Diplo game thought experiment

[SevenSpirits]

It also goes to the heart of why RB culture is the way it is: "good" play means "playing civ" well. Clever micro, solid buildering, and tactics rooted in a deep understanding of the game design. Perhaps inherited from the tradition of succession games and adventures. Its true that the game design makes aggression predictable (graphs) and easy to counter (whips/drafts) but that doesn't fully explain the meta.

The design is equally compatible with a culture that values aggression and constant war. A pitboss where all the players value aggression and believe they are going to be rushed, and so a pitboss where all the players build early military and constantly harass their neighbours is also an equilibrium. Individual players who try to try to break the meta by building in peace are either crushed or forced to comply with the meta by their more aggressive neighbours. You might further reinforce the meta by labeling the peaceful players as "wimpy" and "stupid" - the expert players in the community know that early aggression is the only way to go! They spend all their time in the forum talking about ways to psych out opponents and all sorts of tactical tricks and traps for warfare and diplomacy.

Its only when you get a group of players who agree that war is stupid that the meta begins to shift to the current equilibrium. If you have enough players with this belief, some of them still get rushed, but one player, somewhere, is left to build in peace for long enough to secure an economic advantage and win the game. Everyone forgets about the other losing teams, remembers the winning team, and adjusts their behaviour accordingly. The shift is reinforced by a value system that rewards peaceful play (that micro was awesome!) and a shift in culture - now everybody is talking about efficient worker movement and optimized city development. They keep elaborate threads that document their economic development and share interesting tidbits about game mechanics. The game hasn't changed but the way players play it has.

That is not true: crazy aggression is not an equilibrium because we are talking about an FFA game of attrition where the defender has an advantage. If the meta is currently to build X amount of military and harass people, then building X military and keeping it all at home for defense trumps it. Building X-1 military and defending is even better. Over time, X keeps getting smaller until it reaches the effective minimum of military that you want to build for some given settings.

Because of demos and power graphs, this isn't even a simultaneous selection game theory matrix where building X+5 military every now and then trumps the normal strategy because you get to conquer someone. People can see your high military before you get to use it, and adjust their X-1 strategy accordingly. It's true, there are rare situations where a player might be able to ascertain an opponent's gambled temporary vulnerability and get a small advantage from it - and that does happen in RB games. But this is not nearly on the scale of an aggressive metagame.

Why is micro/strategic skill (as opposed to diplo skill) so often thought of as synonymous with skill nowadays? Well first of all I am not totally on board with that claim: we do place value on diplo skill, and if there's been less emphasis on that recently, I'm sure that's in large part because we've been playing games where diplo isn't allowed at all! But to the extent that it is true, I hypothesize that it's because we play with maps that are fairly balanced and therefore a good arena for showcasing micro skill -you can see it against an objective measuring stick as well as compare it directly to the other players! But our player pools for games are often incredibly imbalanced, and players' very bad decisions influence the game more than we like to admit. It's hard to discern whether a given diplomatic/positional coup was more due to one player's genius or another player's derpiness.

[suttree]

If the meta is currently to build X amount of military and harass people, then building X military and keeping it all at home for defense trumps it. Building X-1 military and defending is even better. Over time, X keeps getting smaller until it reaches the effective minimum of military that you want to build for some given settings.

Thanks for replying - I was wrong. This is correct.

[suttree]

Why is there an implicit agreement that they're not going to play as random number generators a form of cooperation within the game?

To be precise, I said "A bet needs to communicate some information at least some of the time..." So players are free to play a completely random betting strategy some of the time and, as you note, this might be a good strategy in some situations. You don't, however, have players who play as an RNG all of the time even though this is a valid strategy. Poker in the abstract is a game where players bet on hands of cards with limited information - that's it - but the players, so that there might be winners and losers, make an implicit agreement that at least some of their bets will communicate information. They do this within the game - the players are collectively ruling out a valid strategy for playing the game. It's a small thing, but I thought it a good place to start in describing how players are motivated to cooperate in poker.

A bet communicates information, but it's not a form of cooperation...
What incentive do I have to cooperate with you to fleece other players, and how exactly would that work anyway?

Suppose you stumble across a pool of players who always play as RNGs. You win the game, then, by instead playing the probabilities, betting good hands so you can profit from players who randomly call. At this point you are communicating information through your bets - they reflect the actual strength of your hand. You're right communication is not a form of cooperation, but it does allow for cooperation. If another player wakes up from his RNG slumber and notices that you are communicating information through your bets, he can adapt his play accordingly - both players playing the probabilities and incorporating the information communicated by the other. In fact, so long as the other players don't notice, there is incentive to maintain this arrangement. All the other players lose money to the information sharing players over time. The players are cooperating to make money off the stupid RNG players. They don't need to talk to do this, they just need to pay attention to bets.

I'm arguing that card betting games in general are like this - every time a player makes a bet that reflects the true strength of his hand he is telling the other players "You can trust me. If you pay attention to my bets, you will make more money than if you ignore them." He needs to do this at least some of the time, else he's again playing as an RNG. This is what I mean when I say cooperation is possible in the game of poker. At least some of the time, a players profits (folding to a better hand) because of information communicated by another player. And both players profit if only they pick up on the information. A temporary information sharing treaty if you like.

While cooperation is possible, no particular arrangment/strategy is stable because only one player can win each individual hand. In the example of the players fleecing the RNG's, both players have an incentive (when the pot is sufficiently large) to lie with their bet and so take all the money instead of splitting it.

And so the game evolves: instead of always communicating the strength of his hand, a player devises a betting strategy, sometimes honest sometimes bluffing. As other players pick up on the strategy, it too serves as a temporary arrangement that benefits those who understand it. Thus giving incentive for a meta-strategy - switching between different betting strategies, and so on....

So let's use this to address your comments:

A bet communicates information, but it's not a form of cooperation. If I make a large bet, I'm telling you that I either have a strong hand and I'm daring you to follow me and lose your money, or I'm pretending to have a strong hand and hoping you'll give up and give me the money that's already been waged.

That's communication, but it's not mutually beneficial to any degree and thus it's not cooperation.

Either you're betting the true strength of your hand or you're bluffing. Bluffing will only be profitable if some other player misreads the bet - that is, only if you have previously established a betting strategy that another player is using to guide his decisions. The game is to make a cooperative arrangement -

"This is my betting strategy. Use it to guide your decisions because it will save you money. If the other players don't notice, we'll profit!"

and then break the arrangement when it earns you a large pot.

What incentive do I have to cooperate with you to fleece other players, and how exactly would that work anyway?

If two players are aware of a betting strategy but the other players are not, they have access to more information than the other players and so make more money in the long run. To be specific - suppose I always know when you're bluffing and you always know when I'm bluffing. In the long term, we don't lose money to each other, but the other players lose money to us. I imagine this would sometimes happen with professional poker players if there are inexperienced players at the table..

[SevenSpirits]

I think a large part of people shying away from a more backstabby diplo is the always-far-off-seeming victory conditions. Most games go on until several people are like: man, I am moving way too many units every turn, screw this game, let's start another one. So people often don't quite actually aim to win: the ultimate goal is to reach a state where everyone wants to concede to you!

List of things not to do if you want everyone to crown you the winner:
* Seriously piss someone off.
* Start a long, bloody war with someone you've pissed off.

I think I was going to say more but then I got sidetracked and now it's late, so screw it, you get this version of the post.