I know AI gets bonuses on Diety, but this is ridiculous

[szieker]

Taking a rare effort at Diety. Don't play diety very often (have never won).

I know the AI gets a lot of bonuses, but some unknown AI has managed to get a reformation belief on his religion by turn 65.

How is this even possible? There are only 2 other civs in the entire game that even have a pantheon.

This is a good example of why I dont like the higher difficulty levels. I'd much prefer the AI have the ability to just play smarter, rather than just get unfair and ridiculous cheating abilities.

 

[Scott Jegg]

Probably not that difficult for Poland to do this at any level.

 

[Yzman]

Everyone would love for a "smarter" AI. Unfortunately it isn't practical to do at this point in tech. So they give AI boosts in other ways.

I don't enjoy playing largely unfair games, so I stick to playing at king and occasionally emperor even though I always win.

 

[compwiz1202]

I'd think Ethiopia with wide Stelemania and lucky ruins could too. Maybe even anyone with lucky ruins.

 

[kamikazees]

You'd only need like what, 25 culture ruins to get a Reformation belief by that time, right? No problem! Especially considering that the only direct culture from Piety comes from settling great prophets.

 

[Yzman]

You'd only need like what, 25 culture ruins to get a Reformation belief by that time, right? No problem! Especially considering that the only direct culture from Piety comes from settling great prophets.

Yeah I mean ruins are the only way to get culture! So lets think about this. If I am usually probably half way through tradition by that turn, the AI has 2 settlers and thus can have double the monuments of me, then they could be Poland and get a free social policy or 2.

Yeah not seeing what is so crazy about this. We know they get bonuses to tech and units, happiness etc at that level. They don't get free policies and it seems reasonable to be able to do this, if still a bit lucky.

 

[apotheoser]

Everyone would love for a "smarter" AI. Unfortunately it isn't practical to do at this point in tech. So they give AI boosts in other ways.

Of course it is practical. There just aren't a lot of good AI programmers.

Moreover, designers typically aren't that good at their own games. They can't design a challenging AI if they aren't the best players to begin with.

 

[Yzman]

Of course it is practical. There just aren't a lot of good AI programmers.

Moreover, designers typically aren't that good at their own games. They can't design a challenging AI if they aren't the best players to begin with.

Oh really? Its practical? Are you a master programmer? Also, practical means not only that it can be done, but that it can be done in an efficient amount of time so that budgets on games do not go over.

As an engineer I can understand the problem all too well. Its easy enough to design something, its hard to design something well and there is NEVER enough time to do it.

One of my project managers has a quote above his desk

"Eventually you need to shoot the engineer and finish the project"

My understanding is that programming is the same as this. Its just not practical to make a "smarter" AI in the time these people are given.

 

[Scott Jegg]

Its just not practical to make a "smarter" AI in the time these people are given.

Yes. It takes masses of code to give the AI enough "experience" to deal with every situation presented to it in an efficient manner and make it seem smarter. It not really a case of more "clever" code, although there's certainly plenyt of areas where current algorithms could be improved (unit movement for a start).

 

[kojax]

Yeah I mean ruins are the only way to get culture! So lets think about this. If I am usually probably half way through tradition by that turn, the AI has 2 settlers and thus can have double the monuments of me, then they could be Poland and get a free social policy or 2.

Yeah not seeing what is so crazy about this. We know they get bonuses to tech and units, happiness etc at that level. They don't get free policies and it seems reasonable to be able to do this, if still a bit lucky.

You don't have to unlock the whole tree to get a reformation belief, just 5 policies.


So, 25, then 30, then 60, then 105, then 170? So 390 total culture. 65 x 6 = 390, so they need to average 6 culture per turn.

 

[VisageTDI]

That's quite easy to achieve, especially if they befriended or allied a culture city-state or two.

 

[FrostK]

65? is not that fast actually. U need just 5 policies[inluding initial] to get reformation. In average game one can have 5 in tradition till 65 turn. Of course, piety does not provide culture like tradition, but this is Deity. I think people may easily have piety-4 till turn 65, or even also take reformation if u get 2 culture-cs allies early.

 

[Socratatus]

Taking a rare effort at Diety. Don't play diety very often (have never won).

I know the AI gets a lot of bonuses, but some unknown AI has managed to get a reformation belief on his religion by turn 65.

How is this even possible? There are only 2 other civs in the entire game that even have a pantheon.

This is a good example of why I dont like the higher difficulty levels. I'd much prefer the AI have the ability to just play smarter, rather than just get unfair and ridiculous cheating abilities.

Obviously Diety is made for the BEST of the BEST! It`s for the ELITE gamer that spend ALL their time 24\7 with a calculator playing the computer inside and out at its own game! That`s what it`s measured to. So if you`re not up to it, of course it`s going to seem ridiculous.

I don`t touch Diety cos I know it`s going to be ridiculously hard with the AI getting superman bonus` and that`s not my idea of fair or fun. I don`t play `crunch the numbers` like some do.

I don`t know why you and some others take on the most difficult of difficult levels then post to say it`s too hard\ridiculous. There are some out there who think it`s not hard enough.

 

[Resipsa]

If they have a culture pantheon and a Culture CS ally it can be done

 

[apotheoser]

Oh really? Its practical? Are you a master programmer? Also, practical means not only that it can be done, but that it can be done in an efficient amount of time so that budgets on games do not go over.

I am a programmer, yes. What I am trying to tell you is that hiring your average coder to work on AI is like hiring your average coder to work on art or sound assets.

As an engineer I can understand the problem all too well. Its easy enough to design something, its hard to design something well and there is NEVER enough time to do it.

One of my project managers has a quote above his desk

"Eventually you need to shoot the engineer and finish the project"

My understanding is that programming is the same as this. Its just not practical to make a "smarter" AI in the time these people are given.

This would be true on its own, but the problem is also that writing AI for Game A is not the same as writing the AI for Game B. You have all the normal problems of moving from one project to another, maybe with different tools, but then suddenly too you're talking about an entirely different game with different rules.

We are talking about a game that has been out for years with multiple patches. They've had plenty of time to write good AI. The problem is that (1) they are not capable of doing it at all and (2) it's not really clear that it would be a good business decision. People still buy the game a lot and most people cannot beat higher difficulties as it is. Why invest resources in making it harder?

 

[dragonxxx]

You do realize, that if they do the "perfect" AI, the game would become pretty much unbeatable on higher levels, right?

Imagine deity AI starting with 2 settlers, 2 workers, some units, bonus luck, food and production and on top of that making perfect decisions, while handling military units like an expirienced player.

Sure, there will be probably a handful of ppl who are gonna beat the AI from time to time. The rest will simply not play it ...

So why bother?

 

[civ54lyfe]

i would have guessed that the largest challenge to introducing smarter AI in Civ is that they haven't changed much of the vast base code, or engine for the AI. i would have thought that truly overhauling the AI logic would take a whole new effort, e.g. a Civ6.

 

[JimboOmega]

You do realize, that if they do the "perfect" AI, the game would become pretty much unbeatable on higher levels, right?

Imagine deity AI starting with 2 settlers, 2 workers, some units, bonus luck, food and production and on top of that making perfect decisions, while handling military units like an expirienced player.

Sure, there will be probably a handful of ppl who are gonna beat the AI from time to time. The rest will simply not play it ...

So why bother?

The player can always save/reload or restart when something goes horribly wrong. The AI will never have this luxury.

The AI also always lives in a world with many other AIs, and an optimal AI would pay them much more attention than a handicapped human player anyway .

 

[project_mercy]

You do realize, that if they do the "perfect" AI, the game would become pretty much unbeatable on higher levels, right?

That statement makes no sense.

The fact the AI is coded to cheat is a direct result of an attempt to make something stupid be tactically and strategically difficult. The difficulty is set by Firaxis, it didn't magically come into being. If they changed the AI, they could just as easily change the cheating (and should have, with the last two expansions).

Do you know all those grand-master chess matches where one of the players randomly spawns extra pieces? How about those football matches where one side has extra players on the pitch? You don't, because it doesn't happen. The game's difficulty is based on the skill of the players. The game is otherwise equal and fair, that's why it's a game. In some games when you know the players are just not at equal skills, you provide some sort of handicap for the inferior player. In this case, the AI's cheating is its handicap.

What bonuses the AI gets is purely set by Firaxis in regards to what they felt met their level of difficulty. Or, did, back when they created it.

The issue here is, the AI fails at everything. There is not one thing the AI does well. It can't fight, it doesn't build well, it rarely has any strategy, it's never successfully prosecuted a war past its own borders. It's just 100% terrible.

The OP's dream would be that the AI at least develop some level of either tactical or strategic ability. As it is, the whole game is just horde mode where some of the zombies are not combat zombies. The difficulty level you choose to play on is purely a result of how much clubbing of mouth-breather AI units you can manage in a period of time.

On the lower levels (king or emperor and down) I think it's less of an issue. You know you're going to win, it's just a function of exploring and building and the AI is off doing its own thing. Emperor seems to be about the balance of where the AI's building failures match their bonuses. On the higher levels, it's mostly just a war game that is pretty terrible in comparison to 4x games, because the AI is unable to wage any semblance of war beyond zerging with "overwhelming" forces.

So, I think it's fair for the OP to complain the AI is terrible. It is. How difficult it would be to improve, I have no idea. I imagine it must be pretty painful, else they would have done it; since the terrible AI is functionally the deal-breaker for everyone I know who loves 4x and hates Civ V. But as for the viability of it? I think it's in their court for who's at fault. Total War is way better, Most of the Paradox games are better. When someone decided on changing the franchise to a new unit method, they should have investigated whether they could actually make it work first.

I would propose the true issue is, a lot of people play on Deity who really don't really want to do it, because of achievements and such. Deity really just enhances all the breakdowns in the game. I think things would be a lot better if the only people who kicked into deity knew they were facing a boring seal-clubbing scrum. Then, there would be no reason to complain about it.

 

[bhavv]

You do realize, that if they do the "perfect" AI, the game would become pretty much unbeatable on higher levels, right?

Imagine deity AI starting with 2 settlers, 2 workers, some units, bonus luck, food and production and on top of that making perfect decisions, while handling military units like an expirienced player.

Sure, there will be probably a handful of ppl who are gonna beat the AI from time to time. The rest will simply not play it ...

So why bother?

Because perfect AI without such huge cheating would be better than crappy AI with cheating.

I've beat the game many times on Immortal, but I now only play at prince / king max because I like to be able to wonderspam. AI construction / happy bonuses at higher levels makes it impossible outside of a OCC to get many wonders built before the cheating AI.

Giving the AI so many cheats is fake difficulty, and dumb AI.