Province-single engine piece?
Title: Re: Meta?
Post by: Chris is me on October 11, 2016, 02:25:13 pm
Dominion absolutely has a metagame. It manifests itself in how your opponent plays and how you predict they will play, and which strategies are dominant.

A lot of various strategies are sort of "rock paper scissors" with each other. Some more blatantly than others, but usually to some extent. Lots of engine players tune themselves to win the mirror, or to build to a certain fixed point, when they should be more fluid and reactive. That fixed point they build to is a function of both the metagame and the strength of a traditional / average board.

In the scenario you're laying out, you definitely do let the time and rate of your opponent's greening override your original plan, being careful not to let Points Panic get to you. If they started greening, greening strictly reactively as soon as they start isn't always the ideal though. You're already behind - are you close enough to a milestone point that you can build and catch up? Maybe it's going Prov Duchy instead of double Prov - sooner than you planned but soon enough to overtake a single Prov player with a head start. If you cut your losses and just go single Prov, your only hope is that they trip and you don't.
Title: Re: Meta?
Post by: GendoIkari on October 11, 2016, 02:52:45 pm
What do you mean by a meta? I know what a meta game means as it relates to board games, but I'm not clear in how it relates to your question.
Title: Re: Meta?
Post by: werothegreat on October 11, 2016, 02:56:21 pm
What do you mean by a meta? I know what a meta game means as it relates to board games, but I'm not clear in how it relates to your question.

More in reference to CCG's.  Like with Hearthstone, you might decide to include a weapon-destroying card if most people are playing decks with weapons in them, or you might include board clears if most people are playing decks with lots of minions.  With regards to Dominion, it could be something like, what strategy do you go for, given that it's more like your opponent will go for an engine?  Or maybe more people are going for big money.  The common strategies of other players outside of your current game informs your choices.
Title: Re: Meta?
Post by: dedicateddan on October 11, 2016, 04:26:10 pm
In dominion, players start the game with the same 10 cards, so there's no "metagame" of starting decks

What does vary, however, is the style in which the game is played. Some players favor single-province play and others favor huge engines

What's correct on a given board depends on the specific cards and interactions on the board. Engine play gets stronger when there's a large pool of alt-VP to pull from. It gets weaker if 4-5 provinces make up the bulk of the VP

In the end, it takes a lot of attention to get these little things right
Title: Re: Meta?
Post by: Chris is me on October 11, 2016, 04:30:02 pm
The metagame is obviously not present in constructed decks, but in strategies chosen and the expected decisions you assume opponents will make, as well as the expectations of speed and performance you subconsciously set when strategizing. One obvious example of metagame driven strategy benchmarks is the evolution of BM strategies as primitve yardsticks - Gear can win in 14 turns, so you have to beat that, etc.

I think what's "correct"on a given board, in addition to what dan has outlined, is heavily dependent on your opponent'a strategy in all cases. You need to win before your opponent does, and there are tradeoffs in reliability, payload, explosiveness, three pile risk versus speed. The correct balance of these tradeoffs is opponent dependent, and before they have acted and you have profiled their strategy, your presumptions of their strategy constitute the metagame.
Title: Re: Meta?
Post by: trivialknot on October 11, 2016, 04:39:00 pm
In an optimal strategy, you don't need to make any presumptions of your opponent's strategy.  You watch what they do, and react to that.  If you react to what you predict your opponents will do, then they can just not do the thing.

However, people don't play optimally, therefore there is meta.  The fact that there is meta is proof that people aren't playing optimally.

The most obvious kind of meta is when newbies think Thief is really devastating, so the best strategy is to prepare for some Thieves.  However, I suspect that the metagame online is more subtle, since the players are better?  I dunno, I play offline only.
Title: Re: Meta?
Post by: SCSN on October 11, 2016, 05:01:20 pm
The most obvious kind of meta is when newbies think Thief is really devastating, so the best strategy is to prepare for some Thieves.  However, I suspect that the metagame online is more subtle, since the players are better?

I generally still prepare for my opponents going mass Thief.
Title: Re: Meta?
Post by: timchen on October 11, 2016, 05:16:10 pm
Meta originates from hidden information in a 2p game. Since there is no hidden information in dominion (opponent cannot build deck without being seen) in principle there should be no meta.

In reality when the play is far from optimality or when the strategy landscape is very frustrated then there can be meta; i.e., what kind of strategy people tend to play. In theory they are just non optimal strategies but the difference is just too small to matter or it is just very hard to play in the optimal way.

StarCraft is an ideal example that has both. For one you can't know for sure what your opponent is doing so your strategy really depends on the expectation of what he is likely doing. On the other hand a part of the reason why meta exists, especially when the game is young, is just because that the game has not been figured out and it takes a lot of practice to play a certain strategy good enough to know whether it is really good enough.

In the case of dominion, there certainly are strategies that beat one another, but since it is turn based, the optimal strategy exists between these strategies. The perceived meta can only be from the population adhering to playing in some non-optimal way.
Title: Re: Meta?
Post by: Chris is me on October 11, 2016, 09:37:44 pm
There's PLENTY of hidden information in Dominon, that's ridiculous. You have no certainty over what cards the opponents will buy in the future, even if you know all their options, and you can use knowledge of the metagame as one of many factors in predicting their likely strategy. That alone is sufficient evidence of metagame influences in Dominion.
Title: Re: Meta?
Post by: DG on October 11, 2016, 10:48:44 pm
In an optimal strategy, you don't need to make any presumptions of your opponent's strategy.  You watch what they do, and react to that.  If you react to what you predict your opponents will do, then they can just not do the thing.

This is of course only true in 2 player games. In multiplayer games you do have to make assumptions about how your opponents will play. It is an advantage to know the style of play that your opponents will employ.
Title: Re: Meta?
Post by: tristan on October 12, 2016, 06:23:01 am
Is there a meta in online Dominion?
Nope. Meta is usually referrs to some out of game thing going on, i.e. deck-building in CCGs/LCGs.
Players choosing different strategies and greening at different moments DURING the game, well, that's just the game and hardly a metagame.
Title: Re: Meta?
Post by: Chris is me on October 12, 2016, 07:17:39 am
Is there a meta in online Dominion?
Nope. Meta is usually referrs to some out of game thing going on, i.e. deck-building in CCGs/LCGs.
Players choosing different strategies and greening at different moments DURING the game, well, that's just the game and hardly a metagame.

But think of it this way. Would players be choosing Jack BM strategies or Rebuild strategies or Hermit / MS strategies, and all the detailed choices that come with them, without the outside influence of online discussion boards or learning from the examples of other players on the ladder? Certainly not. And it's unlikely that this play is perfectly, universally optimal, so the way these strategies manifest are themselves products of the Dominion metagame.
Title: Re: Meta?
Post by: tristan on October 12, 2016, 07:35:02 am
Is there a meta in online Dominion?
Nope. Meta is usually referrs to some out of game thing going on, i.e. deck-building in CCGs/LCGs.
Players choosing different strategies and greening at different moments DURING the game, well, that's just the game and hardly a metagame.

But think of it this way. Would players be choosing Jack BM strategies or Rebuild strategies or Hermit / MS strategies, and all the detailed choices that come with them, without the outside influence of online discussion boards or learning from the examples of other players on the ladder? Certainly not. And it's unlikely that this play is perfectly, universally optimal, so the way these strategies manifest are themselves products of the Dominion metagame.
Sure, if you consider everything that happens outside of a game as metagame that's true. But then every game has a metagame and the concept becomes meaningless.
I view metagame as a strategic notion, i.e. considering e.g. what cards my opponents might put into their decks while preparing my own deck in an LCG. Chess is also an obvious example although here folks simply say opening preparation instead of metagame.
Title: Re: Meta?
Post by: Kirian on October 12, 2016, 01:19:37 pm
So here's an example of what I was talking about:

http://forum.dominionstrategy.com/index.php?topic=16382.new#new

Opponent buys a Province on T6.  And wins.
Title: Re: Meta?
Post by: gloures on October 12, 2016, 02:32:41 pm
i don't think two player dominion has much of a metagame. What I see frequently happening is that there are two competing strategies and one is stronger than the second only if the other player doesn't  mirror, so in the end both players end up somewhat suboptimally. Multiplayer dominion on the other hand should have a lot of meta. If three players go for the same engine, there might not be enough components for all of them, and a fourth player going BM might end up with a clear road to victory, in another case there might have a big engine available, but if all three other players go for something like workshop/gardens, your engine will never have a chance to get running. So being first player in a game like this (specially since a strategy like workshop/gardens generally requires commitment from turn 1) you really have too ask yourself what do you think the other players will do.

You seem too be talking about when to green, but that's not meta, that's evaluating strategy. You gotta evaluate if the optimal strategy in a given board is a late blooming engine that can overcome an strategy that greens early, or if the early greener can potentially build enough of an early lead that it becomes insurmountable. Of course since dominion has a decent  luck factor still you might always pick the best strategy and still lose, so evaluating when luck was a big factor is also an important thing when trying to improve your game.
Title: Re: Meta?
Post by: timchen on October 12, 2016, 04:11:08 pm
There's PLENTY of hidden information in Dominon, that's ridiculous. You have no certainty over what cards the opponents will buy in the future, even if you know all their options, and you can use knowledge of the metagame as one of many factors in predicting their likely strategy. That alone is sufficient evidence of metagame influences in Dominion.

Future information does not count as hidden information. Think about chess. Or any board game. If you know for certain what your opponent will do then there is not much of a game right?

You seem to confuse between concepts. Meta is called meta because it is something that affects the game that is outside the game. Think about rock-paper-scissors. If you just play one game, you will do as good as you can by always play rock. But if your opponent knows this then you will lose all the time.

This is what meta is about: you have several mutual exclusive strategy options which counters each other (while nash equilibrium would be a probability mixture of them). How you choose one among is really by preference, and is the hidden information I am talking about. The population preference is the current metagame.

Not so much in Dominion. The population can certainly have a preference on different strategies, but it is pure suboptimality. For a given board, one can always react to the way the other player is playing. The optimal strategy is a script of actions taken when seeing what opponent plays, that will give you the highest win percentage. This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

 
Title: Re: Meta?
Post by: Kirian on October 12, 2016, 04:43:45 pm
Not so much in Dominion. The population can certainly have a preference on different strategies, but it is pure suboptimality. For a given board, one can always react to the way the other player is playing. The optimal strategy as a script of actions taken when seeing what opponent plays, that will give you the highest win percentage. This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

OK, I like this... definition?
Title: Re: Meta?
Post by: Chris is me on October 12, 2016, 04:48:10 pm
Not so much in Dominion. The population can certainly have a preference on different strategies, but it is pure suboptimality. For a given board, one can always react to the way the other player is playing. The optimal strategy as a script of actions taken when seeing what opponent plays, that will give you the highest win percentage. This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

OK, I like this... definition?

I just don't agree. Strategic choices become better or worse based on predictions of opponent choices, not purely reactions to ones that have already been made, and there are a nontrivial amount of boards with dueling / countering strategies and choices where popular and social influence will have an effect on the game. I do not believe every board has a grand, purely optimal strategy, and as long as humans are playing the game it will never be reached anyway.
Title: Re: Meta?
Post by: AdrianHealey on October 12, 2016, 04:55:02 pm
Given my understanding of the word 'meta', there is no, nor can there be, a meta in dominion.

One easy way to grasp it, is: meta depends on things outside the particular instance of the game you are playing. There is no such relevant thing in dominion.
Title: Re: Meta?
Post by: timchen on October 12, 2016, 04:58:02 pm
I guess it is a definition, but it truly is (at least in principle) what the optimal strategy is.

I don't dispute that in a 2p game that is complicated enough such as Go, even though in principle an optimal strategy exists, there is still playing style shifting around as time goes forward. That is sometimes also called meta too, but it is understood as from our inability to play this game optimally too.

I just realized that in LoL we also called a similar thing meta. The current meta is induced by the game balance changes but in principle an optimal b/p strategy should also exist. The situation is blurred since there is this factor that different players can play a given hero to very different effectiveness as well. But in principle I guess I would say again, meta should not exist if we understand the game perfectly. Or maybe it is just called meta because we consider the b/p as a metagame perhaps.

For dominion I just don't think the game is complicated to the point that one should consider how his opponent would like to play more than trying to play optimally. To me that is the real implication whether there is meta or not in a game.
Title: Re: Meta?
Post by: timchen on October 12, 2016, 05:13:37 pm
Not so much in Dominion. The population can certainly have a preference on different strategies, but it is pure suboptimality. For a given board, one can always react to the way the other player is playing. The optimal strategy as a script of actions taken when seeing what opponent plays, that will give you the highest win percentage. This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

OK, I like this... definition?

I just don't agree. Strategic choices become better or worse based on predictions of opponent choices, not purely reactions to ones that have already been made, and there are a nontrivial amount of boards with dueling / countering strategies and choices where popular and social influence will have an effect on the game. I do not believe every board has a grand, purely optimal strategy, and as long as humans are playing the game it will never be reached anyway.

Let's think about an example.

Suppose for a given board you perceive a popular strategy that your opponent is likely to play.

Now, is it a dominant strategy on this board that has no counter?
If so, then it is at least close to the optimal strategy so there is no preference to begin with.

If it has a counter, and the counter is a dominant strategy that has no counter, then it is pure suboptimality.

If it has a counter, but the counter has some other counter as well - then one should not blindly commit to play the counter. There is no necessity either. In the game one can see what the opponent is doing and go for the counter if necessary. This is clearly a better strategy than blindly going for the counter. Anticipating your opponent's play is itself suboptimal (especially if there is a counter to the play you are going to commit; and if not it is close to the optimal play anyway and anticipation does not change the way you play.) there is no necessity to counter it before it is gradually played out.

So in all situations, no prior knowledge, unless you know something with certainty, will change the way you play. That is how I think there is no meta.

Title: Re: Meta?
Post by: JW on October 12, 2016, 05:35:11 pm
Consider the following simplified example: Suppose you are going first and are deciding on one of two options for your first buy: one that's building the engine in a way that's better against an engine mirror, and the other that's building the engine in a way that’s better against a money strategy. You know that the engine is better, but your opponent may not know the same, and your first buy won’t influence your opponent’s choice of engine/money. Then your decision will depend on, among other things, a) how much better one option is than the other against an engine/money opponent; and b) your subjective view of the probability that your opponent will play engine as opposed to money.

Whether element b) qualifies as a “metagame” I leave to the judgment of the reader.
Title: Re: Meta?
Post by: Chris is me on October 12, 2016, 06:33:53 pm
If it has a counter, but the counter has some other counter as well - then one should not blindly commit to play the counter. There is no necessity either. In the game one can see what the opponent is doing and go for the counter if necessary. This is clearly a better strategy than blindly going for the counter. Anticipating your opponent's play is itself suboptimal (especially if there is a counter to the play you are going to commit; and if not it is close to the optimal play anyway and anticipation does not change the way you play.) there is no necessity to counter it before it is gradually played out.

The implicit assumptions here are what I generally find fault in. You're assuming committing to the counter is particularly costly, and that the counter can be gone for with equal effectiveness later on versus early. If the counter is costly to your deck, that tilts things in the direction of a more conservative playstyle. If the counter can be easily gone for later with minimal opportunity cost to skipping it now, then yeah, why bother committing?

But oftentimes, there are cards that are viable enough to not be total trash, but that you will get further mileage of based on your evaluation of how likely they are to go for that strategy. Here's a dumb example off the top of my head. Fool's Gold is on the board, and it seems like viable payload. The speed at which you have to get Fool's Golds, and the amount you contest them, is probably one of the strategic facets most influenced by the Dominion metagame. If you both ignore it, you can build up and trash a bit and pick up a handful at once near the end as needed. If one of you rushes and the other mostly ignores it, the BM deck the Fool's Gold rusher can give has semi-serious potential to mess up some engines. If you both rush it, you split them evenly and essentially delay the start of the build phase around it.

Consider a Talisman open in that scenario. It's not a useless card; maybe the Buys are scarce on this board and there's lots of low cost engine components, but no great double Prov payload. Knowing how Fool's Gold rushes versus engine payload games play out in the current metagame, predicting based on attributes of the player, etc. will help you decide if you wanna open Talisman or something like Cutpurse instead. I think on smaller scales these kind of decisions happen all the time.
Title: Re: Meta?
Post by: timchen on October 12, 2016, 07:40:39 pm
Hmm, I am not sure why do you think there is implicit assumptions about the counter. Also not quite sure what costly means. For the purpose if the counter strategy does not deviate from the strategy you are going to play anyways then there is not much to discuss, isn't it? Also I am not assuming it can always be gone from later in the game. If not, then the optimal strategy is likely needed to have incorporated some key elements of the counter.

Good to talk about an explicit example though. Still I am not sure what is the point you are trying to get to. In this case, maybe you are saying money-uncontested FG is strong, comparable to engines without it? Engines hoping to get a few FGs in a turn later can only work when the opponent is not getting them initially but becomes a dominant strategy when both players are not going for it initially?

Sure, but how does these factor into your own strategy decision?

For me the first thing to consider is that whether I need to contend FG. I don't need to know whether the current "meta" strategy likes FG or not; I will assume my opponent plays well enough and will contend it if it
is strong. So if I end up with a conclusion that an engine cannot win without contending FG against a FG-money strategy, then I will incorporate this part into my strategy. Getting a talisman can be a good way to contend a few FGs later in the game and it has some other uses, so that can be something I open with.

I agree, situation can get pretty complicated, but overall it is just a pretty big decision tree consisting of what can the opponent go for and what options does one have. Thinking about certain meta is restricting oneself into a smaller subset of this tree; it can save your thinking time and can give you better results if the opponent plays as you anticipate. However, that is not the advantage of knowing the meta, but is instead from the meta being suboptimal and your opponent still following it.

"Knowing how Fool's Gold rushes versus engine payload games play out in the current metagame" this I am not sure what you are talking about... how it plays out, it will play out, how does that depend on any meta?
Title: Re: Meta?
Post by: tristan on October 12, 2016, 08:20:16 pm
So here's an example of what I was talking about:

http://forum.dominionstrategy.com/index.php?topic=16382.new#new

Opponent buys a Province on T6.  And wins.
So your opponent palys different then you, buys a Province on turn 6 and wins. Still missing the metagame part.
Title: Re: Meta?
Post by: Mic Qsenoch on October 12, 2016, 08:26:14 pm
Any time I have seen streams or read game reports where a person seems to be using "meta" considerations about what their opponent is likely to do I've thought they made a worse choice because of it.
Title: Re: Meta?
Post by: eHalcyon on October 12, 2016, 09:03:08 pm
There's PLENTY of hidden information in Dominon, that's ridiculous. You have no certainty over what cards the opponents will buy in the future, even if you know all their options, and you can use knowledge of the metagame as one of many factors in predicting their likely strategy. That alone is sufficient evidence of metagame influences in Dominion.

Future information does not count as hidden information. Think about chess. Or any board game. If you know for certain what your opponent will do then there is not much of a game right?

You seem to confuse between concepts. Meta is called meta because it is something that affects the game that is outside the game. Think about rock-paper-scissors. If you just play one game, you will do as good as you can by always play rock. But if your opponent knows this then you will lose all the time.

This is what meta is about: you have several mutual exclusive strategy options which counters each other (while nash equilibrium would be a probability mixture of them). How you choose one among is really by preference, and is the hidden information I am talking about. The population preference is the current metagame.

Not so much in Dominion. The population can certainly have a preference on different strategies, but it is pure suboptimality. For a given board, one can always react to the way the other player is playing. The optimal strategy is a script of actions taken when seeing what opponent plays, that will give you the highest win percentage. This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

So are you saying that Rock-Paper-Scissors has a meta?  Because your last paragraph could apply to it just as well:

The population can cetainly have a preference on different strategies, but it is pure suboptimality.  For a given series of RPS games, the optimal strategy is to pick your action with perfect randomness every round.  This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

But in practice, people don't play optimally, they don't do perfect random.  So your RPS strategy can exploit that, depending on the meta.  Doesn't that also apply to Dominion?  People don't play optimally, they may favour certain strategies based on their own preferred style or perceived strength of cards.  Isn't that a meta?
Title: Re: Meta?
Post by: Witherweaver on October 13, 2016, 03:05:37 pm
Any time I have seen streams or read game reports where a person seems to be using "meta" considerations about what their opponent is likely to do I've thought they made a worse choice because of it.

Maybe you could track what people seem to use such 'meta' considerations in their game play and, when you play against them, use it to your advantage...
Title: Re: Meta?
Post by: Chris is me on October 13, 2016, 03:39:07 pm
There's PLENTY of hidden information in Dominon, that's ridiculous. You have no certainty over what cards the opponents will buy in the future, even if you know all their options, and you can use knowledge of the metagame as one of many factors in predicting their likely strategy. That alone is sufficient evidence of metagame influences in Dominion.

Future information does not count as hidden information. Think about chess. Or any board game. If you know for certain what your opponent will do then there is not much of a game right?

You seem to confuse between concepts. Meta is called meta because it is something that affects the game that is outside the game. Think about rock-paper-scissors. If you just play one game, you will do as good as you can by always play rock. But if your opponent knows this then you will lose all the time.

This is what meta is about: you have several mutual exclusive strategy options which counters each other (while nash equilibrium would be a probability mixture of them). How you choose one among is really by preference, and is the hidden information I am talking about. The population preference is the current metagame.

Not so much in Dominion. The population can certainly have a preference on different strategies, but it is pure suboptimality. For a given board, one can always react to the way the other player is playing. The optimal strategy is a script of actions taken when seeing what opponent plays, that will give you the highest win percentage. This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

So are you saying that Rock-Paper-Scissors has a meta?  Because your last paragraph could apply to it just as well:

The population can cetainly have a preference on different strategies, but it is pure suboptimality.  For a given series of RPS games, the optimal strategy is to pick your action with perfect randomness every round.  This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

But in practice, people don't play optimally, they don't do perfect random.  So your RPS strategy can exploit that, depending on the meta.  Doesn't that also apply to Dominion?  People don't play optimally, they may favour certain strategies based on their own preferred style or perceived strength of cards.  Isn't that a meta?

I don't know how you define optimal, but random is not the RPS strategy which maximizes your win rate in competitive play. That's why people who competitively play RPS sustain win rates of more than 33%.

RPS absolutely has a metagame. That is basically the entire strategic depth of RPS, the metagame.
Title: Re: Meta?
Post by: Witherweaver on October 13, 2016, 03:51:41 pm
I'm sure Wine Merchant would have something to say about meta.
Title: Re: Meta?
Post by: madeofghosts on October 14, 2016, 05:25:38 am
I don't know how you define optimal, but random is not the RPS strategy which maximizes your win rate in competitive play. That's why people who competitively play RPS sustain win rates of more than 33%.

RPS absolutely has a metagame. That is basically the entire strategic depth of RPS, the metagame.

...I feel better about my life.
Title: Re: Meta?
Post by: markusin on October 14, 2016, 07:31:38 am
There's PLENTY of hidden information in Dominon, that's ridiculous. You have no certainty over what cards the opponents will buy in the future, even if you know all their options, and you can use knowledge of the metagame as one of many factors in predicting their likely strategy. That alone is sufficient evidence of metagame influences in Dominion.

Future information does not count as hidden information. Think about chess. Or any board game. If you know for certain what your opponent will do then there is not much of a game right?

You seem to confuse between concepts. Meta is called meta because it is something that affects the game that is outside the game. Think about rock-paper-scissors. If you just play one game, you will do as good as you can by always play rock. But if your opponent knows this then you will lose all the time.

This is what meta is about: you have several mutual exclusive strategy options which counters each other (while nash equilibrium would be a probability mixture of them). How you choose one among is really by preference, and is the hidden information I am talking about. The population preference is the current metagame.

Not so much in Dominion. The population can certainly have a preference on different strategies, but it is pure suboptimality. For a given board, one can always react to the way the other player is playing. The optimal strategy is a script of actions taken when seeing what opponent plays, that will give you the highest win percentage. This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

So are you saying that Rock-Paper-Scissors has a meta?  Because your last paragraph could apply to it just as well:

The population can cetainly have a preference on different strategies, but it is pure suboptimality.  For a given series of RPS games, the optimal strategy is to pick your action with perfect randomness every round.  This grand optimal strategy does not depend on any preference of the opponent so there is no meta.

But in practice, people don't play optimally, they don't do perfect random.  So your RPS strategy can exploit that, depending on the meta.  Doesn't that also apply to Dominion?  People don't play optimally, they may favour certain strategies based on their own preferred style or perceived strength of cards.  Isn't that a meta?

I don't know how you define optimal, but random is not the RPS strategy which maximizes your win rate in competitive play. That's why people who competitively play RPS sustain win rates of more than 33%.

RPS absolutely has a metagame. That is basically the entire strategic depth of RPS, the metagame.

I think this was another topic somewhere, but can you clarify something on this? Is pure random not the best strategy because there are players who are not playing pure random that can be exploited? Is it because even if you play many matches against a pure random opponent, and only a pure random opponent, there is a strategy that beats it somehow?
Title: Re: Meta?
Post by: Chris is me on October 14, 2016, 08:38:20 am
I'm not the biggest RPS expert, but for a few reasons:

Title: Re: Meta?
Post by: markusin on October 14, 2016, 09:36:37 am
I'm not the biggest RPS expert, but for a few reasons:

  • RPS is usually played best of 7, so past throws and psychology come into play, as well as "popular" opening algorithms (Rock Rock Rock, or Paper Scissors Scissors)
  • Humans are extremely bad at making "random" decisions unassisted. They just can't do it.

Oh, I meant in the RPS programming competition DXV had brought up at some point. That involves computer bot programs playing RPS against each other, and even there the random bots apparently don't win the championship.
Title: Re: Meta?
Post by: Witherweaver on October 14, 2016, 09:37:26 am
Humans are extremely bad at making "random" decisions unassisted. They just can't do it.

Ha!  We have three threads that say otherwise!
Title: Re: Meta?
Post by: eHalcyon on October 14, 2016, 07:37:10 pm
I'm not the biggest RPS expert, but for a few reasons:

  • RPS is usually played best of 7, so past throws and psychology come into play, as well as "popular" opening algorithms (Rock Rock Rock, or Paper Scissors Scissors)
  • Humans are extremely bad at making "random" decisions unassisted. They just can't do it.

That was my point though.  The meta develops because players aren't optimal and can't do perfect random.  But against a player that is perfectly random, you can't do better than pure random yourself.  Why doesn't this apply to Dominion, where players are  also unable to be optimal?

If you define "optimal" and "meta" to account for that sort of exploitation (of other players psychology, etc.) then it should also apply for Dominion, in which case Dominion can certainly have a meta.
Title: Re: Meta?
Post by: markusin on October 14, 2016, 07:48:30 pm
I'm not the biggest RPS expert, but for a few reasons:

  • RPS is usually played best of 7, so past throws and psychology come into play, as well as "popular" opening algorithms (Rock Rock Rock, or Paper Scissors Scissors)
  • Humans are extremely bad at making "random" decisions unassisted. They just can't do it.

That was my point though.  The meta develops because players aren't optimal and can't do perfect random.  But against a player that is perfectly random, you can't do better than pure random yourself.  Why doesn't this apply to Dominion, where players are  also unable to be optimal?

If you define "optimal" and "meta" to account for that sort of exploitation (of other players psychology, etc.) then it should also apply for Dominion, in which case Dominion can certainly have a meta.

What about Dominion with more than 2 players? Surely there is sort of meta that can develop there where the best strategy is more likely to be dictated by what the other players do.
Title: Re: Meta?
Post by: Awaclus on October 15, 2016, 12:10:15 am
I'm not the biggest RPS expert, but for a few reasons:

  • RPS is usually played best of 7, so past throws and psychology come into play, as well as "popular" opening algorithms (Rock Rock Rock, or Paper Scissors Scissors)
  • Humans are extremely bad at making "random" decisions unassisted. They just can't do it.

That was my point though.  The meta develops because players aren't optimal and can't do perfect random.  But against a player that is perfectly random, you can't do better than pure random yourself.  Why doesn't this apply to Dominion, where players are  also unable to be optimal?

If you define "optimal" and "meta" to account for that sort of exploitation (of other players psychology, etc.) then it should also apply for Dominion, in which case Dominion can certainly have a meta.

What about Dominion with more than 2 players? Surely there is sort of meta that can develop there where the best strategy is more likely to be dictated by what the other players do.

The best strategy in Dominion with more than 2 players is to not play it.
Title: Re: Meta?
Post by: pacovf on October 15, 2016, 01:28:13 am
I don't know how you define optimal, but random is not the RPS strategy which maximizes your win rate in competitive play. That's why people who competitively play RPS sustain win rates of more than 33%.

I get what you mean, but surely 50% is the number you wanted to use here?

I'm not the biggest RPS expert, but for a few reasons:

  • RPS is usually played best of 7, so past throws and psychology come into play, as well as "popular" opening algorithms (Rock Rock Rock, or Paper Scissors Scissors)
  • Humans are extremely bad at making "random" decisions unassisted. They just can't do it.

Oh, I meant in the RPS programming competition DXV had brought up at some point. That involves computer bot programs playing RPS against each other, and even there the random bots apparently don't win the championship.

Not knowing any specifics about RPS bot competitions, the moment somebody submits a bot that isn't perfectly random, then the optimal strategy ceases to be a random bot, because random bots won't be able to exploit any deficiency in the programming of that one non-random bot.
Title: Re: Meta?
Post by: Awaclus on October 15, 2016, 05:28:51 am
I don't know how you define optimal, but random is not the RPS strategy which maximizes your win rate in competitive play. That's why people who competitively play RPS sustain win rates of more than 33%.

I get what you mean, but surely 50% is the number you wanted to use here?

If your strategy is full random, your win rate is 1/3 (and so are your lose rate and tie rate).
Title: Re: Meta?
Post by: AdrianHealey on October 15, 2016, 06:13:54 am
I could be wrong, but it doesn't seem to hard to go full random in RPS.

Just throw X amount of dices (x=how many games you are playing) and do whatever the result of those dices tell you.
Title: Re: Meta?
Post by: Titandrake on October 15, 2016, 12:35:57 pm
I could be wrong, but it doesn't seem to hard to go full random in RPS.

Just throw X amount of dices (x=how many games you are playing) and do whatever the result of those dices tell you.

When RPS is normally played, you don't get to flip coins or roll dice before making your throw. You can mentally roll dice, but that turns into the same problem. People are bad at simulating random behavior.

I guess you could roll a bunch of dice before, memorize how they landed, and call on that stream of random numbers whenever you're asked to play RPS, but that seems like way too much work. (And it also seems less fun than trying to predict your opponent's throw.)
Title: Re: Meta?
Post by: Awaclus on October 15, 2016, 12:40:14 pm
I guess you could roll a bunch of dice before, memorize how they landed, and call on that stream of random numbers whenever you're asked to play RPS, but that seems like way too much work. (And it also seems less fun than trying to predict your opponent's throw.)

It is, however, a strategy that has been actually done by some competitive RPS players.
Title: Re: Meta?
Post by: pacovf on October 15, 2016, 01:50:42 pm
If your strategy is full random, your win rate is 1/3 (and so are your lose rate and tie rate).

That's only true for a single round. Draws only happen (1 + 7*6 + 7*6*5*4 /4 + 7*6*5*4*3*2 /(6*6)) / 3^7 ~ 18% of the time for a best-of-7 match, and the probability decreases as the number of rounds increases, which, for bots, can be arbitrarily large.
Title: Re: Meta?
Post by: GendoIkari on October 16, 2016, 08:48:41 am
I'm not the biggest RPS expert, but for a few reasons:

  • RPS is usually played best of 7, so past throws and psychology come into play, as well as "popular" opening algorithms (Rock Rock Rock, or Paper Scissors Scissors)
  • Humans are extremely bad at making "random" decisions unassisted. They just can't do it.

That was my point though.  The meta develops because players aren't optimal and can't do perfect random.  But against a player that is perfectly random, you can't do better than pure random yourself.  Why doesn't this apply to Dominion, where players are  also unable to be optimal?

If you define "optimal" and "meta" to account for that sort of exploitation (of other players psychology, etc.) then it should also apply for Dominion, in which case Dominion can certainly have a meta.

This is less true in Dominion because in RPS, the best strategy is 100% based on what strategy your opponent chooses. In other words, knowing what your opponent will do, plus a very basic and obvious strategy for using that information, guarantees a win every time. In Dominion, this isn't true. With very rare boards, you can have A>B>C>A, but not normally. Normally if A is best, then A is best no matter what your opponent does.
Title: Re: Meta?
Post by: Amac on October 16, 2016, 05:19:41 pm
In an optimal strategy, you don't need to make any presumptions of your opponent's strategy.  You watch what they do, and react to that.  If you react to what you predict your opponents will do, then they can just not do the thing.

However, people don't play optimally, therefore there is meta.  The fact that there is meta is proof that people aren't playing optimally.

The most obvious kind of meta is when newbies think Thief is really devastating, so the best strategy is to prepare for some Thieves.  However, I suspect that the metagame online is more subtle, since the players are better?  I dunno, I play offline only.

If there is no dominant strategy, the optimal strategy is dependent on the plays of the other player(s). This is the core of game theory. Meta is trying to optimize if there is no dominant strategy: either beating a large as possible set of strategies (in ccg's, as you can't change your starting deck you just need to build a deck that beats the largest set of predicted strategies. Meta comes in when you have to predict the probability of a certain deck being played: One includes TeCH to try and beat the strategies with largest probabilites), or trying to adapt to your opponents strategy.

Adapting to your opponents strategy is all what dominion is about. If your opponent is playing a rush, you want your engine to go off earlier than if he plays big money. If Tunnel and Militia are on the board (and no other Tunnel enablers), you don't want to buy Tunnel except if your opponent buys 5 Militia, for instance. This is just adapting the strategy set to what the other player buys.

Anyway, I guess there is meta, it predominantly comes from the approach of greening. There is some optimal way to green, but it is not easy to understand it. The meta evolution made greening an ever-changing process. I don't know if there are other large meta-changes. Maybe the acceptance of JOAT and Rebuild after time (and simulations) as good cards can considered to be changing the metagame. But also the inclusion of new sets changes the approach of some cards. I take Upgrade as an obvious example. At first, everybody accepted Upgrade as a bona-fide copper-trasher if it wasn't used for anything better and the copper wasn't completely necessary. But then came Poor House, a card costing 1. It suddenly changes the approach of the card when that was on the board: the copper-trashing ability then is posed with the obvious problem of needing to gain a Poor House.

The small problem is that because of the 'we only use 10 cards'-approach, these additions can change the game, but don't necessary do it. Still, having the cards around in the selection procedure does change the value of some cards. This is most obvious with Alchemy. Most Alchemy cardss are better when fewer sets are included in the selection procedure.

So, there is meta in both ways: The decision making of the players given a certain set is meta, as well as the change in used cards in the selection procedure. It is not as huge as in a CCG like Hearthstone or Magic, or something like Pokemon battling. Also contributing to that is that there is no banlist or something: In Pokemon we consider Mewtwo to be way overpowered battling in Ubers, whilst Tauros is declined to the depth of Never Used. Both metagames with complete different banlists. Dominion doesn't consider these banlists, or rather, has a randomized banlist consisting of all of the kingdom cards except 10 of them. If one considers that, there are an enormous amount of possible metagames. Therefore, we can't analyse them (well, we can, but we don't consider all individual kingdoms obviously), and subsequently don't feel like considering a metagame analyzing kingdoms at all (or at least I do). But I guess they each include their own meta, strictly speaking.
Title: Re: Meta?
Post by: allanfieldhouse on October 18, 2016, 02:50:36 pm
I think of the meta as generally being "the strategies that people use". Sure if there's one dominant strategy, everyone will just play that, and it wouldn't really make sense to use the word meta. But that doesn't apply to Dominion.

In Dominion you can see exactly what your opponent buys/plays, so there's less of a meta than there is in a game with hidden information like Starcraft (or Magic where you don't have the same cards as your opponent).

It's still important to have a general idea of what your opponent is probably aiming for in addition to what you can see from their actual buys.
Title: Re: Meta?
Post by: eHalcyon on October 18, 2016, 05:38:11 pm
I'm not the biggest RPS expert, but for a few reasons:

  • RPS is usually played best of 7, so past throws and psychology come into play, as well as "popular" opening algorithms (Rock Rock Rock, or Paper Scissors Scissors)
  • Humans are extremely bad at making "random" decisions unassisted. They just can't do it.

That was my point though.  The meta develops because players aren't optimal and can't do perfect random.  But against a player that is perfectly random, you can't do better than pure random yourself.  Why doesn't this apply to Dominion, where players are  also unable to be optimal?

If you define "optimal" and "meta" to account for that sort of exploitation (of other players psychology, etc.) then it should also apply for Dominion, in which case Dominion can certainly have a meta.

This is less true in Dominion because in RPS, the best strategy is 100% based on what strategy your opponent chooses. In other words, knowing what your opponent will do, plus a very basic and obvious strategy for using that information, guarantees a win every time. In Dominion, this isn't true. With very rare boards, you can have A>B>C>A, but not normally. Normally if A is best, then A is best no matter what your opponent does.

I contend that even though there's less impact in Dominion, it's not totally absent.  Sometimes you have boards with A>B>C>A, so that already supports my point, even if it's rare.  But more commonly -- and more importantly! -- I believe that players usually won't perfectly identify and execute the optimal strategy.  Moreover, I reject the idea that Dominion is merely "multiplayer solitaire".  Your opponents choices matter and should be considered in your strategy.
Title: Re: Meta?
Post by: Amac on October 19, 2016, 08:32:37 am


I contend that even though there's less impact in Dominion, it's not totally absent.  Sometimes you have boards with A>B>C>A, so that already supports my point, even if it's rare.  But more commonly -- and more importantly! -- I believe that players usually won't perfectly identify and execute the optimal strategy.  Moreover, I reject the idea that Dominion is merely "multiplayer solitaire".  Your opponents choices matter and should be considered in your strategy.

The fact that we can't even determine for sure what the optimal play would be in quite some situations here on the site, and that the game is not 'score the most points', but 'score more points than your opponent', does imply there must be metagaming (possibilities) of some sort.

Even chess has metagaming (so-called 'plugging' is the most common way), and there have been many, many years of analysis on chess.

I'm inclined to say that every game that lacks a clear dominant strategy has possible hidden information and therefore implies the existence of metagaming.