Recraft AI so military victories are not required at higher levels

[Chris Withers]

I strongly believe that the AI needs to be modified at the higher levels (Emperor and above). The only/predominant way to keep up & win at these levels is to attack early & often. I'm all for combat, but to me the game is more fun at the lower AI levels where I can do a combination of military builds & infrastructure builds in the first two ages.

The problem seems to me to be very simple to solve. The AI levels the playing field at the higher levels mostly by building stuff faster. It builds wonders & etc faster so there is no chance for a cultural victory. Even worse, it gets way ahead in tech so one is forced to attack quick and often. Once one has beaten down rivals over a period of time one could choose to then go for a spaceship or cultural victory, but this does not change the fact that it is essentially a military victory.

The underlying problem is that the AI does not operate its workers & military units as efficiently as a good player can. There are many ways the s/w can be improved here, but the AI will never be able to do as well as the human player (see below).

So why not change it so only workers (and settlers) and military units are harder to build for the human player? If we have fewer workers operating efficiently then that equates to the AI having more workers operating more efficiently. If we have fewer military units because they are harder to build then we won't go to war as much. There will then be more incentive to build infrastructure. With this change, the s/w can change so we don't fall so far behind in tech. Keep tech growth equal for the AI & the human player, which is then needed of course since we won't have as many military units to go force it away from our rivals. For example, as one goes up each AI level military units might cost X% more each level. Settlers would need to cost more at each level since the human player is better at picking out the best sites to settle, in the exact best hex, and in a sequence that makes sense for the total map situation.

The reason the AI will never do as well militarily as a good human player is the human player thinks in terms of campaigns and strategy. The AI looks at the immediate picture. So the human player is more omnipotent than however much programming time can be afforded to put into the AI. To see the whole picture and conduct a proper military offensive is easy for us, but is difficult to program.

So the human player will put together a nice stack, and probably a combined arms one at that (defensive val units, lots of offensive fellers, some artillery, etc). After critical mass is attained we will make a logical sequence of moves & battles to win the campaign (take some cities & get great peace terms). Exactly how the human player does this will vary depending on the total board situation. In contrast the AI sort of wings it. It throws units at one, making poor tactical and strategic decisions.

Back to the main theme in case I didn't state it well. This whole "hit 'em early and hit 'em often" required strategy at the high AI levels is very disappointing to me. That sort of game should be called "World Conquest" (even if one uses conquest to build a spaceship first ...). I'd rather play "Civilization" where there must be more depth to a victory than just whalloping your rivals on the battlefield. I've no problem with going to war, but to not also have to build up the civilization infrastructure to support it is ahistorical. A simple analogy is the Roman society could support the large and good army because their cities had the aqueducts & sewer systems and farming to create and sustain that army. I had more fun while learning and doing a combo of building city improvements and doing military campaigns. Those main articles in The War Academy (very well written btw) on how to play at the high levels somewhat prove my point. "Get over it" about not building early wonders, and "only build workers/settlers/military" in the early going from those articles shows exactly what's wrong with Civ3 at the higher levels. I haven't "gotten over it"; I do think we should be able to build some of this stuff. I have plenty of war-games I’d rather play if that’s all Civ is going to be at the high levels. I guess I could regress some levels and play that way, but then I'd always win ...

 

[DoTheMath]

I agree with your idea, but I'm concerned stacked combat increases the human advantage.

In the 20th Century, there are certainly enough diplomatic alternatives, MPPs and "Cold War" alternatives. CIA & KGB installing the leaders of their choice, and then convincing the people in their sphere of influence that it was "the right thing to do", rather then simply what their leaders decided to do, whether their motives were good or bad.

 

[oldStatesman]

I had more fun while learning and doing a combo of building city improvements and doing military campaigns. Those main articles in The War Academy (very well written btw) on how to play at the high levels somewhat prove my point. "Get over it" about not building early wonders, and "only build workers/settlers/military" in the early going from those articles shows exactly what's wrong with Civ3 at the higher levels. I haven't "gotten over it"; I do think we should be able to build some of this stuff. I have plenty of war-games I’d rather play if that’s all Civ is going to be at the high levels. I guess I could regress some levels and play that way, but then I'd always win ...

Excellent point...compliments from a newby just learning the game. At first I was all fired up to learn to try and tackle Deity and Sid...but now I am really disillushioned after learning that the only way to beat these levels is micromanagment galore and ALL WAR. I LIKE to build things. I have fun warring sometimes it, don't get me wrong it has it's place in the game; but as you said if it is going to be an all war game I have several of those that are quite detailed and much much much more realistic than Civ ever will be...I put those down ages ago, as they got boring, ...I was hoping Civ was different, and at the lower levels it is, but the AI is too easy, too predictable, too, well, stupid.... right now I am still keeping great interest because of things like learning new Civs and building the different wonders and seeing how differnet UU's act... but once that pales and if the higher levels are all about War with no alternative strategy then I guess I will lay this one down too...

 

[Sirian]

Deity is beatable peacefully.

Here's one example.


- Sirian

 

[Doc Tsiolkovski]

Let's hope the same is true for CIV .
After all, peaceful high level Civ3 victories are about manipulating the AI - and we all hope the CIV AI isn't as easy to trick.

 

[oldStatesman]

Deity is beatable peacefully.

Here's one example.


- Sirian

Thanks!

Still, something so far has made me feel lukewarm about a UN victory...I have had 2 and it is just so anti-climactic. Feels more like winning by a throw of the dice after playing for scores of hours than a real earned victory.

But great example of how to trade...maybe some of the problem with higher levels is the Victory conditions themselves...they along with the AI advantages seem to almost force a domination or conguest victory from what I have observed in other games; here's hoping they will add some dimension to the other types in Civ4.

Thanks again; I am glad to see not everyone here espouses war first last and always!

 

[Chris Withers]

Sirian, that's a great game and I'm going to follow it in more detail when I get a chance. Thanks for taking the time to put all that stuff you have together.

However, I feel it just proves my point about how the AI is crafted wrong. You are way behind in tech, culture forget it - they are getting all the wonders. You do not have a normal civilization compared to the others. I should modify my opinion in that one can win diplomatically instead of just militarily, but I do not call how these high level victories are achieved "Civilization". The AI should be reconstructed so there are multiple avenues to victory and where one's civilization can follow a normal path throughout all the ages.

"OldStatesman's" comments above are right on IMHO.

 

[Sirian]

I wrote out my thoughts on the AI in some of the AI threads that were hot in this forum last summer. If you can find them, you can see what some of the other community members have already said on the topic.

- Sirian

 

[Chris Withers]

Ah, I found it. It is an excellent thread, all interested should read it (and in its entirety). Here it is:

http://forums.civfanatics.com/showthread.php?t=93015

It seems like we agree with the basic tenet in that as one moves up levels the game becomes less total-Civ like and one gets "robbed" in many of the ages where one has an aberrant/atypical civilization. What we want at the higher levels of Emperor and above is to be able to play all the ages, on par with the AI civs (and vice-versa), and have civs that are able to approximate theirs in all aspects (culture, military, tech, etc).

I think your idea has great merits. It may need to be combined with what I am saying above and/or with other ideas. My thiought is to make military units especially (but also workers) cost more for the human than for the AI. This is because we can handle these fellers much better than they will ever be able to program.

I wonder if the Civ4 programmers are taking any of this up, anyone know? I sure hope so, they sure need to do something radical in the AI. They want a enduring "franchise" as that presentation says. A franchise needs a good "buzz" from the high end players, the ones active in sites like this for example. This "buzz" works its way into many of the other people that may or do buy the game. If we are not in total enjoyment of our Civ experience then that has to hurt sales at some point ...

 

[Sirian]

I wonder if the Civ4 programmers are taking any of this up, anyone know?

Firaxis has been stingy with info about the game thus far. E3 is coming up, though, and some gamer mags are supposed to be previewing as well.

Soren Johnson indicated that feedback from fans about Civ3 had a major impact on a lot of his Civ4 design choices. For instance, getting rid of "whack-a-mole" pollution and replacing with something else.

"One third old, one third improved, one third new." Or something like that.

Well, if the third that they ditched to make room for the new bits is well and truly the weakest third of the old game, then Civ4 may hold quite a bit of promise.


When I interacted with Soren over patch issues for Civ3 and expansions, he listened a lot more than he talked. He just doesn't say much. But he ran the patching for Civ3 and he stuck with it until the vanilla game got polished up pretty nicely. I really appreciated that. He didn't just listen to fans, he actually understood our feedback. Patches for most games tend not to make significant gameplay improvements, but only to smooth over the bugs, if that, but Civ3 was different.

We're not going to know what Civ4 holds until we play it, of course, but the devs seem to be in touch with the fans in a healthy way.


- Sirian

 

[opus95]

Right you are. I agree completely. More emphasis should be put on developing an advanced civilization ... otherwise the only way that England can win would be to emulate a Mongolian-styled campaign of conquest and destruction.

In real life, England did it (creating a world-wide empire) not by sacking and destroying advanced civilizations, but by converting some by superior ability and science, and in other places by settlement in areas of obviously inferior cultures. And, of course exeptionally effective diplomacy, which probably isn't feasible in the game as it is.

 

[DoTheMath]

To create more of a builder's strategy, you might want to cut the cost of buildings, especially libraries. (Although conquest will still get the player more access to shield production).