What is the depth/complexity of Civ 5 compared to other games?

[seanflokstra7]

Definitely at least 25 for Civ 5.

 

[Calouste]

IIRC , no one's been able to do a good computer AI to handle GO but there are plenty of very good Chess ones.

As a 4X game, I'd say Civ and other 4X games are nearer to Go in complexity than Chess, as evidenced by the general difficulty in making competent AI in any 4X game without giving AI massive bonuses or hiding the AI's weakness through gameplay limitations

I play Civ5 at immortal and although I haven't played serious competitive chess in a while, I had a rating of around 2100. Chess is a lot more complex than Civ5. Most decisions in Civ5 are straightforward. IMO Civ5 is about 4 or 5 on that scale. Civ5 is more complicated than chess, but that doesn't make it more complex.

 

[dexters]

4X games like Civ are complex by their very nature. There are many layers of complexity while the AI is also playing and acting a part to create fun. Also, as has been noted there are already grandmaster level chess AI's none of the 4x AI can approach it and I don't think it's all because of economic reasons or lack of attention from the devs. It's a legitimate problem. The solution would likely be something like the monte Carlo method as it substitutes Cpu horsepower for lots of rule based behavior

Sent from my Galaxy Nexus using Tapatalk

 

[Funky]

The possibility of creating a good AI for a game has very little to do with its complexity. Strong chess AI's are available not for lack of complexity but because chess is a closed system. That alone says nothing about complexity. Chess is vastly more complex than any computer game, let alone Civ 5.
I know people who have played and studied chess their whole lives with high ambitions. The most succesful of them achieved an FM title (FIDE master) in their mid-30s. That is pretty darn strong. But they are still far behind IMs (international masters). Most IM's are far behind GMs (grandmasters). And the difference between standard GMs and the world's chess elite is in itself huge. It is virtually impossible to become a GM in chess if you don't possess a great amount of talent, even if you study the game 10 hours a day for many decades. Compare that to Civ 5. Moving up in chess takes years, making progress in Civ 5 takes hours.

 

[Vladesch]

some Chess AIs are at "Grandmaster level"

It's really gone beyond that.
All of the popular chess engines far surpass any human player. (grandmaster)
Top human player ELO is 2881 followed by 2830, 2787, 2785, 2783
Computer programs are 3270, 3248, 3238, 3173, 3164, 3159.
There are 32 different programs with higher ratings than the top human player.

I suspect a large part of the reason Go programs are so weak in comparison is lack of people actively working on writing programs for it.

 

[Funky]

There are 32 different programs with higher ratings than the top human player.

That is true. One might add though that there are some chess players who have specialized on computer chess. They study the AI and its algorithms and are able to consistently outwit the best programs by navigating into positions which computers are notoriously bad at. Some of these players have ratings as low as 1900 (which is a pretty good rating in itself, but nowhere near the lowest offical title of an FM, which requires 2300). They use the fact to their advantage that computers don't actually understand chess, and have difficulties i.e. with closed positions with all the pieces still on, which most humans with a general chess knowledge would have no problem finding a plan in. Another factor in comparing humans to computers is that computers don't suffer from fatigue or psychological impediments, which are crucial elements of human chess.
So while it can be said that chess programms have generally exceeded the playing strength of humans, ratings aren't everything. And you need a human to tell you what is actually going on in a given position when it comes to forming a plan.

 

[Pegasus_77]

Credibility of the background info. I think the civilopedia is also a masterpiece, which is teeming with various civilizations who existed in human history and a lot of other elements.If only i had it when i was a little kid, in that sense, I could have spent more time on the history rather than fooling around.

 

[Pegasus_77]

I don't know much about Go, other than that it is also a tremendously complex game. But the atom comparison applies to chess as well. There are 1079 atoms in the universe, while a chess game of average length has 10100 to 10120 possibilities.

I was totally stunned when i saw the possibilities are more than the atoms in the universe.
I do not understand.are you referring to the atomic bombs in the world or the atom in the physics ?

 

[GoStu]

Interesting paper. I like the idea of separating players in distinct groups by >60% win probability. I'm not actually all that surprised to see something like Magic rank that low.

I'm not a Magic player, but one of my former roommates was pretty into it and I picked up some info off him. In terms of distinct brackets, I'd give it "guy who knows how to play", "guy who knows how to play and can build a decent deck", and "guy who knows how to play quite well and builds very nice decks". With the luck element on what you draw, it'd be hard to split hairs finer than that: you can have a very nice deck but get screwed if someone's built their deck counter to yours, or if you just can't get the card you need.

Civ I'd probably place around 4 or so. Despite having eight difficulties, there's probably only about four distinct skill levels of player; the difficulty setting will determine simply how easy a time they have. A player that can beat Prince/King reliably can probably do well in Emperor, it'll just be less of a cakewalk.

It's making me wonder about my other strategy games that I love so much. I wonder what the number would be on Starcraft, or Achron?

 

[Funky]

I was totally stunned when i saw the possibilities are more than the atoms in the universe.
I do not understand.are you referring to the atomic bombs in the world or the atom in the physics ?

I mean the atoms in the universe. I'm glad there aren't 1079 atomic bombs in the world!
I agree it sounds crazy at first, but the exponential increase of the number of possibilities in chess is very rapid.

 

[vorlon_mi]

I'm not sure that combinatorics and mathematics are the best way to compare the games. Consider a duel sized map for Civ5. More tiles than a chessboard, but many fewer starting units initially. As each player chooses which units to build, and when, and where to move them, I have to believe that the combinatorics of Civ5 outstrip chess.
Now, consider a standard sized Civ5 map, with 8 tribes and 16 CityStates. An order of magnitude more hexes than chess, more starting units, and the decision trees about which units are built, and when, just start to explode. I don't think the math is an accurate assessment of the depth of skill required to play.
Chess and Go include both strategy and tactics, but there are no probabilistic elements. A pawn can take a queen *every time*, while the classic "spearman vs. tank" battle has been a source of frustration since Civ2. The decision-making in Civ5 must include likelihoods and probabilities in tactics; the decision-making in chess includes likelihoods and tendencies on the strategy level... what is this human likely to do? What has he/she done in the past?
The map characteristics change from game to game in Civ5, whereas the board is *always* the same in chess. White always has the same set of first moves available every game, whether newbie or grandmaster; this is not the case on the first move in Civ5

 

[GoStu]

I don't buy much into the "huge numbers of possible positions" argument. I don't buy it very much for Chess, and I certainly don't buy it in Civ.

In Chess, there's certainly a massive number of unique positions, but most of them will never or just can't arise during gameplay unless something goes really weird/wrong/special. Chess has fairly established openings, endgame strategies, and everything in between:
http://en.wikipedia.org/wiki/Chess_opening
http://en.wikipedia.org/wiki/Chess_endgame

As to Civilization, while there's more possible positions, they're certainly less distinct. For the most part the huge number of tiles and the randomness of setup ensures a relatively homogeneous board. Starting terrain is notable only as being well above or well below average, and a few variations based on specific resources. Examples of these might be:
- How to play out of a flat tundra start
- How to maximize a good desert position
- What to do if you have no iron, and there's none anywhere in sight (my current game)
- What to do if you've got the Zulus for neighbors.
- What should I do with all this Wine and Incense?

The precise locations of where exactly the hills are or if you've got an extra two horses matter very little. Similarly, the exact positioning of a unit isn't a major factor in a larger war. We could go into detailed and torturous examples about that one time there was a hill just in the wrong spot so your trebuchet couldn't get line of sight, but in the broad picture there was still a way around.

 

[monad]

I think some people here overestimate the difficulty of Civ and confuse the complexity of a game with its depth. While Civ is indeed very complex (in terms of features, layers, set-up, micro and macro resource management), the optimal strategies are often straightforward and very similar each game. Furthermore, the opponent's actions can be neglected most of the time and a passive victory can be achieved. The game lacks a strong direct confrontation between the players, involving an in-depth analysis to find the optimal winning strategy.

This is the opposite of games like chess and go, where a 2h game is a continuous mental analysis and computation, knowing that a non-ideal move can lead to a crippling disadvantage later in the game. Civ has too much randomness (starting location usually decides the game between similar skilled players), the optimal path is not too difficult to find and there is a lot of margin for mistake, compared to chess and go. Regarding the table in the OP, Civ is at most 11-12.

 

[Naokaukodem]

the 2010 version of civ 5 was far deeper than the version that exists today

the tradeoffs involved much more strategic decision-making and there was a much wider range of viable ways to grow your empire

since release, firaxis has systematically removed depth+complexity from the game while adding miscellaneous mechanics that add breadth and make the game less balanced

ultimately, civ 5 is broken because the challenge is about exploiting an idiotic AI and essentially finding weaknesses/patterns in the code. the decision-making at this point has very little to do with the game mechanics / game design

the game actually isn't that bad from a design standpoint. it's just that the depth only exists in the theoretical realm of human vs human (which is basically non-existent because of the broken multiplayer implementation)

That post sums up pretty much all my thought about Civ5. However, I'm not qualified to compare vanilla to expansions as I'm sticking with vanilla. I'm now playing at Immortal and find great challenge in it.

However, I think that Firaxis could have made easier our way to greater difficulty levels. Now the majority of players rely on youtube let's plays in order to do so, or on strategy wrotes in this forum. While it can please the player who have plenty time to lose, I think that an average player should be able to beat the highest DL on his own. Most players beating Deity have reverse-engineered the game in order to see how the AI would behave in which situation, and take profit from it. Unless they outtested extensively how the AI was behaving in a number of situations, like in bargains, (I learned that a lux could be traded for 240 gold in normal speed in a let's play, and now i always use it, but strangely I never had the idea to see how much the AI was ready to give us apart from what it proposed already) but i'm not believing it too much.

One thing is sure : the best players use holes in the AI behavior in order to take profit from it. For example, the AIs play FFA with forbidden alliances (or at least not designed until a victory, just more circumstantial), with no consideration of the global geo-politics, while you have a whole set of alliance possibilities on your own. Beating the AI is like exploiting its weaknesses on purpose, not really play at a game with set rules like it could have been on multiplayer if only multiplayer would have been a good experience. (it's not, now the rare games with no DLCs or expansions you are kicked because they are reloads, the multiplayer lobby is huge crap, not to mention the hosting freeze and other disconnections)

 

[Moriarte]

The possibility of creating a good AI for a game has very little to do with its complexity. Strong chess AI's are available not for lack of complexity but because chess is a closed system. That alone says nothing about complexity. Chess is vastly more complex than any computer game, let alone Civ 5.
I know people who have played and studied chess their whole lives with high ambitions. The most succesful of them achieved an FM title (FIDE master) in their mid-30s. That is pretty darn strong. But they are still far behind IMs (international masters). Most IM's are far behind GMs (grandmasters). And the difference between standard GMs and the world's chess elite is in itself huge. It is virtually impossible to become a GM in chess if you don't possess a great amount of talent, even if you study the game 10 hours a day for many decades. Compare that to Civ 5. Moving up in chess takes years, making progress in Civ 5 takes hours.

So, you compare human vs. human in chess to human vs. AI in CIV? Isn't the comparison a bit flawed? It will take you same decades to rise through the ranks to be a grand master of CIV, less, if you have natural talent. I don't see how it is different to chess, if you make the only true comparison for both games: human vs. human.

It also seems to me that it is infinitely easier to code a grand master level AI in chess than to do the same for civ. The reason for that is you have to code a grand master general, a grand master economist and so on. (for civ)

 

[Babri]

the 2010 version of civ 5 was far deeper than the version that exists today

the tradeoffs involved much more strategic decision-making and there was a much wider range of viable ways to grow your empire

since release, firaxis has systematically removed depth+complexity from the game while adding miscellaneous mechanics that add breadth and make the game less balanced

ultimately, civ 5 is broken because the challenge is about exploiting an idiotic AI and essentially finding weaknesses/patterns in the code. the decision-making at this point has very little to do with the game mechanics / game design

the game actually isn't that bad from a design standpoint. it's just that the depth only exists in the theoretical realm of human vs human (which is basically non-existent because of the broken multiplayer implementation)

This is a joke, right? Not funny at all.

 

[Naokaukodem]

This is a joke, right? Not funny at all.

I don't think it is to take literally as to bold the sentence. I think it is just to make hear a voice.