useless AI in svn version

[rashaverak]

Hello and sorry for such a strong words but the AI is really useless.

Monarch difficulty, AI is probably not able to support advanced start, with 25000 points all had only 1 city except of one, but that is the 'minor' problem. Even when they researched tribalism they do not start to create any cities. Together with the hopeless way how they attack and how they do not defend their cities the AI state is awfull.

I have to ask do you even still play your own mod ? From point of view the mod is unplayable. 'Stable' v35 release crashed so often I switched to svn just to find out the AI is in this state. I returned to try the mod after about 2 years and while there a few new features, the overall state is really poor.

One of you who develope said you cannot only fix you need to develope new features to keep the developers happy. I am developer for 12 years. I agree that to develope new feature is much more fun than fix crashes or optimize something. But if you keep the mod in this way and just develope new anew non working things which ai doesnt support you will loose players and effectively lead this mod to dead. You are loosing me for sure because after I returned this was my 4th game (and I am playing huge maps with slowest research so it is some time spend in useless games), I lost mood for any other attempts.

Such a game is not fun at all.

edit: Please dont get me wrong, it is great what you are doing, I love the mod playing it since MoM and then when it become c2c but that is the reason why it feels so sad to cannot play it anymore. The ai problems are known for weeks according to the bug threads and there seems to be no care about it ... while other minor things get fixed. I understand that only a few people in the team can handle it and it is complex, but there seems to be not much care.

When I had crash every turn and I was told it is not priority to fix such things, well I was surprised and I could say something how wrong attitude that is, but I got over it, this ai problem is just too much.

Good luck.

 

[olduser]

One of you who develope said you cannot only fix you need to develope new features to keep the developers happy. I am developer for 12 years. I agree that to develope new feature is much more fun than fix crashes or optimize something. But if you keep the mod in this way and just develope new anew non working things which ai doesnt support you will loose players and effectively lead this mod to dead.

I can only agree here but apparently it is better to just add infinite amounts of new stuff. If you suggest that modders should test the new Stuff they add In-Game before committing it to the svn it is bad because they loose interest. If you build in some xml error checking it is bad because they have to get their xml right and then they loose interest

The result of this was visible with the v35 release, untested new stuff(in that case some animals) was added just before the release and caused those crashes.

 

[Arakhor]

As far as I can tell, the AI is somewhat overwhelmed, yes. The Prehistoric Era was actually pretty good the last time I played, but then I only play on Noble (or Prince if I'm doing well), so I don't expect too much from the AI. AI competence drops off very quickly as you move through the years though.

 

[darkkhon]

Sadly I have to agree with the first statement.The AI was not great a few versions ago,but now is just really bad.It barely expands(really really slowly) and although I have almost reached medieval era I haven't seen a single city being taking over by force,despite the fact that the AI has tons of units.Although I do appreciate the efforts made regarding the graphic part,in it's current state the mod is unplayable.

 

[Thunderbrd]

I'll go on record saying, yes, I know the AI sucks right now. I'm not sure it sucks worse than LAST version because it's basically all the same AI with the same flaws. It's just that some recent changes to major underlying game values have revealed the AI's inability to compensate for all potential settings. As for the settler problem, it looks to be something that's been with us for a while and I'm not really sure what exacerbates it now but it does need addressing.

It may seem as if I and other members are only working to ADD things still but where I'm currently working is on what is also a currently unpolished and distorted portion of the game. That project began before last version's release and if you constantly derail your current projects to go after the next identified major problem you'll never get anything done.

The whole of the AI structure, particularly in unit behaviors and build choices, is certainly a BIG and MAJOR project but does need doing and very soon.


Also, it's not that many bugs are of no priority, it's just the timing of when in the cycle they'll get primary focus.

 

[Sparth]

Maybe AI has problem with gold? Currently im playing little longer 1st time after v35 and I have some problems with gold in prehistoric era (marathon/prince/huge map). With 3 cities only.

 

[Arakhor]

Recently the Prehistoric Era has been slightly better for gold once you start expanding, yes, but at Ancient and onwards, you can still run 100% science permanently if you see fit.

 

[Sparth]

Recently the Prehistoric Era has been slightly better for gold once you start expanding, yes, but at Ancient and onwards, you can still run 100% science permanently if you see fit.

Yea I like situation when im struggling with lack of money. Ok ill see in later eras.

 

[JosEPh_II]

Recently the Prehistoric Era has been slightly better for gold once you start expanding, yes, but at Ancient and onwards, you can still run 100% science permanently if you see fit.

Sorry old friend but that last part "you can still run 100% science permanently if you see fit" is just not true in All game speeds or Difficulty levels.

JosEPh

 

[Arakhor]

That's quite possible. Maybe I should have specified that I am (of course) only speaking in my experience, which is mostly Marathon and on difficulties up to Prince.

 

[Dancing Hoskuld]

Playing Noble/Snail you can't keep your science at 100% until late Medieval from my experiance. I do play with Rev on which basically means that if you go below 60% science your cities will revolt. I have had all my cities revolt in one game where I did not get the currency tech soon enough.

 

[rashaverak]

playing prince/monarch prehistoric on slowest speed is 100% research and actually I never got to a phase when I would have to lower it if I care. But I never got later than to industrial era (but there it is like permanent +3000 gold per turn with 100% research)

I would backup the idea that ai has problem with gold, but not in the way they do not have the gold but that they keep too much, that could also explain why the advanced start computer seems to not spend the money on anything

 

[Sparth]

I would backup the idea that ai has problem with gold, but not in the way they do not have the gold but that they keep too much, that could also explain why the advanced start computer seems to not spend the money on anything

In short words: AI is too greedy?

 

[Noriad]

I prefer to keep science at 100% at all times. Since most income from buildings is already either science or money, changing the sliders has much less effect than in Civ4:BTS anyway.

I make up the difference by building Lesser Wealth. Usually, up to half my cities are building Lesser Wealth, the rest are building cheap income buildings. Hammer producing buildings first, then money/science buildings, then food buildings. You save time by using the in-city filters and queue up a large number of buildings. When less than a quarter or so of my cities are building Lesser Wealth it is time to settle/conquer another city. The new city will be a net money loss, but you get additional science and culture, plus more terrain resources, plus the opportunity to build more of the cheap buildings.

Also I am reluctant to build the more expensive buildings, like cost 500 hammers for a 1 hammer + 1 happiness increase in income. This is just a bad investment. If it produces a resource I will build one, the rest of the cities skip it and build wealth or science instead. I keep enough money stored to survive 3 or so turns without building wealth, the rest I will spend by having less cities build wealth.

I also keep a fairly large army (up to the point were unit cost starts to increase rapidly) of strength-3 melee units, rogues and hunters roaming the countryside to subdue animals. I build a hunting instruction before building strength-3 units. My tech tree route follows this; first I research hard hammer percussion (for stone tools building) then the tech that gives hunting instruction, then piercing tech to get spiked clubmen. Then I build spiked clubmen up to the unit limit where costs start to increase rapidly.

Since hunters have a higher chance to subdue, they get to attack the more valuable animals. If you are careful with them, your strength-3 units shouldn't die so they rack up quite a bit of xp and become exceptionally strong (many end up becoming level 8-10.). It also saves you the hammer cost of building new units.

Careful means no attacking unless you have 98+% chance to win, or if chance is above 85% I park it on a 50-75% defense bonus square and let the enemy attack me. Careful also means only moving to squares with a defensive bonus unless you start on a hill and see that the immediate surroundings are safe. Careful also means not ending a movement within the first ring of an enemy/barbarian city, or only on a 75% defense tile within second ring. Once I got obsidian weapons tech I upgrade my high level melee units to Obsidian Swordsmen.

My rogues tend to stay close to enemy territory to kill as many enemy units as possible, since each kill nets up to 20 money. Also the captives (military) can be quite useful in speeding up building, if I have powerful units and a barbarian city nearby I tend to thin out their armies just for the captives and use those to build the expensive buildings I skipped earlier.

P.S. I play Deity, snail speed, nightmare on, Large PerfectWorld2F map with "start anywhere".

 

[Sparth]

Thanks for valuable tips

 

[strategyonly]

P.S. I play Deity, snail speed, nightmare on, Large PerfectWorld2F map with "start anywhere".

Do you like the Nightmare Mode better??

 

[Noriad]

Do you like the Nightmare Mode better??

I have no idea at all what Nightmare mode does except that it states that it makes the game harder. Which is what I like. "start anywhere" also makes the game harder as it spawns civs further away so you can't just spiked-clubman rush a close neighbour early in the prehistoric era.

P.S. The 90% maintenance autobuildings basically increases your effective city limit by (almost) 2 as long as you don't change from anarchism civic to chiefdom.

 

[IronClaymore]

I have no idea at all what Nightmare mode does except that it states that it makes the game harder. Which is what I like. "start anywhere" also makes the game harder as it spawns civs further away so you can't just spiked-clubman rush a close neighbour early in the prehistoric era.

P.S. The 90% maintenance autobuildings basically increases your effective city limit by (almost) 2 as long as you don't change from anarchism civic to chiefdom.

You can check out the different xml's to see the difference nightmare mode does. I don't know if any of the values have changed between v35 release and svn, but for the release version nightmare mode makes deity difficulty have: slightly less free units, much higher research cost, massive increase to basic city maintenance values, bit more inflation, seriously higher revolution index, less animals but more barbarians and barb cities, and the AI hate you more and make stuff faster. Oh, and you might start off in worse terrain than you would otherwise. This is compared to regular deity mode.

 

[rashaverak]

tested another prehistoric game on emperor dificulty. After computer invents tribalism they start to build cities, but it seems pretty slow, slower than v34. But I am not 100% it is really slower just seems so, but they stop at 3 cities and do not build anymore unless they invent (cannot remember it is the civic before monarchy) this is definately a change, before they could manage more cities before, you can imagine how long it takes them to get the research with just 3 citiees while player is able with the same civics to manage 7-8 cities.

edit: and this behaviour is the same if you have the options to limit cities by civics on/off that doesnt matter which is also pretty weird (with the option off it is even easier to totally outnumberred them, I had even some message I never before saw that I am somehow over populated, I was about 45% of worlds population I think, with 24 AIs .... in prehistoric era) I dont want to play deity, I am not so good at that, but currently my usually fun difficulties (prince-immortal) are really too easy mode

 

[MacCoise]

tested another prehistoric game on emperor dificulty. After computer invents tribalism they start to build cities, but it seems pretty slow, slower than v34. But I am not 100% it is really slower just seems so, but they stop at 3 cities and do not build anymore unless they invent (cannot remember it is the civic before monarchy) this is definately a change, before they could manage more cities before, you can imagine how long it takes them to get the research with just 3 citiees while player is able with the same civics to manage 7-8 cities.

There is such a severe maintenance penalty going from Anarchism to Chiefdom which might play a role here

Do they switch to Chiefdom and then find themselves broke?

PS Is there any benefit to switching to Chiefdom is happiness penalties are turned off? Two free citues is a pretty big bonus