|
Gort posted:she should be less co-operative with such a player since they're a threat. Why? The best way to deal with a potential threat is to be friends with them. If some big mean motherfucker sits next to you at a bar, you don't punch him in the face, you buy him a drink.
|
# ? Aug 5, 2016 01:42 |
|
|
# ? May 14, 2024 16:51 |
|
Baronjutter posted:Sorry for Civ5 posting but I've been trying to play on "terra" maps because I had so much fun with the overseas land rush in Civ4 with the same setup, but no one loving colonizes off their home continent. Even in our starting continent there's all these very close by islands with great luxuries and no one builds more than 5-6 cities then just stops. Is the Ai programmed to always go "tall" ? Civ5 drives me nuts with how much land is empty in the late game. I really hope the world better fills up in 6. Keep in mind that 'continent' means something different in Civ VI. It's not just 'part of the same landmass', but are rather defined and named areas in the game. I'm not sure what the algorithm is, but even Pangaea maps are reported as having multiple different continents. So things like the American bonus for fighting on the home continent or the English Redcoats bonus for foreign continents can come into play differently even for Pangaea. As for the AI/spreading wide in Civ 6, who knows? With local happiness/amenities, at least, I haven't seen a reason to not go wide yet, unless they keep the Tech/Culture penalties.
|
# ? Aug 5, 2016 02:39 |
|
Gully Foyle posted:Keep in mind that 'continent' means something different in Civ VI. It's not just 'part of the same landmass', but are rather defined and named areas in the game. I'm not sure what the algorithm is, but even Pangaea maps are reported as having multiple different continents. So things like the American bonus for fighting on the home continent or the English Redcoats bonus for foreign continents can come into play differently even for Pangaea.
|
# ? Aug 5, 2016 02:42 |
|
Gully Foyle posted:Keep in mind that 'continent' means something different in Civ VI. It's not just 'part of the same landmass', but are rather defined and named areas in the game. I'm not sure what the algorithm is, but even Pangaea maps are reported as having multiple different continents. So things like the American bonus for fighting on the home continent or the English Redcoats bonus for foreign continents can come into play differently even for Pangaea. They said they try to make continents have 2-3 starting civs in them. So I'd imagine they just draw a rough circle around those civs and call it a day. Flavor-wise it's supposed to be the border between Europe and Asia.
|
# ? Aug 5, 2016 03:02 |
|
Earlier I was thinking this means an end to landlocked maps, but I guess the algorithm could designate areas (divided by mountains, rivers, deserts, dense jungle) as "continents", as with Europe and Asia. Don't know how Vicky would cope with that but w/e. Some AIs just don't play as well as others. In Civ V I've seen Wu Zetian go big only once; every other time, she seems to die or get reduced to a rump state early.
|
# ? Aug 5, 2016 03:15 |
|
Victoria hating people she who don't share a continent with her is a seriously neat bit of design. It means she's very likely to attack those savages over the ocean, but once she's actually taken a city or placed a colony, she'll go all out for peace. Resulting in a very "British" empire with bits spread all over the world. Spreading so thin would be lethal for most civs, but then, england gets extra trade routes and redcoats who fight better overseas, so perhaps it evens out in the end.
|
# ? Aug 5, 2016 06:36 |
|
Fister Roboto posted:Why? The best way to deal with a potential threat is to be friends with them. If some big mean motherfucker sits next to you at a bar, you don't punch him in the face, you buy him a drink. Your real-world analogy is not relevant to a game where there's only one winner and everyone else loses. If another player is out in front in Civ, you'd better either overtake them peacefully or take them down violently, or you just lost. The last thing you should be doing is co-operating with them and making them even more likely to win.
|
# ? Aug 5, 2016 07:54 |
|
having the highest population doesn't necessarily mean you are winning
|
# ? Aug 5, 2016 08:21 |
|
Aerdan posted:having the highest population doesn't necessarily mean you are winning It's still stupid that they wrote an AI procedure to detect when someone is doing well on an important metric in the game and reward them for doing so. That's rich-get-richer, and there's no logical reason for an AI opponent to do it.
|
# ? Aug 5, 2016 08:27 |
|
Gort posted:Your real-world analogy is not relevant to a game where there's only one winner and everyone else loses. If another player is out in front in Civ, you'd better either overtake them peacefully or take them down violently, or you just lost. The last thing you should be doing is co-operating with them and making them even more likely to win. Maybe people don't want the AI to play to win, because it's super boring to play a game where everyone gangs up on you the moment you take the lead and diplomacy doesn't actually matter.
|
# ? Aug 5, 2016 08:28 |
|
I'm a fan of this new AI style. I've always played CIV with victory conditions turned off.
|
# ? Aug 5, 2016 08:30 |
|
Kajeesus posted:Maybe people don't want the AI to play to win, because it's super boring to play a game where everyone gangs up on you the moment you take the lead and diplomacy doesn't actually matter. I don't think we've actually seen a Civ AI that plays to win. Mostly they don't know about victory conditions and at best will stumble into a space victory eventually. Normally the AI "wins" by causing the player to ragequit.
|
# ? Aug 5, 2016 08:30 |
|
I want the AI to play to win. But if most players want the AI to roleplay then they should gear it towards that. If the game is going to be as moddable as they claim it will be, then hopefully I'll be catered for by the modders - lord knows there's enough nerds on CivFanatics willing to make a change like that. I would imagine it would be an easy mod to make too: just delete the diplomatic modifiers that make no sense, and add two new ones - a negative modifier for whoever is in the lead, and a positive modifier for whoever isn't (strength scaled according to degree of difference). That ought to foster the crab-bucket mentality. I guess a lot of people find that boring but I enjoy the challenge of working out how to be the first crab out.
|
# ? Aug 5, 2016 09:14 |
|
Gort posted:It's still stupid that they wrote an AI procedure to detect when someone is doing well on an important metric in the game and reward them for doing so. That's rich-get-richer, and there's no logical reason for an AI opponent to do it. The logical reason for this is because is because the computer isn't trying to beat you at the game, it's trying to make it more interesting. If it wanted to win the game would just be a millennium-long perpetual war between whomever is next to each other on the scoreboard.
|
# ? Aug 5, 2016 09:32 |
|
Another factor is that you're unlikely to be able to "fulfill" all of the AI civ's agendas at the same time. Several of them seem to require you to go into wholly different directions, and I wouldn't be surprised if some were outright mutually exclusive. So some AIs might like you by default, but at the same time others will probably also dislike you as well. In that sense, it's actually not all that different to a bunch of human players. You might be able to play nice with some of them, perhaps because you have aligning short-term goals or just because you get along well with the player, but others will probably be more wary of you because you're in a threatening position, or perhaps just because they think you're kind of a dick. So in a way, the agenda system just replicates such a system of semi-natural faction alignments and power blocks. It just does so in a somewhat more flavourful way, rather than basing it purely on strategical factors.
|
# ? Aug 5, 2016 09:52 |
|
JeremoudCorbynejad posted:I want the AI to play to win. But if most players want the AI to roleplay then they should gear it towards that. It could be both. Maybe near the mid game, the game can pick one or two AIs who have a shot at winning and turning off their "roleplay modifiers", while letting the rest be the diplomacy puzzles that the community wants. What annoys people is not ANY AI playing to win, it's diplomacy just shutting down altogether.
|
# ? Aug 5, 2016 12:12 |
|
Baronjutter posted:Sorry for Civ5 posting but I've been trying to play on "terra" maps ... Is the Ai programmed to always go "tall" ? Since I don't think anyone answered your question, the mechanics of V pretty much makes anything more than 4 cities non-optimum, and the AI has 'some' understanding of the concepts that produce that number, and so they rarely stray from half-dozen or less. Add in the extra difficulty of an overseas city and it quickly becomes 'not worth the effort'. And yes, we all absolutely hate the fact that 40% of the map isn't within anyone's borders by end game.
|
# ? Aug 5, 2016 12:43 |
|
Baronjutter posted:Sorry for Civ5 posting but I've been trying to play on "terra" maps because I had so much fun with the overseas land rush in Civ4 with the same setup, but no one loving colonizes off their home continent. Even in our starting continent there's all these very close by islands with great luxuries and no one builds more than 5-6 cities then just stops. Is the Ai programmed to always go "tall" ? Civ5 drives me nuts with how much land is empty in the late game. I really hope the world better fills up in 6. Eh, on higher difficulties the AI positively shits out cities everywhere, but yeah, some places just remain empty, because Civ5 generally encourages going tall with 4 cities.
|
# ? Aug 5, 2016 13:08 |
|
Rexides posted:It could be both. Maybe near the mid game, the game can pick one or two AIs who have a shot at winning and turning off their "roleplay modifiers", while letting the rest be the diplomacy puzzles that the community wants. What annoys people is not ANY AI playing to win, it's diplomacy just shutting down altogether. I'm thinking of a similar solution. I think that each AI should roleplay and then, at some point, only play to win the specific victory type it is most likely to win that game. That way, many AIs will still work with you and roleplay while you go for your science victory because all they are programmed to care about the culture victory, and you'll get some serious competition from the fellow science victory aimed AIs.
|
# ? Aug 5, 2016 14:11 |
|
The problem with having the AI play to win is that computers are terrible at that kind of long term planning compared to a human, at least without each AI civ's turn taking 10 times as long as it goes though all the permutations. It's much simpler to write an AI that behaves according to simple "if x then y" roleplay rules than one that can strategize, alter plans as obstacles are met, recognize advantageous opportunities and everything else a human would do to win.
|
# ? Aug 5, 2016 15:18 |
|
Christo posted:The problem with having the AI play to win is that computers are terrible at that kind of long term planning compared to a human, at least without each AI civ's turn taking 10 times as long as it goes though all the permutations. It's much simpler to write an AI that behaves according to simple "if x then y" roleplay rules than one that can strategize, alter plans as obstacles are met, recognize advantageous opportunities and everything else a human would do to win. There are some straight forward "if x then y" rules that could be used for an AI programmed to win, though. For example, "if player is leading in tourism, don't open borders to them". Or "if spaceship parts are available, then build them, stupid" Which is just as simple as, but better than, "if population then be impressed" The AI doesn't have to be Deep Blue to achieve the lofty goal of not undermining its own victory.
|
# ? Aug 5, 2016 15:39 |
|
Also, very few players actually want to be playing next to a completely handicapped nation flailing around making GBS threads itself no matter how well it roleplays its diplomacy. In order to satisfy players they're going to have to program a certain level of baseline competence with the game system anyhow, which is the main thing "play to win" players actually want. Programming a high level competitive AI wouldn't really be feasible but most people would settle for an AI that understands what its long term goals are and how to work towards them, even if it can't really keep up with a human.
|
# ? Aug 5, 2016 16:26 |
|
Every time someone says how easy it would be to make an AI that plays Civ well, I have to wonder what reason they think Firaxis would have for making an awful AI. Why would they put in a bad AI if writing a good one is easy?
|
# ? Aug 5, 2016 16:28 |
|
TooMuchAbstraction posted:Every time someone says how easy it would be to make an AI that plays Civ well, I have to wonder what reason they think Firaxis would have for making an awful AI. Why would they put in a bad AI if writing a good one is easy? A lot of it is making a game your AI can play, rather than the other way around. I remember playing Medieval Total War 2, where an AI attacking a city had to build siege weapons (ladders, towers etc) and roll them up to specific places in order to attack. Unfortunately it got confused doing this a lot, and often ended up running up-and-down in front of the infinite-ammo arrow towers, taking hideous casualties. Everyone said the AI was terrible as a result. Fast forward to Shogun Total War 2, where they made a big selling point out of the great AI. They got rid of most of the siege engines - instead troops could just climb walls with grappling hooks they all carried. The AI performed a lot better as a result. When they went from Civ 4 to Civ 5, they introduced a lot of elements that effectively punished the AI. Balance was generally poor, which meant the AI fell into traps. One-unit-per-tile combined with units being pretty slow meant that it could waste a lot of turns shuffling its units around. Being able to destroy units with no retaliation (ranged units/air units/cities) meant a player could punish them for this, and the poorly-balanced promotions on such units made it even worse. High difficulties gave the AI more units instead of better ones, exacerbating the unit traffic jam problem. Terrain could be very punishing, such as water which multiplies the damage a unit takes when it's embarked. There were tons of things they could have done to make the game easier for the AI to play it, but I don't think they were particularly committed to supporting the game - patches were very slow in coming, and dried up years before the release of the next game in the series.
|
# ? Aug 5, 2016 16:39 |
|
Gort posted:Your real-world analogy is not relevant to a game where there's only one winner and everyone else loses. If another player is out in front in Civ, you'd better either overtake them peacefully or take them down violently, or you just lost. The last thing you should be doing is co-operating with them and making them even more likely to win. This is the problem that people have with the Civ 5 AI. It's not that it "plays to win", but rather that it's programmed to be antagonistic to the player even when it doesn't make sense to be. Yeah, it's a video game with victory conditions, but it also has to have some grounding in realistic expectations. I think you might also be overstating how good a "hey bro nice work" message is.
|
# ? Aug 5, 2016 16:43 |
|
Gort posted:A lot of it is making a game your AI can play, rather than the other way around. I remember playing Medieval Total War 2, where an AI attacking a city had to build siege weapons (ladders, towers etc) and roll them up to specific places in order to attack. Unfortunately it got confused doing this a lot, and often ended up running up-and-down in front of the infinite-ammo arrow towers, taking hideous casualties. Everyone said the AI was terrible as a result. Fast forward to Shogun Total War 2, where they made a big selling point out of the great AI. They got rid of most of the siege engines - instead troops could just climb walls with grappling hooks they all carried. The AI performed a lot better as a result. This is a smart post. And yeah the main difference between the AI's combat ability in Civ4 vs Civ5 is that the AI could handle "stack troops --> move troops" but was totally incapable of managing the Civ5 system. Designing game systems with your AI's capability in mind is very important.
|
# ? Aug 5, 2016 16:44 |
|
I've been playing a multiplayer card game a lot recently, and cozying up with the strongest player before betraying them at the end is one of the most common strategies used by actual humans trying to win. It makes way more sense than attacking the strongest player and dying immediately.
|
# ? Aug 5, 2016 17:45 |
|
Clarste posted:I've been playing a multiplayer card game a lot recently, and cozying up with the strongest player before betraying them at the end is one of the most common strategies used by actual humans trying to win. It makes way more sense than attacking the strongest player and dying immediately. What form does "cozying up with the strongest player" usually take?
|
# ? Aug 5, 2016 18:09 |
|
As someone who totally ignores victory conditions and just enjoys playing and seeing what narratives form I just want the AI to behave like other countries. I wish we were much more beholden to our people or internal groups and their opinions and the direction we took out countries was much more based on that rather than some pre-planned optimal victory strategy. Like going to war with a long time ally would get your people extremely upset, betraying a country you have rich trade ties with would see your merchant/capital class potentially revolt. I guess I'd just love to see more eu4/paradox style diplomacy and internal politics where everyone is guided by their own politics and goals rather than meta-game level "victory conditions". But the AI should absolutely realize if anyone is getting too powerful or too ahead score wise. In EU4 there's a whole aggressive expansion and coalition system where if you start to become powerful all the countries near you will make a big anti-you alliance. It's a great way of stopping snow-balling victories, and it's a great way as a player to work together with the AI to unseat another AI that's too far ahead. I was just playing a CivV game where the dutch had some ridiculous run-away lead after conquering a neighbour early on. Me and everyone else were sitting around 500 score while they had 800. Then they went and declared war on another weaker neighbour while everyone else sat around. Then they were sitting at 1000 while everyone else was 500, yet there were lots of us bordering him. In EU4 every single player near him would have been in an anti-dutch coalition which would have been instantly triggered the moment he attacked someone, or the moment the coalition decided it was time to take him down a peg. I tried to form an alliance like this myself but none of the idiots wanted to go to war against the dutch with me. At the same time you can be a huge powerful country and people won't hate you because of it, only if you're a threat/competition. Other great powers will be competitive/scared of you, but smaller countries often end up deciding just being friends is the best strategy, and if you have a good reputation for not eating your friends this can be quite useful, allowing you to peacefully vassalize and even annex smaller powers. Baronjutter fucked around with this message at 18:26 on Aug 5, 2016 |
# ? Aug 5, 2016 18:20 |
|
Baronjutter posted:As someone who totally ignores victory conditions and just enjoys playing and seeing what narratives form I just want the AI to behave like other countries. I wish we were much more beholden to our people or internal groups and their opinions and the direction we took out countries was much more based on that rather than some pre-planned optimal victory strategy. Like going to war with a long time ally would get your people extremely upset, betraying a country you have rich trade ties with would see your merchant/capital class potentially revolt. I guess I'd just love to see more eu4/paradox style diplomacy and internal politics where everyone is guided by their own politics and goals rather than meta-game level "victory conditions". Same here. I usually play Civ 5 while Im still having fun and the drop it when it starts getting boring (usually around the information era). I rarely finish my games. And when playing like that is very annoying that the AI will gang up on you and irrationally hate you just because you are doing well
|
# ? Aug 5, 2016 18:24 |
|
It basically sounds like you are just saying "I want more Paradox games".
|
# ? Aug 5, 2016 18:25 |
|
I think people are over estimating the impact agendas will have on the AI, especially on victory conditions.
|
# ? Aug 5, 2016 18:35 |
|
I've said it before, but I think a lot of it could be solved at the diplomacy screen and the way the AI reacts to players requests or communications. Sure mechanically they should bring back vassalization and remove some of the warmonger stuff, but just having the AI react to the Y/N things the player does. If I'm stomping the AI and they are down to 2 cities (2 cities I DON'T want mind you), have their entire army destroyed, and just want htem to stop hoarding all the oil, they STILL will refuse to trade 20% of their oil for 5000 gold. Why? Why is the AI like this? Granted, that is an extreme example, but I feel like other games like Galactic Civ, Endless Legend, and yes, Paradox games, allow the AI players a bit more nuance when dealing with diplomacy. They won't chop off their head to spite their nose, so to speak. Allow more cooperation. Don't necessarily have the AI play to win, but have them play to not-lose. Sure they aren't going to go exactly 4 cities and have perfect civics, but if someone is getting huge, have a few variables in there that make them less likely to roll over-and less likely to suicide for no reason.
|
# ? Aug 5, 2016 19:07 |
|
Jastiger posted:less likely to suicide for no reason. Holy poo poo this. In my last civ5 game I had some random 300 score country suddenly declare war against two 500ish score countries and quickly get destroyed. Why would they do that? What AI code would possibly tell them this is a good thing to do? The AI should never declare war unless them plus their allies have more military power than their target.
|
# ? Aug 5, 2016 19:19 |
|
Chomp8645 posted:It basically sounds like you are just saying "I want more Paradox games". If there's a strategy game maker other strategy game makers can learn mechanics from, it's Paradox. Civ games tend to be a shitload more friendly and accessible though.
|
# ? Aug 5, 2016 19:34 |
|
I would love it more developers took some ideas from Paradox but Jesus do not include their loving business model in that process.
|
# ? Aug 5, 2016 19:36 |
|
Chomp8645 posted:I would love it more developers took some ideas from Paradox but Jesus do not include their loving business model in that process. What do you mean?
|
# ? Aug 5, 2016 20:11 |
|
Chomp8645 posted:I would love it more developers took some ideas from Paradox but Jesus do not include their loving business model in that process. I just had to buy about 8 DLC's in order to have the correct content to use most of the big civ 5 mod packs. Everyone does a million DLC's these days.
|
# ? Aug 5, 2016 20:12 |
|
Baronjutter posted:I just had to buy about 8 DLC's in order to have the correct content to use most of the big civ 5 mod packs. Everyone does a million DLC's these days. If it's a DLC complaint, keep in mind stuff like being unable to multiplayer Civ 5 if all players don't have the same DLC, while with Europa 4 everyone gets the DLCs of the player hosting. Much more generous, and makes a lot more sense since otherwise you're splitting your multiplayer playerbase.
|
# ? Aug 5, 2016 20:14 |
|
|
# ? May 14, 2024 16:51 |
|
Is there a thread (or even a video tutorial) to learn Crusader Kings II? I love the idea of the game, I love Civ 5 and the Total War series, but I tried playing once and had no clue and just quit.
|
# ? Aug 5, 2016 20:20 |