AI victory strategy

[tebriz]

Does AI really has any long term victory strategy? It seems for me that it doesn't...
For example when I test a game (with AI autoplay) with only conquest or domination victory condition enabled - there is quite a lot of situations when one super power emerges, but all other civilizations don't unite against it as it would be wise for them. I see many games when one civilization has almost half of the entire world, but other civilizations still fight among themselves while that big civilization continues its expansion.
I think there must be something to fix it (especially in games with only conquest or/and domination victory conditions) - whenever a super power emerges, all other powers must be aware that if they don't stop wars among themselves and unite against big enemy, he will conquer entire world and win the game.

 

[Dancing Hoskuld]

Our expert on AI is not available until next year. Not sure that anyone else can answer your question.

 

[Thunderbrd]

They're probably right now programmed more towards thinking for their own benefit they can take down weaker opponents and they are acting with a reasonable sense of intimidation when considering the larger enemy. I can't say that for sure but you do make good points and overview level strategies such as this have not gotten all that much attention since the better ai project that took place a long time before this mod. To me, this is something to work on once most of the mod elements are otherwise in place as so many new projects to come would force constant reworking of large scale strategies.

 

[tebriz]

That is the issue - AI is programmed more toward thinking for their own benefit. But this seems to be a short term benefit, not long.
I have tested lots of games with many starting civilizations (using AI autoplay)- and all the games end around Industrial era, with one single civilization conquering more than half of the entire world. I have even tried to turn on "NoVassalStates" option in order to prevent this, but still not a big difference.
AI definitely has to learn something about "balance of power" here
I think this is quite an important issue, much more important than many other projects which are being worked on now. This issue needs more attention.

 

[Dancing Hoskuld]

Unfortunately it will probably not get the attention you think necessary until another AI programmer joins the team.

 

[tebriz]

Maybe till then we can make something very simple - tweak it that if one civ gets big, it starts to get negative attitude from all other civs... bigger it gets, more negative attitude it receives. Is it hard to implement?

 

[Dancing Hoskuld]

No idea.

 

[Yudishtira]

That is the issue - AI is programmed more toward thinking for their own benefit. But this seems to be a short term benefit, not long.
I have tested lots of games with many starting civilizations (using AI autoplay)- and all the games end around Industrial era, with one single civilization conquering more than half of the entire world. I have even tried to turn on "NoVassalStates" option in order to prevent this, but still not a big difference.
AI definitely has to learn something about "balance of power" here
I think this is quite an important issue, much more important than many other projects which are being worked on now. This issue needs more attention.

To me this is a big plus. I don't want AIs to act like their playing a game and exploiting every loophole and gaming every exploit. I want them to act like they are civs in the game world - which after all is what they are.

The Mediterranean didn't gang up on Rome (so as to not lose the game) - they looked after their "small agenda" as best they could, and Rome ended up dominating for its allotted span. Same with the Mongols in Asia, same with every region, in every era including the present one. That's the game I want - one that doesn't know it's a game!

 

[tebriz]

I would totally agree with you, but only if C2C was a really open-ended simulation of the world and not just an great game. Even if it is an awesome mod, there is still a long way until C2C will feel like a real simulation of the world with dynamic rise and fall of empires where civilizations fall not only when they are conquered by other power, but also because of internal factors... Something like Rhye's and Fall of Civilization, but without preset spawn dates ofc, and not only that... Current Barbarian Civilization spawn system doesn't really give any new long term competitive civilization (whatever BUG option you try)... Basically all civilizations which were not conquered by others are still in game in later eras... and you don't see many new actors on world stage... Empires don't split and etc.... I would really like to see all of these implemented, but this is already another story
If all these concepts would be implemented in C2C, then I would totally agree with you that each civ in game should act as a real life civ with proper small agenda. But until then C2C remains just a strategy game, where the only aim of a civilization must be winning the game.

 

[Yudishtira]

Basically all civilizations which were not conquered by others are still in game in later eras... and you don't see many new actors on world stage... Empires don't split and etc....

All these have happened in just my current game!
- civ destroyed by revolutions before I even met it properly
- barb complex of 4 cities becoming a civ
- new civ spawned from barbs and then destroyed by more barbs
- original civ merged its one remaining city with me, but its revolution split-offs, with multiple cities, continue as viable

If they don't happen for you, it's not the fault of the mod...

With out any of RFC's artificial presets, I'm enjoying a real alternate history sim.

But I sense that's a side issue: you want AIs that will "game" you. C2C's AIs are certainly improving, but I couldn't say whether they're going in that particular direction - only that I hope not (nothing personal , just the opportunity cost)...

Then again, C2C tries to have options to suit everyone, so I hope it can accommodate you (as well as me obviously).

 

[Thunderbrd]

Throughout your conversation so far both of you have expressed very interesting polar viewpoints that I'd personally like to see both sides expressed fully within the scope of the game. I don't at all find them incompatible views. It just takes a lot of skill and work and focus on that region and it's something I'm sure we will address at some point. Despite the 'critical' nature that you point it out to be, it's still something that would take such constant adjustment to keep right during other project developments that it really needs to be done after things calm a bit. It'd be like using the fine sand paper before whipping out the grinder... no point to doing finish work until the shape of the crafted piece is pretty much in place.

 

[tebriz]

To Yudishtira,
regarding your previous comment I would like to add that there were quite a lot of cases when real life empires acted to preserve balance of power in region. One good example is European history of last centuries.

Regarding you last comment:

civ destroyed by revolutions before I even met it properly
I have also seen this happen a few times, but this looks more like a bug, not a working game mechanism. Have you ever seen this happen beyond prehistoric era? I haven't... This only happens when civilization is very small, usually with just one city... rebels conquer their only city and it collapses. What I would like to see is big empires stagnate because of internal problems and eventually collapse. I would like to see more challenges regarding internal problems of my empire, and current revolution system is too easy to handle in my opinion.

barb complex of 4 cities becoming a civ
This is true - you just need to not overcrowd map with lots of starting civilizations and you will see this happen. So, only thing you need for this is enough empty land on map. But once that empty land is filled with new spawn civilizations you don't get any new "rise" of civilizations and game loses its dynamism.

new civ spawned from barbs and then destroyed by more barbs
I have also experienced this, but again, it is possible on maps with lot of empty space where barbarians rule. So, you can't rely on barbarians to make game more dynamic later on when there is no space left for them, and you definitely need other mechanisms in order to keep game dynamic.

original civ merged its one remaining city with me, but its revolution split-offs, with multiple cities, continue as viable
I have also seen this happen many times... But it is not enough to consider game being truly dynamic.

I never said that I would like to see any artificial presets of RFC in C2C. I just like mechanism of RFC for domestic stability, its dynamism, and competitiveness of new born civilizations. I would like to see these concepts of RFC's implemented in C2C, but without artificial presets ofc.

To Thunderbrd,
Thank you for your comment. Unfortunately I am not a programmer to help mod, but at least I would like to give some fresh ideas