GrogHeads Forum

Digital Gaming => Computer Gaming => Topic started by: FarAway Sooner on March 29, 2014, 12:53:22 AM

Title: A New Model for Diplomacy in 4x Games
Post by: FarAway Sooner on March 29, 2014, 12:53:22 AM
I've seen very few 4x games in any genre where the diplomacy is what I would consider good.  The AI is often capricious and arbitrary, and seldom rationale.  There might be a long list of modifiers you might be able to see, but there's little dynamic to them.  Improve relations by bribing or gifting, or by signing a treaty.  Relations get to a certain point, and you can sign a Non-Agression Pact, followed by a non-bedwetting agreement, consummated by an Alliance.

If you betray an ally, you might get a reputation ding for the rest of the game, as you are untrustworthy.  But I bet if you spend a few hundred gold,  you can offset that penalty.  It's kind of silly.

But that's not the way diplomacy works in the real world.  The United States couldn't declare war on England even if it bloody wanted to.

Why hasn't anybody ever come up with a more intriguing model for diplomacy in a 4x game, where attitudes towards other countries often evolve despite your government's intentions?  And where betraying an ally earns you more than just a slight distrust penalty from everybody else?

Am I alone in thinking there must be some other model out there, somewhere, that might be vaguely more interesting and strategic than "If you give me enough gold, I'll sleep with the wild pigs outside your second-largest city"?
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Nefaro on March 29, 2014, 01:33:33 AM
When you find one with a more human-like Diplo model, let me know.  I don't think developers are willing to spend the time producing an intricate & believable Diplomacy AI and the oodles of extra time to test and tweak it. 

I would also like to have something more advanced than just a point threshold system.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: tgb on March 29, 2014, 06:08:33 AM
Quote from: FarAway Sooner on March 29, 2014, 12:53:22 AM
I've seen very few 4x games in any genre where the diplomacy is what I would consider good.  The AI is often capricious and arbitrary, and seldom rationale.  There might be a long list of modifiers you might be able to see, but there's little dynamic to them.  Improve relations by bribing or gifting, or by signing a treaty.  Relations get to a certain point, and you can sign a Non-Agression Pact, followed by a non-bedwetting agreement, consummated by an Alliance.

If you betray an ally, you might get a reputation ding for the rest of the game, as you are untrustworthy.  But I bet if you spend a few hundred gold,  you can offset that penalty.  It's kind of silly.

But that's not the way diplomacy works in the real world.  The United States couldn't declare war on England even if it bloody wanted to.

Why hasn't anybody ever come up with a more intriguing model for diplomacy in a 4x game, where attitudes towards other countries often evolve despite your government's intentions?  And where betraying an ally earns you more than just a slight distrust penalty from everybody else?

Am I alone in thinking there must be some other model out there, somewhere, that might be vaguely more interesting and strategic than "If you give me enough gold, I'll sleep with the wild pigs outside your second-largest city"?

Paradox system probably comes closest, imo, although they aren't 4X games, strictly speaking.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: bbmike on March 29, 2014, 07:13:05 AM
For me the best 4x diplomacy experience I've had seems to be Alpha Centauri. The next best might be GalCiv2. I'm really disappointed in the diplomacy experience in Civilization 5. That's one game you would think that they should really should be trying to create a more realistic diplomacy model.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: MetalDog on March 29, 2014, 07:26:03 AM
I'm not sure you would want a 'real life' system for diplomacy.  Human beings are capricious, stubborn, volatile, selfish, etc.  There are no guarantees even if they GIVE you a guarantee. 

And how exactly would you model that?  Randomly accept or reject any proposal you might put to it?  Even knowing you had overwhelming odds in your favor, the AI might still decline.  It's human nature not to be bullied into something.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Rayfer on March 29, 2014, 08:32:58 AM
Perhaps multi-player is the only truly 'realistic' diplomacy model? On second thought, maybe it's the worst? I'm not sure.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: bbmike on March 29, 2014, 10:36:38 AM
I don't think you could model a 'real life' diplomacy for these games. But you should be able to at least have system where friends, enemies, alliances, ect. are possible. If in Civ 5, for example, my ally for the entire game suddenly decides that the end of the game is near and attacks me just to 'win the game', that is not diplomacy. Also, games that say they have diplomacy and only give you two options (war or peace) that also is not diplomacy.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: skeptical.platypus on March 29, 2014, 11:56:25 AM
I think it's unreasonable to expect a realistic diplomatic model in games because they ARE a zero-sum game, real life isn't, and realistic diplomacy reflects that.

Let me turn that around -- how different would diplomacy work in the real world if every single player (individual or entity, you pick the scale) knew there was only one winner?

Now think about how diplomacy would work if the only winner was bacon.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: skeptical.platypus on March 29, 2014, 12:16:54 PM
I imagine I'm basically describing a game of game theory here, but if it were be possible to have multiple "win" options for both AI and human players, and those win options were not necessarily mutually exclusive, wouldn't diplomacy becomes a more critical tool? I have no idea what the programming challenges* would be, or how many human beings would play a game where one of the ways to win was not to lose.


(*I do understand that involves limericks. Or algorithms. I forget which, but it is easy to confuse the two. (I just noticed -- did Al Gore invent Rithm?))
Title: Re: A New Model for Diplomacy in 4x Games
Post by: FarAway Sooner on March 29, 2014, 12:36:02 PM
Just to be clear:  I'm not asking somebody to model all the personality interactions between two leaders, add in all the competing interests between two nations, factor in all the differing cultural traits between two peoples, and generate a model that measures the dynamic interactions between all those things.

I just want something where I'm not free to do whatever I want, within the singular confines of a point-threshold system or a "wait one turn" limitation.  And I want something that treats diplomacy as more than just the interactions between two individual people.

As a good first step, I'd love to see some sort of system which differentiates between my people's attitudes towards a different country and my government's attitudes towards that country. 

FDR wanted to declare war on Germany in 1939 but he didn't have the political consensus to do so.  The Italian army in large part fought like dogs from 1939-1942 because they were fighting against people whom they considered cultural peers, on behalf of people whom they'd fought against bravely and tenaciously 25 years earlier.  The US had to do an abrupt and embarrassing about-face when we relabeled the mujahedin from anti-Soviet freedom fighters to Moslem extremist terrorists.  I'd like to face some internal political consequences when I go to war with somebody who was recently an ally.  I'd like to be hemmed in by the tactically/politically expedient choices that I made 5-10 years ago.

There ought to be some way to model popular sentiment towards other countries--and popular sentiment towards a war--independently from an arbitrary "war fatigue" penalty if I've been at war too long.

tgb
, how does Paradox handle diplomacy?
Title: Re: A New Model for Diplomacy in 4x Games
Post by: panzerde on March 29, 2014, 01:01:39 PM
Quote from: skeptical.platypus on March 29, 2014, 11:56:25 AM
Now think about how diplomacy would work if the only winner was bacon.

Kevin Bacon or tasty, tasty strips of fried pig fat? Because I thinks there are very different answers there.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Nefaro on March 29, 2014, 03:02:15 PM
Quote from: tgb on March 29, 2014, 06:08:33 AM

Paradox system probably comes closest, imo, although they aren't 4X games, strictly speaking.

They have more options, but are still based on a simple +/- point system, with thresholds.  Not so much a diplo-specific decision making AI, but having point thresholds that are fairly predictable.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: bobarossa on March 29, 2014, 03:05:47 PM
Quote from: FarAway Sooner on March 29, 2014, 12:36:02 PM
I just want something where I'm not free to do whatever I want, within the singular confines of a point-threshold system or a "wait one turn" limitation.  And I want something that treats diplomacy as more than just the interactions between two individual people.

As a good first step, I'd love to see some sort of system which differentiates between my people's attitudes towards a different country and my government's attitudes towards that country. 

FDR wanted to declare war on Germany in 1939 but he didn't have the political consensus to do so.  The Italian army in large part fought like dogs from 1939-1942 because they were fighting against people whom they considered cultural peers, on behalf of people whom they'd fought against bravely and tenaciously 25 years earlier.  The US had to do an abrupt and embarrassing about-face when we relabeled the mujahedin from anti-Soviet freedom fighters to Moslem extremist terrorists.  I'd like to face some internal political consequences when I go to war with somebody who was recently an ally.  I'd like to be hemmed in by the tactically/politically expedient choices that I made 5-10 years ago.

There ought to be some way to model popular sentiment towards other countries--and popular sentiment towards a war--independently from an arbitrary "war fatigue" penalty if I've been at war too long.

tgb
, how does Paradox handle diplomacy?
If you are talking the EU games, if your country does not have a causus belli, then you take a stability hit (reduces income, increases revolt risk, etc).  Your government needs to work to get this causus belli before you declare war.  In HOI games, you have to lower your 'neutrality' level until it reaches a threshold based on other country's 'threat' level to you.  Otherwise, you can't declare war.  You can use spies to increase a country's threat level.  Depending on game version, you can have a minister that lowers your neutrality a little each month.

Speaking of interesting real-life diplomacy, I just finished reading The Winter War (Edwards).  Because of MR pact, Hitler stood by while SU invaded Finland even though German population supported Finland.  Even Goring (who had married a Swedish noblewoman) sent armaments to Finland on the sly.  Italy also tried to send munitions but Hitler forbade it passing through Germany. 

And as Nefaro pointed out while I was typing, EU games have a Bad Boy rating system that goes up when you annex provinces and goes down as time passes.  You have to balance absorbing enemies with the chance your rating gets high enough to cause neighbors to declare war on you.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Nefaro on March 29, 2014, 03:19:44 PM
I realize that, being a game, it will require numbers to quantify things.  I would just like to see a deeper system with more than one or two numbers keeping tally on diplomatic standings and being used for AI decisions. 

This may require a whole battery of various values and a lot more AI work but it would be quite nice to have a grand strategy game with a more intricate diplomatic system.  With a good AI programmer, they could make a more sophisticated diplo system.  The Paradox games probably are some of the most advanced in this area, but they're still pretty thin on depth when it all still comes down to a single Relations number and an extra switch or two (in EU4).  Still makes it feel like you're just pushing that AI nation's whole wants & needs around via jockeying a single number.

Hell, I was ecstatic to find out about PrON's Diplo Crisis mini-game, which broke this mold.  Evidently Pdox liked it too, because they tried adding something similar to Vic2 in a later expansion.  Dunno how well that system works in Vic2, as I've not played it far enough to find out yet.  Still.. breaking out of the box like that is something we need to see more.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: FarAway Sooner on March 30, 2014, 12:06:54 AM
Yeah.  That's all I'm looking for, is an effort to break out of the box and get beyond one main number and two or three qualifying rules.  What might such a system look like?

I'd be happy to just see a game where my people's attitudes towards an enemy were tracked separately from my own status with them (i.e., War, Peace, Non-Aggression Pact, etc.).  As was the case in the EU games, if you go to war with somebody where there's not a proper casus belli, you pay the price in domestic turmoil. 

The EU system actually abstracts that stuff, but still at least makes it more plausible.  Adding in the occasional "story line" random event might make for interesting decisions.  Do you use a particular international incident as a chance to maximize short-term resources, or do you play for an outcome that will actually antagonize your own public the most so it's easier to really go to war later on?

That sort of thing would be more interesting than allowing me to pivot on a dime if I feel like it.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Nefaro on March 30, 2014, 01:26:31 AM
To EU4's credit, they did add an extra generalized stance (Neutral, Wary, etc) as evidenced by the little symbols next to your relations rating.  However, it may only be an extra permanent bonus or penalty to the usual Relations rating.  I think it can change if you pull some dastardly stuff with them, like breaking an alliance or something, and it lasts much longer.  So it's a start.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: tgb on March 30, 2014, 08:49:44 AM
One of the later Paradox games (I forget if it was Vicky 2, EU3, or CK2) introduced random events that could create a CB. It was the players' decision whether or not to act on it.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Bletchley_Geek on March 30, 2014, 07:11:13 PM
Quote from: FarAway Sooner on March 29, 2014, 12:36:02 PM
Just to be clear:  I'm not asking somebody to model all the personality interactions between two leaders, add in all the competing interests between two nations, factor in all the differing cultural traits between two peoples, and generate a model that measures the dynamic interactions between all those things.

I just want something where I'm not free to do whatever I want, within the singular confines of a point-threshold system or a "wait one turn" limitation.  And I want something that treats diplomacy as more than just the interactions between two individual people.

As a good first step, I'd love to see some sort of system which differentiates between my people's attitudes towards a different country and my government's attitudes towards that country. 

FDR wanted to declare war on Germany in 1939 but he didn't have the political consensus to do so.  The Italian army in large part fought like dogs from 1939-1942 because they were fighting against people whom they considered cultural peers, on behalf of people whom they'd fought against bravely and tenaciously 25 years earlier.  The US had to do an abrupt and embarrassing about-face when we relabeled the mujahedin from anti-Soviet freedom fighters to Moslem extremist terrorists.  I'd like to face some internal political consequences when I go to war with somebody who was recently an ally.  I'd like to be hemmed in by the tactically/politically expedient choices that I made 5-10 years ago.

There ought to be some way to model popular sentiment towards other countries--and popular sentiment towards a war--independently from an arbitrary "war fatigue" penalty if I've been at war too long.

Paradox is probably the closest thing to what you're asking for - as others have already noted. EU4 diplomacy is probably the best I've experienced in a strategy game (the Coalition mechanics, where states get loosely associated to bring down another state perceived to be a "threat", typically France, are indeed a step forward to bring more meaningful interactions). Nonetheless, it still has some shortcomings (in my opinion):

Title: Re: A New Model for Diplomacy in 4x Games
Post by: Martok on April 01, 2014, 03:29:04 PM
Armada 2526 Supernova and Fallen Enchantress: Legendary Heroes are two 4x titles where I feel the diplomacy isn't completely whacked, but even they're both deeply flawed.  (Anyone who's played the former can testify to AI suddenly declaring a trade embargo on you for seemingly capricious reasons...) 


As has already been stated here, I think the biggest problem is that no one has been willing to devote the resources to create an AI that truly handles diplomacy well.  Not that I blame anyone -- I'm sure it would be difficult to justify investing the resources necessary to do so, as the quality of AI players is one of those "intangibles" that's hard to quantify the value they add to a strategy game. 




Quote from: FarAway Sooner on March 30, 2014, 12:06:54 AM
I'd be happy to just see a game where my people's attitudes towards an enemy were tracked separately from my own status with them (i.e., War, Peace, Non-Aggression Pact, etc.).  As was the case in the EU games, if you go to war with somebody where there's not a proper casus belli, you pay the price in domestic turmoil. 
Armada 2526 has something like this.  It sadly doesn't have any true effect, however (i.e., declaring war on a nation your people like lowers your colonie's collective happiness/loyalty). 

Title: Re: A New Model for Diplomacy in 4x Games
Post by: MuzzyDev on April 02, 2014, 02:11:02 PM
I thought Civ 4 had some pretty decent diplomacy.  There were a number of factors that made relations positive or negative: frequency of trade, disputed borders, similar religions, helped them in a war, etc.   You had a pretty decent sense when your relations were good or on the rocks--most of the time.  But a few civs' actions seemed to come out of left field.  And that was OK.   Real life diplomacy is unpredictable.

Modeling diplomacy in a strategy game can be a nightmare, precisely because so much of it is predictable--until it's not.  Real life failures in diplomacy frequently happen when a nation expects--for whatever reason--that another nation will act a certain way, and then they don't.  Saying a game is unrealistic because, for example, a nation you were at war with becomes an ally later is not necessarily true.  It may be for our game, which has a 10 year span for gameplay, but Civ, for example, is 4000 years!  Enemies can certainly become allies and then enemies again multiple times over a span like that.

We're working on some interesting levers for diplomacy in The Great War.   The WWI era was very self-serving.  Lots of capricious and irrational behavior.  Some nations were literally up for auction, highest bid earned them an ally.  We want to allow players to use diplomacy as another way to deal with nations as much as going to war.  You will be able to buy influence.  You will be able to try to counter the influence that other nations are attempting.  There will also be betrayals.  Think that country has your back?  They may bail on you if you get into a war.  That's how diplomacy worked during that period. 

I can't speak for other developers, but we are very opposed to designing a game where you can't choose to turn on a dime.  We don't want players to think: "Oh, the designers don't want me to do that."  Because that sucks.  It's much better to let you do what you want, and have to deal with the consequences.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: bbmike on April 02, 2014, 02:50:58 PM
Quote from: MuzzyDev on April 02, 2014, 02:11:02 PM
Saying a game is unrealistic because, for example, a nation you were at war with becomes an ally later is not necessarily true.  It may be for our game, which has a 10 year span for gameplay, but Civ, for example, is 4000 years!  Enemies can certainly become allies and then enemies again multiple times over a span like that.

Which is perfectly acceptable if there is a reason for those diplomatic stances to change. And a valid reason is not "oh, there's only 50 turns left for me to win so I'm going to declare war on whoever is ahead".  Granted diplomacy is hard for a game like Civilization because there are basically two types of players. Those that want more of simulation-ish game and those that want a war/fight to win game.
I agree that diplomacy seemed to be much better in Civ4.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Mr. Bigglesworth on April 02, 2014, 05:16:26 PM
It's easy to make a good model in excel so there is no reason for games not to have one. Well, there is one, the game usually sells on the pizzaz you see not the guts of the system. Of course grogs have different wants than regular gamers.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: FarAway Sooner on April 02, 2014, 06:36:07 PM
An old buddy of mine (who was one of the two lead designers on the Imperialism titles) once said, "Most of the time, when people complain about AI behavior, the root problem is in the way the underlying system was designed.  There's some stuff that AI will just not be good at, regardless of how strong your programming skills are.  The real trick is to build a system that's easy for AI to respond intelligently to."

I wonder whether that's not the case for AI Diplomacy.  The EU IV system sounds a lot like what I'm looking for honestly (and I'm okay that it's a little abstracted, relying on particular minister's skill sets rather than requiring you to specify whether you want to fashion a border incident or stage a false terrorist attack on your own people).

But, on another level, I wonder if we just haven't had much of any originality in how Diplomacy is designed in most games?  It's like we're still using hex-based combat and Attack/Defense Ratios, with a six-sided die, to determine results based on a Combat Results Table (CRT, for those of you who remember!).

I'm not sure what a different system would look like, but I sure wish somebody would try!
Title: Re: A New Model for Diplomacy in 4x Games
Post by: MuzzyDev on April 02, 2014, 07:37:45 PM
LOL!  Easy?  That is one word I've never heard applied to AI in a strategy game.  It does start with good models and systems in the design and on spreadsheets, but that's just the beginning.  If Sid Meier has a difficult time making diplomacy work in a game, I can't imagine that it's particularly easy.  He's got some pretty serious grog credentials.  Paradox has some brilliant people who are grogs, and it's not easy for them, either.

The real limitation to something like diplomacy tends to be performance.  If 50+ nations have to calculate how to react to many possible diplomatic options, the amount of calculations can quickly spin up to where it's really borking your performance, especially a few hundred turns in, when hundreds of units are fighting each other all over the world.  .

I think Sooner's friend is correct, the root problem is the core system.  But you don't make computer games all at once.  Often a flaw is only apparent when a lot of the system is already in.  You can restart but it's always a challenge at the best of times to finish a game before you run out of money.  Good engineers and artists are not cheap.  And in this genre, it's unlikely we're going to sell 6 million copies.

Plus, no matter how successful you are, one multiplayer game will blow away the best AI.  Nothing beats playing against other meatbrains.

Chris--
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Mr. Bigglesworth on April 02, 2014, 07:52:02 PM
I bet Sid is a better grog than me. He understands the game system and the programming. I can't program 10 lines. I see diplomacy as a Operations Research problem. That I can do in excel easily.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: bbmike on April 02, 2014, 07:54:40 PM
Quote from: FarAway Sooner on April 02, 2014, 06:36:07 PM
An old buddy of mine (who was one of the two lead designers on the Imperialism titles)...

:o Have you seen my avatar? You need to rekindle that old friendship ASAP and find out why/where/when Imperialism III is! Polk and the Spieth boys are brilliant!  8)
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Bletchley_Geek on April 02, 2014, 08:05:55 PM
Quote from: MuzzyDev on April 02, 2014, 07:37:45 PM
LOL!  Easy?  That is one word I've never heard applied to AI in a strategy game.  It does start with good models and systems in the design and on spreadsheets, but that's just the beginning.  If Sid Meier has a difficult time making diplomacy work in a game, I can't imagine that it's particularly easy.  He's got some pretty serious grog credentials.  Paradox has some brilliant people who are grogs, and it's not easy for them, either.

The real limitation to something like diplomacy tends to be performance.  If 50+ nations have to calculate how to react to many possible diplomatic options, the amount of calculations can quickly spin up to where it's really borking your performance, especially a few hundred turns in, when hundreds of units are fighting each other all over the world.  .

I think Sooner's friend is correct, the root problem is the core system.  But you don't make computer games all at once.  Often a flaw is only apparent when a lot of the system is already in.  You can restart but it's always a challenge at the best of times to finish a game before you run out of money.  Good engineers and artists are not cheap.  And in this genre, it's unlikely we're going to sell 6 million copies.

I reckon it's a problem with the interface between the game mechanics and the AI. If you can have an interface that offers the AI a simplified representation of the relevant aspects of the game world, along with the capability of 'simulating' the outcomes of diplomatic actions on that simplified representation, then you can apply a huge number of algorithms and techniques derived from the idea of Monte Carlo Tree Search

http://link.springer.com/chapter/10.1007%2F978-3-642-12993-3_3

If you can apply those techniques to a game like Cataan, you can apply it to pretty much any Civ-like game. It wouldn't be hard to devise a way to map the state of a Civ-like game, into the state of a Cataan-like game. The thing is that, in order to apply this concept, one needs to define the architecture of the game in a very particular way from day zero. Retrofitting these interfaces on an existing code base can prove quite difficult.

Regarding performance: I'd say that a turn-based game can be harder, since all decisions (and plays) are made (resolved) in a sequential manner. In a WEGO game, the decision making can happen in parallel (and the resolution of the selected actions is usually *much* faster) making usage of multi-core architectures a no-brainer. In a real-time game, where you basically break down the notion of time into a 'infinite' sequence of time slots, and each slot is 'filled' with some events which are posted by the AI (or by the player, when he issues an action request from the UI), you can schedule the 'AI decision making' events in a way that suits your multi-core architecture best (so the evaluation of the actions is asynchronous). The EU 4 engine exploits this to great advantage.

Not to mention that you don't need *optimal* AI play to provide 90% of the players with a more than decent challenge. You can always limit the 'depth' - how far into the future the AI projects the outcomes of its actions - of the reasoning to your liking, as well as the 'width' of this reasoning (by avoiding to consider all possible outcomes, and restrict either to 'good' outcomes or 'sample' these outcomes on the basis of a time budget).

Listening to Soren Jonsson on 3MA - which is credited with the AI in Civ 4 - I can't help thinking that if he represents the 'state-of-the-art', then the art is indeed quite rudimentary.

Quote from: MuzzyDev
Plus, no matter how successful you are, one multiplayer game will blow away the best AI.  Nothing beats playing against other meatbrains.

Every year that goes by, the dominance of the 'meatbrains' is contested in a different game. I'm not a friend of Kurzweil like pipe dreams, but I think we overrate ourselves quite often. The 99% of chess players in the world can't even dream of competitive play with the top 1%. If you can write a program that can offer an interesting challenge to that 1%, is probably going to be quite challenging for the 99% as well.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Mr. Bigglesworth on April 02, 2014, 09:34:10 PM
And yet players typically develop some basic rules of thumb, they dont go calculating huge tables of permutations. What scared devs, I think, is how the options multiply. You don't have to go through all options, they key is tree pruning.

Even a chess engine might go searching out 5-10 moves for each likely branch. If you add the human type rules of thumb to a pruned search you get through fast with not a lot of number crunching.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: FarAway Sooner on April 02, 2014, 10:09:31 PM
MuzzyDev, in all fairness, even if I remember his word choice correctly, he probably meant "easier" rather than "easy".  He himself admitted that, in Imperialism 2 (their strongest title ever) the AI struggled with late-game decisions around generating and spending revenue.

Quote from: bbmike on April 02, 2014, 07:54:40 PM
Quote from: FarAway Sooner on April 02, 2014, 06:36:07 PM
An old buddy of mine (who was one of the two lead designers on the Imperialism titles)...

:o Have you seen my avatar? You need to rekindle that old friendship ASAP and find out why/where/when Imperialism III is! Polk and the Spieth boys are brilliant!  8)

Mike, I didn't place that avatar until you mentioned it.  Frog City Software closed their doors around 2004, selling out to another firm which was quickly purchased by Take Two Interactive and then shut down a year later.  They loved their work and they loved games, but the revenue stream was too irregular for some folks just starting their family in a high-cost city (Ted and his wife, who was the lead Project Manager for the duration of Frog City's run).

In addition to being a great game designer, Ted also had the best flick I think I've ever seen in 25 years of playing Ultimate Frisbee.  I haven't talked to them since I left the Bay Area, but a close mutual friend of ours passed away from cancer late last year, so I hope to look 'em up the next time I'm back in San Francisco for business travel.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Anguille on April 03, 2014, 12:00:07 AM
Quote from: FarAway Sooner on March 30, 2014, 12:06:54 AM
I'd be happy to just see a game where my people's attitudes towards an enemy were tracked separately from my own status with them (i.e., War, Peace, Non-Aggression Pact, etc.).  As was the case in the EU games, if you go to war with somebody where there's not a proper casus belli, you pay the price in domestic turmoil. 
Armada 2526 has something like this.  It sadly doesn't have any true effect, however (i.e., declaring war on a nation your people like lowers your colonie's collective happiness/loyalty).
[/quote]

If i am not mistaken, it's the case in Star Trek: birth of the federation. No?
Title: Re: A New Model for Diplomacy in 4x Games
Post by: James Sterrett on April 03, 2014, 09:38:00 AM
Quote from: Mr. Bigglesworth on April 02, 2014, 09:34:10 PM
And yet players typically develop some basic rules of thumb, they dont go calculating huge tables of permutations. What scared devs, I think, is how the options multiply. You don't have to go through all options, they key is tree pruning.

Even a chess engine might go searching out 5-10 moves for each likely branch. If you add the human type rules of thumb to a pruned search you get through fast with not a lot of number crunching.

Interesting article on the state of AI for Go: http://www.newyorker.com/online/blogs/elements/2014/03/the-electronic-holy-war.html (http://www.newyorker.com/online/blogs/elements/2014/03/the-electronic-holy-war.html)

Even tree pruning takes number crunching, and the less clear-cut the measurement of success is, the harder it gets.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Mr. Bigglesworth on April 03, 2014, 12:00:30 PM
That does not apply. Chess and go are fairly flat. Chess the center has more value, go the edges have more value. In any strategy game based on a real map there are huge gradients in value for areas with choke points, resources, capital cities, all kinds of things. Combine that with leadership 'personality' Expansionist/ aggressive translated to zero sum game. Cooperative/ alliance builder you can use Shapley value.

You can include balance of power philosophy for some states. Some states only care about getting the most now. Next year is another problem, maybe even another politician's problem. Others think of long term growing trade.

Outlook changes based on trustworthyness. It changes again based on the level of threat to the state.

In simple terms there are all kinds of things that let you prove any possibility from certain branches cannot possibly yield more. Therefore the whole branch is wiped out, no need to search it.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Bletchley_Geek on April 03, 2014, 07:00:16 PM
Quote from: James Sterrett on April 03, 2014, 09:38:00 AM
Quote from: Mr. Bigglesworth on April 02, 2014, 09:34:10 PM
And yet players typically develop some basic rules of thumb, they dont go calculating huge tables of permutations. What scared devs, I think, is how the options multiply. You don't have to go through all options, they key is tree pruning.

Even a chess engine might go searching out 5-10 moves for each likely branch. If you add the human type rules of thumb to a pruned search you get through fast with not a lot of number crunching.

Interesting article on the state of AI for Go: http://www.newyorker.com/online/blogs/elements/2014/03/the-electronic-holy-war.html (http://www.newyorker.com/online/blogs/elements/2014/03/the-electronic-holy-war.html)

Even tree pruning takes number crunching, and the less clear-cut the measurement of success is, the harder it gets.

Not a bad article, but contains a huge lie:

QuoteTo say that Go is more complex than chess, though, is a little like saying that one infinity is larger than another. While technically true—and mathematically possible—it does not fully explain why computers, which can't fully compute chess or Go, have become good at one and not the other. "A few hundred orders of magnitude don't matter when you're up in the ten to the one hundred and twenty," Murray Campbell, a member of the I.B.M. Deep Blue team, told me. The key lies in Go's structure. Deep Blue was able to exploit a weakness in chess's armor: at the grandmaster level, to tell who is winning, you add up the pieces on the board and consider their positions. Campbell explained that, to win, you just stay ahead the whole time, vastly reducing the number of moves for computers to consider.

The truth is that present-day computer chess uses two kind of 'heuristics' to evaluate moves: positional heuristics (basically, an algorithm that assigns a numeric value to a given game configuration) and, more importantly, a cleverly crafted 'pattern database' that maps game states into a vast - but represented in a very compact form - database of chess end games (think about all the chess problems you find in the newspaper, compiled together). What positional heuristics do (or rather, ensembles of such heuristics) is to speed up the search up to the mid game. The reason for this is that, as less pieces are on the board, the values of such positional heuristics become less informed (which means that is harder to tell apart good from bad states). The solution to that was enabled by 'pattern database heuristics', that basically rate states according to how similar they are to game states enabling a known end game (that is, a sequence of moves where the player has victory guaranteed).

This strategy doesn't work for Go for several reasons. First, the game tree is exponentially bigger than that of chess, so weaknesses in positional heuristics (all heuristics, by definition, have a 'blind spot') early in the search can lead you astray very easily. And then, Go "end game" isn't as well documented as that of Chess, so nobody has come out with good pattern database heuristics (and there are well-founded doubts about the technical feasibility of building such pattern databases in the first place).

In order to succeed in Go, Go programs had to abandon the algorithmic framework used in Chess since Claude Shannon's wrote his seminal paper on the subject (presenting the Min-Max algorithm). The problem with Min-Max (and it's evolutions like Alpha-Beta Min-Max or NegaScout) is that is too systematic even with pruning rules. That systemacity - which guarantees optimal play provided enough time to compute the moves - had to be abandoned. And here we come with the Monte Carlo approaches: which are derived from the methods to solve the "n-armed bandit problem" (that is, how to figure out the strategy that maximizes profit when playing a slot machine).

These methods are based on 'rolling out' a base strategy (which in the case of a board game, encodes the kind of positional or lookahead reasoning of chess heuristics), and adjusting this base strategy by sampling possible executions (i.e. the program applies the strategy until reaching a terminal - win or loss - state) and adjusting the expected pay offs accordingly. The 'magic' of these methods lies in being able to account for the variance in payoffs, so the computer program prioritizes those strategies which 1) held the promise of better payoffs and 2) haven't been explored so often. And the level of play of these programs is proportional to the number of strategy evaluations they can make per unit of time.

The fact they play differently as is common amongst humans allows to make an interesting 'meta-physical' question. Games - all games - can be formalized into a mathematical model (which can be more or less 'workable'). We humans play these games, but usually only get a glimpse of all the possible developments of the game. We are taught to play these games in very particular ways: as in Chess, strategies are to a great degree a cultural construct built around the game. The Monte-Carlo go-playing program doesn't leverage this cultural baggage (which is useful in order to play the game effectively) and explores ways of playing which have never crossed our minds. It's a bit like these programs to look for you birthday date into the decimal digits of the number pi - you're guaranteed to find it, infinitely often. But it is also true that you can find there patterns nobody has thought of yet (and assigned to them any meaning).
Title: Re: A New Model for Diplomacy in 4x Games
Post by: FarAway Sooner on April 03, 2014, 10:13:26 PM
Bletchley, in simple terms, then, why is it so easy to find certain exploits that AI will fall for so consistently in most turn-based strategy games?
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Mr. Bigglesworth on April 03, 2014, 11:15:32 PM
In simple terms humans are experts at pattern recognition.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Bletchley_Geek on April 04, 2014, 01:14:12 AM
Quote from: FarAway Sooner on April 03, 2014, 10:13:26 PM
Bletchley, in simple terms, then, why is it so easy to find certain exploits that AI will fall for so consistently in most turn-based strategy games?

Quote from: Mr. Bigglesworth
In simple terms humans are experts at pattern recognition.

What Mr. Biggles said basically: we're very good at figuring out the intentions behind the moves of others. That allows us to pre-empt future opponent moves or to setup a 'trap' to bait them. Purely reactive - that is, they only consider the current state of the game - AIs are just incapable of imitating the kind of reasoning we do when we contrast the moves of our opponents against the state the game was in and where it makes sense for the opponent to take it. Expert human players tend to be very good at masquerading their actual intentions, misleading or distracting their opponents.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: bbmike on April 04, 2014, 06:38:07 AM
Quote from: Mr. Bigglesworth on April 03, 2014, 11:15:32 PM
In simple terms humans are experts at pattern recognition.

You watched Cosmos the other night!  ;D
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Mr. Bigglesworth on April 04, 2014, 07:14:20 AM
I wish, I don't have cable. Internet only.
Title: Re: A New Model for Diplomacy in 4x Games
Post by: Martok on April 06, 2014, 09:14:14 AM
Quote from: Anguille on April 03, 2014, 12:00:07 AM
Quote from: Martok on April 01, 2014, 03:29:04 PM
Quote from: FarAway Sooner on March 30, 2014, 12:06:54 AM
I'd be happy to just see a game where my people's attitudes towards an enemy were tracked separately from my own status with them (i.e., War, Peace, Non-Aggression Pact, etc.).  As was the case in the EU games, if you go to war with somebody where there's not a proper casus belli, you pay the price in domestic turmoil. 
Armada 2526 has something like this.  It sadly doesn't have any true effect, however (i.e., declaring war on a nation your people like lowers your colonie's collective happiness/loyalty).

If i am not mistaken, it's the case in Star Trek: birth of the federation. No?
BOTF only shows another empire's feelings towards you.  Armada does that, *plus* displays your own people's feelings towards that empire as well.