Soren Johnson's Blog

[Psyringe]

I missed this earlier. No, it is not an exaggeration, it happens sometimes. Given that you play games more slowly and have probably tallied FAR fewer immortal/deity games, maybe you've not seen it. You get situations where Sury expands to 15 cities in the same timeframe that pericles expands to 3 cities when each has the land available. Then surry peacevassals pericles, kills another civ off, cap vassals a 3rd civ before 1200 AD and when you finally meet him on the other continent, he has 40+% of the world's land and pop and is closing in on infantry. Good luck with that.

I actually haven't reached immortal/deity yet. I hope that it takes a while until I get there because I expect immortal/deity games to be less enjoyable than the emperor games I'm currently playing. (A pet theory of mine is that games are more enjoyable when you're learning to master them than after you actually have mastered them; as a result, I tend to stay away from strategy guides and detailed analyses, because I actually enjoy the game longer when I have to develop my strategies myself). But I digress, that's another discussion.

We were discussing whether "games decided at turn 0" was an exaggeration. From the games I played, I'm certain that I couldn't predict how they play out. One reason for this may be that you analyzed the game far more than I did (as I said in another thread, imho you analyzed it to a point that your knowledge about it became detrimental for your enjoyment). Another reason may be that I play on super-huge maps with 30+ civs, and never use settings like "fair distribution of resources" (I actually _want_ them unfairly distributed, I want some AIs to have better starting positions than others, so that some AIs can become strong enough to challenge me in the late game). I've experimented with games with fewer strategic resources; I guess limited availability of iron or oil on the map could outweigh AI predispositions in determining which of them grows strong.

I can only say that I don't see the predictable pattern that you see. I once had Gandhi as a runaway winner conquering/vassalizing two continents (to bring one example, I can come up with others), so I don't see the consistent "suiciding" behavior that you claim to be inherent in certain AIs.

One part of your argument strikes me as odd though. You complain that the Civ4 AIs lead to predictable gameplay, yet you _want_ (and favor) predictable gameplay by excluding luck, strong random events, etc. as much as possible. Perhaps I'm misunderstanding something, but this doesn't seem consistent to me. Personally, I _want_ unpredictable gameplay and I set my games up in a way that I tend to get it, and I have fun playing them. It may be a niche playstyle (as is yours, I guess), but it's consistent and I'm enjoying it.

 

[TheMeInTeam]

I can only say that I don't see the predictable pattern that you see. I once had Gandhi as a runaway winner conquering/vassalizing two continents (to bring one example, I can come up with others), so I don't see the consistent "suiciding" behavior that you claim to be inherent in certain AIs.

Starts and favorable positioning relative to other suicide AIs can overcome anything. Give a high level gandhi 20%+ of the world's land to work with and his 15 unitprob will do him just fine (since he has over double the cities of some AIs)

One part of your argument strikes me as odd though. You complain that the Civ4 AIs lead to predictable gameplay, yet you _want_ (and favor) predictable gameplay by excluding luck, strong random events, etc. as much as possible. Perhaps I'm misunderstanding something, but this doesn't seem consistent to me. Personally, I _want_ unpredictable gameplay and I set my games up in a way that I tend to get, and I have fun playing them. It may be a niche playstyle (as is yours, I guess), but it's consistent and I'm enjoying it.

You are confusing lack of luck-based outcomes with predictability. You don't have to have imbalanced trash in a game for your opposition to bring you unpredictable situations.

 

[Psyringe]

You don't have to have imbalanced trash in a game for your opposition to bring you unpredictable situations.

I know, it just seems to me that your visions of ideal gameplay and an ideal AI don't match. But maybe I'm misunderstanding you, that's why I ask.

It seems to me that you favor games with:

- balanced starting positions (no "imbalanced trash", as you labeled it)
- "good" AI in the sense Soren described it, i.e. an AI that doesn't try to roleplay a character at the cost of foregoing good strategies for its position

This favors a setup with equally strong AIs, each working from an equally strong basis. From my experience with other games, this usually leads to situations where the AIs are deadlocked with each other (i.e., no AI is so much stronger than another one that it can actually overcome it), while the human player is smart enough to do so. Once the human player has conquered on or two AIs, he has a stronger position than anyone else in the game and will usually win. That's why I'm favoring "imbalanced crap" in my games - it helps creating two or three AIs that grow strong enough to challenge me even in the late game, even if I conquered five or six others.

So I'm curious, how would your ideal game and ideal AI be designed, and how would you prevent the balance trap I described above?

 

[TheMeInTeam]

I know, it just seems to me that your visions of ideal gameplay and an ideal AI don't match. But maybe I'm misunderstanding you, that's why I ask.

It seems to me that you favor games with:

- balanced starting positions (no "imbalanced trash", as you labeled it)
- "good" AI in the sense Soren described it, i.e. an AI that doesn't try to roleplay a character at the cost of foregoing good strategies for its position

This favors a setup with equally strong AIs, each working from an equally strong basis. From my experience with other games, this usually leads to situations where the AIs are deadlocked with each other (i.e., no AI is so much stronger than another one that it can actually overcome it), while the human player is smart enough to do so. Once the human player has conquered on or two AIs, he has a stronger position than anyone else in the game and will usually win. That's why I'm favoring "imbalanced crap" in my games - it helps creating two or three AIs that grow strong enough to challenge me even in the late game, even if I conquered five or six others.

So I'm curious, how would your ideal game and ideal AI be designed, and how would you prevent the balance trap I described above?

Similar to how you describe, but with enough variability in AI strategy that they can emerge over each other. This can be done by making multiple approaches viable, but some weak to others, and then having the AI gamble, or through the actual game mechanics themselves.

Games where AIs can back stab each other rarely end in deadlocks from what I've seen, though, unless defense is just too strong relative to offense.

 

[lschnarch]

Guys, you've lost me.

Are you still talking about Civ4? Because the description of "runaway" civs, occupying whole continents extremely fits my experiences with Civ5.
And the remark of all Civs playing the same style seems to be quite Civ5, too.

So, what game are you talking about?

 

[r_rolo1]

I think TMIT is talking of both civ IV and V

 

[Psyringe]

So, what game are you talking about?

We started with Civ4 (and a bit of Civ5), but migrated to "ideal" games and AI design. Which, come to think of it, isn't even off-topic considering that Soren doesn't limit his AI typology to Civ4 either.

 

[Brawndo]

2. "The units are the most important thing on the map, not the terrain, the terrain shouldn't 'hug' the player's attention". Also, interesting to hear that they scrapped a whole system of realistic looking terrain in the middle of development because the players couldn't easily see which tile was what and therefore had less fun. As Soren says:



Since I had a _lot_ of problems with both things in Civ5 (terrain tiles bleeding into each other, and units being hardly recognizable on the terrain), I wish they had continued to follow this rule instead of concentrating on "organic looks". This may be a preference issue, but Civ4 clearly matched mine better in this case.

Couldn't agree more. I would go back to colorful graphics and 2-3 figure units in a heartbeat if we could, because the Civ 5 yellow-green map is completely cluttered to the point that I feel I am mostly playing with icons

 

[aimlessgun]

I'd be willing to place a bet you didn't run the game on immortal+.

Two (reasonably common) outcomes from the asserted "fun" AI in civ IV:

- Human spawns next to creampuffs or in a world filled with them. Using diplomacy, he eliminates any possible chance of any AI declaring on him anywhere, and cruises to his choice of culture, diplo, or tech + conquer type victories (Easy games, sometimes even easy on deity, sort of)

- Human spawns next to a somewhat-unreasonable AI or an idiot like toku. "fun" AI with personalities on the other continent: Gandhi, Isabella blocked into minimal land, Sury, and washington. When you meet Sury around 1100-1200 AD, he is at war with and finishing off washington, with gandhi dead or capvassal'd and izzy as a peacevassal. Sury, before war, had 20% of the world's land (in a 7 civ game!), but with all of his peace/war vassals and conquests, now has over 40%. You have...maybe toku's land, hard fought, or a simple 14% TOPS, but sometimes 10% or less if the RNG feels fresh.

Playing on exactly the same difficulty with exactly the same ability on exactly the same map script, you can run roughshod over scenario 1 and win easily, and have just about 0 chance in game 2. The problem with "fun" AIs is that when you implement them, SOME of them try, and the rest don't. That quickly results in stupid runaway situations where several AIs essentially GIVE land free of charge (via underexpansion or peacevassaling senselessly) to a super AI...meaning these "fun" AI are now effectively dogpiling you...based on a random chance factor of who-spawns-where turn 0.

Some "fun". V's AI struggles mostly because 1) diplomacy is opaque and convoluted and 2) while it tries to win, not all of its actions are yet consistent with that goal. Civ V's AI is actually MARKEDLY better than stock IV's AI, and that's by design. It mainly suffers from non-dynamic strategic choices in V (war and/or huge empires are pretty much the only strong tactic) and the fact that 1upt is a lot harder tactically than "put 150 cavalry onto a tile and march at each other or humans' cities).

From your 2 scenarios, scenario 1 is alive and well in Civ5. In scenario 2 in Civ5 the military AI is so bad it doesn't matter

Leading to Civ5 suffering from non-dynamic strategic choices: war is the best option because the AI is awful at it. The horrible military AI overshadows every balance discussion because the game might be completely different if they got it to play well (and that might be impossible). You can play a completely peaceful game and crank out a diplo win relatively quickly (maybe not sub-turn 200, though I think it should be possible): maybe if the combat AI was improved a notch diplomacy and UN victory become the best strategy.

Random note: the programming of the UN victory shows that they still have the Civ4 mentality of the AI playing for 'fun' and not to win. Right now the AI only throws weak challenges at an imminent diplomacy victory (they might buy off a few city states using only a fraction of their vast gold reserves). If the deity AI 'played to win' on the peaceful options right now it would be a beast (except it can't defend itself).

Anyways not really sure what my points are here...what are we talking about in this thread?

PS: very interesting talk, thanks for the link OP

 

[Lightzy]

I think the most important part in his presentation was that Civ is a game about TILES.

CIV5 is most most most definitely not about tiles.
They're all the same, they all suck, and they all need trade posts on them for correct use.

 

[TheMeInTeam]

From your 2 scenarios, scenario 1 is alive and well in Civ5. In scenario 2 in Civ5 the military AI is so bad it doesn't matter

Leading to Civ5 suffering from non-dynamic strategic choices: war is the best option because the AI is awful at it. The horrible military AI overshadows every balance discussion because the game might be completely different if they got it to play well (and that might be impossible). You can play a completely peaceful game and crank out a diplo win relatively quickly (maybe not sub-turn 200, though I think it should be possible): maybe if the combat AI was improved a notch diplomacy and UN victory become the best strategy.

Random note: the programming of the UN victory shows that they still have the Civ4 mentality of the AI playing for 'fun' and not to win. Right now the AI only throws weak challenges at an imminent diplomacy victory (they might buy off a few city states using only a fraction of their vast gold reserves). If the deity AI 'played to win' on the peaceful options right now it would be a beast (except it can't defend itself).

Anyways not really sure what my points are here...what are we talking about in this thread?

PS: very interesting talk, thanks for the link OP

Play MP with a mind for anything except military, and see how well you do on average.

While the bad war AI certainly contributes to the strategic imbalance, it is DEFINITELY NOT the only factor at play, and it wasn't in IV either.

And yes, runaway situations due to situationally (or just arbitrary) AI play are present and a problem in both games. A runaway AI should emerge via good play, not because its neighbors were such laughingstocks that they essentially gave their land away (and sometimes they literally do that in both games).

And civ V IS a game about tiles. It does not have the ludicrously strong special resource tiles relative to IV, but there is still disparity between terrain types - hammers are scarce and several resource tiles offer attractive yields over base values, especially early on. Next time you play, pull up a mid-game empire and see how much of your empire's total output is coming from worked tiles vs something else. Unless you are abusing maritime (which definitely needs to be nerfed), I'll wager the vast majority of your output is STILL being supported by tiles.

Now, one could certainly argue that the yields are underwhelming and that such affects the pacing of the game. That might even be a valid argument (although progression through the tech tree, large military, etc are certainly possible). However, it is NOT valid to claim this isn't a tile based game. Even WITH maritime, total yield is often from worked tiles more than other sources. With a nerfed maritime, where else are you getting anything? Pop 1 merchan navy, or order SPs? How are you getting there? Probably growing for pop/gold/science multiplier production, all of which lean on tiles.

 

[aimlessgun]

Play MP with a mind for anything except military, and see how well you do on average.

First off, I've never played Civ multiplayer, so my comments on it now are just going to be speculation

Are you saying the multiplayer balance was better in Civ4?

Is is not possible to be defensive and go for another win condition? I guess with units as ridiculous as horses defense would be at a disadvantage, though if you made it to artillery it would flip for a while.

I should probably try out an MP game eh.

 

[TheMeInTeam]

First off, I've never played Civ multiplayer, so my comments on it now are just going to be speculation

Are you saying the multiplayer balance was better in Civ4?

Is is not possible to be defensive and go for another win condition? I guess with units as ridiculous as horses defense would be at a disadvantage, though if you made it to artillery it would flip for a while.

I should probably try out an MP game eh.

No, civ IV was not any better. Everyone I played against on game spy...even bad players...would know to go military because thats #1 priority in every civ game. At first, collateral initiative favored the defender, but once you had things like naval stacks forking your cities, air power, or involved a tech lead, things could rapidly get out of hand.

Admittedly, my V MP experience is limited, because UI controls lag, the turn timer is RIDICULOUSLY SLOW, and waiting for other people to play turns is like watching paint dry for me (I can seriously play 2 games at once at that speed without trouble). However, military survival is still #1 priority and I don't see how someone staying small to max out culture SP is going to stay alive against sheer production + tech of a larger empire, nor do I see a game actually finishing in a space win before someone interrupts it (doubtful a game would last even that long - nukes are available sooner after all).

The reality of this TBS is that military survival has dominated and still dominates, to the point that it is the strongest tactic. Some prefer it this way, but the game was never balanced such that other VCs are as strong as military VCs. If you don't believe me, look at civ IV's HoF times. If you're not cheesing AP, military wins are always the fastest X_X.

 

[Grimberht]

Please explain how people that enjoy the game are not "normal." Are you using yourself as the barometer of "normalcy?" Are the people that like the game better than "normal?" Are they worse than "normal?"

Feel free to elaborate rather than just tossing a baited hook into the water, hoping to catch a fish.

It's one thing to dislike the game, or even to insult the game. But to insult people that have nothing to do with the game itself other than they find joy in it. Well, that really just tells everyone more about you than it does about them.

If Civ5 alienates "normal" people like you, even if I did not like it, I would consider buying it just out of principal with hopes that it would keep you "normal" people away.

I'm totally sorry about the impression my post made... might be language related... By "normal" I did not intend to imply that the other was "less" than normal in any way.

I guess it would be more accurate to talk about "builders" which I think is the most frequent kind of players who got hooked by this franchise.

I spend many more hours trying differents way to build up my empire before focusing on different aspect of the decision making, which include in some case almost no military play... now the main decision you have to make are military related. Every other way of experiencing the game now seem totally toss aside and design in some way too "slow" the process of winning the game. I guess what I'm saying the game aint about building a civilization anymore, but about imposing my empire on the others.

I for one, has many others stated all over this forum, never really played this franchise has a pure war game and my post was simply stating that we builder arent the target audience of this new game design.

 

[aimlessgun]

I for one, has many others stated all over this forum, never really played this franchise has a pure war game and my post was simply stating that we builder arent the target audience of this new game design.

Despite this, you can still play this game as a pure builder and not get involved in any wars. It takes a little more work maybe than in previous civs, but you can do it. And the peaceful options are not *that* slow. Obviously they go faster if you make war early to get science, but I wasn't kidding in saying you could probably get a diplomacy victory before turn 200 without having a single war, considering I was able to get a diplomacy win on friggin' OCC on turn 250 without a single war.

 

[Jharii]

I'm totally sorry about the impression my post made... might be language related... By "normal" I did not intend to imply that the other was "less" than normal in any way.

Everything's okay.

There are just so many people taking those kind of pot shots at other people on this forum. Thanks for the clarification.

 

[ohioastronomy]

Despite this, you can still play this game as a pure builder and not get involved in any wars. It takes a little more work maybe than in previous civs, but you can do it. And the peaceful options are not *that* slow. Obviously they go faster if you make war early to get science, but I wasn't kidding in saying you could probably get a diplomacy victory before turn 200 without having a single war, considering I was able to get a diplomacy win on friggin' OCC on turn 250 without a single war.

I think that it's a fair claim, however, that this is the least friendly Civ game for sandbox players. If you look at it that way, the single biggest blunder in Civ 5 was to make the AI players uniformly aggressive. This turns it into a different game than the one that folks like me enjoy playing (beating our prior records, trying out different ways of getting to the finish line.)

And it's one of the worst war games that I can remember in a long, long time. I could only win at Civ 4 on Deity with gimmicks. I can win Civ 5 on Immortal trivially; I have won Civ 5 on Deity, but the game is so tedious with the flawed combat system and endless units that I can't imagine anyone enjoying that level. I certainly don't.

 

[King_Course]

Psyringe, that was exactly what I was thinking while watching that video

And you are not alone with the terrain identification issues, that's clearly a case of where "shiny looks" got preference over good usability.

I utterly disagree on that statement. I think Civ V does an excellent job graphic-wise. Both on ease of eye, ease of delivering information, and aesthetics overall.

Just look at the screenshots they are excellent, tell you everything too I reckon.

The artist didn't describe it as "watercolors" as some here did, but called it an art deco approach. He didnt want the game to be darker and more gloomy as timed passed and you were winning the game. I think he had a good point there, with whole eco-debate raging and playing a map left void of forrests and full of industry.

 

[Mad Hab]

"The units are the most important thing on the map, not the terrain, the terrain shouldn't 'hug' the player's attention". Also, interesting to hear that they scrapped a whole system of realistic looking terrain in the middle of development because the players couldn't easily see which tile was what and therefore had less fun.

That's a thing the original Civ, that some consider 'ugly', did better than the 'more beautiful' Colonization. In Civ you could tell the terrains apart very easily, while in Colonization the forests looked very look-alike...

Cheers,

Mad Hab