Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
If its deity and you are placed next to zulu, mongolia, huns, etc. You're going to die, but some games its actually not even hard. One time I won deity using Portugal and had only three cities, it really all depends, but some times you just know you aren't going to win.
In Deity, you really need to prioritize either being very friendly with your nearest neighbor, or just take them over.
In terms of "absurdity" though... sure, Deity is pretty ridiculous. That's not really a problem with the numbers though, more of a problem with how scaling is done. Giving AIs a headstart, and giving them large yield boosts from the very beginning, is obviously going to catapult them forwards, so you'll just be playing catch-up for the first half of the game.
We’re talking about the hardest gamemode, it should be ultra hard. Otherwise, the people who love ultra hard would be complaining the hardest gamemode is not hard enough.
Considering there are 8 difficulties, there should be, like there is, many different levels of difficulty — which if the hardest ones were too easy, what is the point of having 8 difficulties then? The 8 difficulties are for all levels of gameplay — super easy to super hard — so excluding ‘super hard’ would make having so many difficulties irrelevant, and the players who want super hard would be out in the rain, or would have to mod their own difficuties in.
Finding a more accurate gamemode should be easier, since there are 6 options between ‘super easy’ and ‘super hard’. I don’t play Diety, nor is there any shame in lowering the gamemode. Nor should there be any pressure for “you must play the hardest gamemode.” But otherwise, wanting the hardest gamemode to be easier because you want to play the hardest gamemode, despite it not being your thing to have it so hard, is like going to a horror movie and complaining it was too scary and all horror movies should be friendlier, despite one themselves not being into horror movies.
I am playing with an A.I. mod currently, it’s in fact the lite version so it isn’t even the best one the modder made, and I have been at war with an AI for the second time, because the first time, even though I had a massive army, it ended in a draw, and the 2nd time I am currently winning, but the war has been lasting over a hundred years, advancing one tile at a time.
It would mean that you'd have to first create an AI that's somehow good enough to go toe to toe with an experienced player - and create it at a stage where it's still unknown which strategies end up being strong and which don't - and then you'd have to create alternative AIs for every difficulty below that. You'd literally waste time programming in "rookie mistakes" for the AI. And then, if you ever release a patch that includes balance changes, you'd need to update every single AI to take those changes into account. The amount of work you'd have to put into that would be absurd.
In a game as complex as Civ, there should always only be a single AI, period. That AI should be as good as you can reasonably make it, and the rest needs to be done with bonuses or, if the AI is good enough, penalties for the other difficulties.
You can argue that the baselevel AI for Civ is far weaker than it could be, and I'd agree with that, but the conclusions that you draw afterwards are simply unworkable. No 4x has ever done it, and no 4x will ever do it - at least not until there's some method that revolutionizes AI-development in a game this complex (and with matches this long).
the A.I. plays worse at lower levels, and improves until Prince difficulty. At Prince, the AI gets the same bonuses and everything as the player does. After that, the A.I. doesn’t get better, it just gets more bonuses. So it does this already, just by the time you get to King, then the A.I. is already the best it can be in the stock game, and just increases via bonuses. So AI does ‘play worse’ at lower difficuties, and improves with difficulty, until it reaches it’s max stock performance at Prince, and King+ just increases performance rather than intelligence.
If you mean that the AI works like that in Civ 5, I need you to provide proof for that. The only things where to my knowledge one could make such a claim is when the AI considers more low-flavor technologies and policies on lower difficulties, which, on average, leads to techs and policies that fit their desired strategy less (if flavors are distributed well). But other than that? I don't think that's the case.
That's... not entirely true either, but that's off topic.
That's also how AI is scaled down below King, it gets yield penalties.
For Civ 5 and 6, the AI nor the player gets any bonuses or penalties on Prince. Below Prince, the player gets bonuses and at the bottom end, the AI gets penalties on top of that.
Above Prince the player still gets no bonuses (but also no penalties) and the AI gets increasingly more bonuses. In both games it has to play by the exact same gameplay rules as any other player however. So yes, it gets a percentage production bonus, but cannot spend the production twice or has its buildings or units cheaper. It also has the vision and fog of war as any player has.
This for example means that if you destroy their production, it's actually destroyed (80% of 0 is still 0 ;) )
In Civ 4 and before the AI actually cheated. It had visibility for example that it shouldn't have had. This makes it seem more competent then it actually is.
Chess games use the exact same resources, yet have several difficulty levels. The only difference between them is how well the AI plays. You may think it "silly" but there are people out there with varying levels of skill to play chess (or Civ or any other game) and these different levels of AI are what allows them to find a setting that matches their skill.
Racing Games - AIs drive slower, not worse.
RPGs: Enemies do less damage, and/or you do more damage.
ARPGs/Hack'n'Slash - AIs deal less damage, and/or you do more damage.
Tower Defense - Your towers do more damage or AIs take more damage
RTS - AIs do fewer actions per second (=play slower), not play worse strategies (although some games do both here)
Shooters - You take less damage, and/or do more damage, AIs don't act more stupid (although some games do both here)
etc. etc.
There are other genres where "playing worse" is the standard, but it's clearly not universally true. It very much depends on the game, because yeah...
...balancing chess by introducing extra pieces completely changes how the game is being played. If we look at another example - Go, the "Chinese Chess" - we actually see that in this case it's very doable. So much so, that the standard method to equalize the board for players of different skill levels, is to let one start with extra pieces[en.wikipedia.org].
Aside from that, Chess AIs are also very easy to program (or rather, train) compared to what you'd need for a high-skill Civ AI that can beat good players without cheating, because Civ 5 is like... 5000 times more complex, and the game plays on completely randomized boards.
Plus, if you have even basic knowledge of the current situation when it comes to AIs, then you'll know that the good chess AIs are neural network based, a technology that once again is completely out of the question for a game like Civ at the point we're at now. There's no way you compute your way through the tens of thousands (or more) of iterations (through the generations) that would be necessary to come up with anything remotely able to beat a good human player.
Overall, your opinion of what should happen is completely incompatible with reality.
To run Civ's AI in such a manner (calculating all possibilities and then limiting the number of turns it can plan) would mean excessively long wait times between turns as the number of choices and resulting branches are vastly larger than chess.