So they still haven't fixed the AI with regard to war.

[Aristos]

If you see an enemy submarine in a one tile fresh water lake for 50 years without moving, you know its a clear case of Lazy Programming.

Then what is it if you see a Huey Teocalli in a one-tile lake?

A battleship armada patrolling a 3-tile lake?

A Petra in the only desert tile of the city?

Etcetcetc...

 

[japanesesamurai]

This is the worst over simplification of a problem I've seen in a while.

First of all this is some seriously complex software, finding niche cases like this can be incredibly difficult and can be the result of hundreds of thousands of lines of logic spread out across a huge code base. Software isn't built as one gigantic script of if else statements..... in fact the worst sort of programming practice is to do what you described above.

There may well be problems with the AI but what you think should be an easy fix here or there is just not true. Especially in a codebase like Civs that's grown for decades. Rest assured it's a much more complicated issue than you'd imagine

Hi!

I’m certainly not disagreeing with you. The UPT tactical aspect of the game is super complex. It is probably the single most complex thing to program properly. I am aware of that.

However, It is not a matter of changing its tactical combat behavior. Its a matter of changing its built behavior. If you limit its unit options, or change its built priorities, it is less likely to make tactical blunders.

For instance:

Map = 60% land tiles

40% = water

If only 7% or less of the total amount of water tiles is ocean (or if ocean tiles are highly fragmented), let AI prioritize ranged ships other than subs. This is a very simple map calculation and a solution to at least minimize builts of “useless ships”.

The other way around, if map has huge chunks of ocean tiles, make subs more likely.

Problem with civ is that AI builds same kind of units in every city regardless of the location/characteristics of that particular city. If city is adjacent to a lake, don’t build subs. Simple.
You can’t build an observatory in a city without an adjacent mountain. AI will not do it because its a simple rule. Same thing. Simple coding.
(What if AI shares the lake with an opponent city? It will still not put a sub in a lake. Who on earth puts a stealth boat in a lake?)

In the eind, AI turns might end up taking a bit longer to calculate, but so be it.

Although a queen is a more powerful piece, sometimes “upgrading” your pawn into a knight, is the better option

 

[japanesesamurai]

Then what is it if you see a Huey Teocalli in a one-tile lake?

A battleship armada patrolling a 3-tile lake?

A Petra in the only desert tile of the city?

Etcetcetc...

Battle ships are ranged units. They can attack nearby land units. And thus are useful

Petra gives a free caravan and trade slot. Heck, if i had a city on a two tile desert island, I would even consider Petra. Delete caravan and build cargo ship for internal food delivery. (Civ 5)

 

[Aristos]

Battle ships are ranged units. They can attack nearby land units. And thus are useful

Petra gives a free caravan and trade slot. Heck, if i had a city on a two tile desert island, I would even consider Petra. Delete caravan and build cargo ship for internal food delivery. (Civ 5)

I would love a MP match against you. For money, if possible.

 

[japanesesamurai]

Ill pass!
Only happened in one case:
I spawned in the middle of nowhere all alone in the ocean on tiny islands.
Only way to be productive and grow properly was to get as much as internal trading as possible. So collosus and petra was priority.
I ended up with tiny island cities with 30+ pop without having strong tiles (lack of hills and fresh water) It was an interesting network of internal sea trade boats for sure

 

[Socrates99]

I would love a MP match against you. For money, if possible.

why is mp challenging a valid response to an argument over AI?

Considering the production bonuses AI gets it's not as much of a potential loss to build units that might be dumb for players to build. Finding frigates or battleships in a landlocked lake near a city center I plan to take is a real dilemma considering how powerful they are vs contemporary units.

I'd have no problem if the AI did this more often even if it is a stupid move for a player.

 

[japanesesamurai]

Whats wrong with having battleships in a lake nearby your city?
They have a range of 3 and are deadly. And infantry are incapable of attacking them properly
They are power monsters if you’re faced with an invasion

 

[Socrates99]

Whats wrong with having battleships in a lake nearby your city?
They have a range of 3 and are deadly. + they ignore differences in heights. Meaning, they can shoot over hills any day

I think the idea is that you're not getting much bang for your buck by building a unit that is permanently locked in one city. It can't help defend other fronts and it can't be used to invade. If you're never attacked its wasted cogs/gold.

Still not a great reason for an AI to not do it though, especially if it's at the capital you're forced to take in order to get a dom victory. I'll take any challenge they give me.

 

[japanesesamurai]

Yes exactly!
Build them if you’re facing invasion

 

[Aristos]

why is mp challenging a valid response to an argument over AI?

It isn't. But a bad use of resources is a bad use of resources, and the fact that the AI is incapable of extracting punishment for extremely inefficient allocation is not an argument in favor of such.

Or maybe I'm in the mood for something more interesting than civ 6's AI... MP or VP, whichever P.

 

[japanesesamurai]

Ah yes but I’m sure we play the game (civ 5) quite differently. I disable science victory (because it ends up being a boring clickfest), diplo victory (cause its easy to bribe and win if theres no other option for a victory) and often disable manhattan project as well because its easy to buy 2 atomic bombs quickly to neutralize en enemy army in one turn. I avoid boring and repetitive play. As in “Upgrade to bombers, bombard from a distance, paradrop, capture city, repeat.”

I only enable cultural victory and domination. On a small strange tectonic map with lots of agressive or culture civs on immortal. And if I start on a single island tundra or snow tile, I dont restart the game. Sometimes i start with Germany and delete my settler on turn 1. Whatever keeps it interesting. So yeah, I probably loose on MP as I dont quite have the optimal strategy for multiplayer games.

That being said, this has only so little to do with the actual content of this thread

 

[Socrates99]

It isn't. But a bad use of resources is a bad use of resources, and the fact that the AI is incapable of extracting punishment for extremely inefficient allocation is not an argument in favor of such.

Or maybe I'm in the mood for something more interesting than civ 6's AI... MP or VP, whichever P.

If the goal is to foil the player I dont think an inland battleship is actually a bad use of resources. It could build bombards instead and the player easily takes them out. Frigates and battleships are hard to counter from land. Like I said, I wouldn't mind seeing it more often than I do.

We dont truly build for defense because the AI isn't a threat but if I was turtling up and wanted what essentially is a floating encampment I might go for it.

The threat would have to be real and the map would have to have the right setup of course. Unfortunately I think GSs fuel req might prevent me from ever doing it while in RnF I might have considered it.

Think of Pedro going for a great person style CV. Hes not going to invade. He has no reason to use his UU aggressively but if he has an aggressive neighbor he could use that landlocked ship(s) to stymie an invasion for an era or two.

 

[Bradypus]

Naval advantage on water maps (comparable between games), UN vs world congress, culture vs loyalty, xp farming/promotions, pillaging (as of GS), overflow (civ 4 bugged it and never fixed), government vs policy cards (timing swaps to optimize is more complex in Civ 6), eurekas/inspirations (no Civ 4 parallel) all come to mind offhand. I'm sure there are others.

I was referring to combat exclusively. Like if you had a slider for each factor in combat (such as composition, numbers, positioning, etc.) and you had to distribute 100 points between those factors in order of importance in relation to the other factors. At the moment I get the impression that Civ 4 must somehow have higher values assigned for each and every combat factor than Civ 6.

 

[orasis]

Firaxis should just hire the dude that created the AI+ mod. It's a damn good mod but imagine what could be done if he could see the source code.

 

[RickSL]

Then what is it if you see a Huey Teocalli in a one-tile lake?

Huey Teocali buffs every lake tile in your empire, so building it in that circumstance can make sense if there are lakes elsewhere.

 

[TheMeInTeam]

I was referring to combat exclusively. Like if you had a slider for each factor in combat (such as composition, numbers, positioning, etc.) and you had to distribute 100 points between those factors in order of importance in relation to the other factors. At the moment I get the impression that Civ 4 must somehow have higher values assigned for each and every combat factor than Civ 6.

That narrows it down. Promotions are more impactful in Civ 6 overall. You could also make a case that collateral initiative is a bit too overcentralizing for most of Civ 4's tech timeline, with nukes completely upending that and effectively winning the game for the civ getting them in any quantity first. I'll give Civ 5/6 credit for being less RNG dependent too, that could be a big deal sometimes in 4.

In most other way's yes Civ 4 is superior tactically.

 

[Aristos]

Huey Teocali buffs every lake tile in your empire, so building it in that circumstance can make sense if there are lakes elsewhere.

Theoretically, but in this case, the AI built it in the ONLY lake tile it had.

 

[kaspergm]

Theoretically, but in this case, the AI built it in the ONLY lake tile it had.

I'm not going to argue against your point, because I agree with you, but you can put forward the argument that AI should build a wonder if it has an available spot for it - even if it's not an optimal spot - and if it hasn't got anything better to build, for two reasons: 1) Because it thereby removes the opportunity for an opponent to benefit from said wonder, and 2) Because it will give it era score and possible adjacency bonuses.

Now arguably I agree that in case where AI only has one 1-tile lake in its entire country, it will - must - have better things to produce than Huey Teocali (and I've seen it do this many times also, plus Petra and Chichen Itza on the only desert/jungle tile within range of a city).

 

[Aristos]

1) Because it thereby removes the opportunity for an opponent to benefit from said wonder, and 2) Because it will give it era score and possible adjacency bonuses.

This "AI" is, I can assure you, absolutely incapable of such a subtlety. That is just too much credit for this thing.

 

[SaMatra]

They still didn't fix day 1 issues so many years later with latest expansion:
1) Trying to heal unit while they're under archer fire, happens all the time, they retreat a bit and go into fortify just to die in few turns by my archers.
2) Building settlers or even wonders while losing war. So many times I was sieging the city, destroying walls over 10 turns only to see the city pop out a Settler instead of something that could help them.
3) Not attacking with units in towns, they sometimes fire archers from town but never ever attack with melee units so that they do damage but not leave the city
4) Sending settlers through enemy territory because war finally opened up borders, also sending unprotected settlers over water all the time, easy captures.
5) Had so many cases where AI could totally take my city in next turn but retreated instead (playing on Immortal to avoid 3-settlers start)

This game is hopeless at this point, I wish somebody would wake the devs up, so much potential is wasted by stupid AI.