1. We have added a Gift Upgrades feature that allows you to gift an account upgrade to another member, just in time for the holiday season. You can see the gift option when going to the Account Upgrades screen, or on any user profile screen.
    Dismiss Notice

MADDJINN GAMES, AI problems

Discussion in 'CivBE - General Discussions' started by jokii, Oct 20, 2014.

  1. jokii

    jokii Chieftain

    Joined:
    Jan 8, 2014
    Messages:
    85
    https://www.youtube.com/watch?v=Kj90R0RWvdQ
    06:55 AI naval fight-suicide

    https://www.youtube.com/watch?v=bzu8u3jPt10
    09:26 taking shots on naval units which will not fire back
    16:29 taking shots on naval units which will not fire back
    33:46 AI naval fight-suicide

    When I saw MadDjinn naval fight in Pre-release build Kavithan Protectorate, I couldn't believe it.....???
    AI can not move and shoot ??? AI BUG - again ???.... after 4 years, two expansion packs, and brand new game...?
    And there is a mod which repaired that same bug for Civ 5 out there...???
    Not to mention all other AI problems... I know that AI coding is very hard, so I did not expect much, BUT THIS ???
    And ships have low defensive strength so you can kill them with one shoot after they move near you...???
    WTF ? Is this a joke ???
    As always, game looks great, but AI is the most important part of the strategy game. Apparently guys from Firaxis games don't agree....
    Luckily there is a multiplayer...Without multiplayer this game would be a junk...
     
  2. Eisenhart

    Eisenhart Warlord

    Joined:
    Aug 18, 2014
    Messages:
    195
    Location:
    NCR, The Philippines
    Pete Murray said on Twitter that they definitely can move and shoot. o_O

    Maybe 'cause that's a non-release build?
     
  3. Miravlix

    Miravlix King

    Joined:
    Jun 28, 2012
    Messages:
    645
    The 1000$ question is, do you think they work on the AI first or are they knee deep worrying about actual code?

    Since you decided to post, I take it you're one of the people that got the answer wrong.

    The AI is the last thing they work on, we won't know until release day if they managed to devote time to it or it will come in a post-release patch.
     
  4. Rakshasa72

    Rakshasa72 Warlord

    Joined:
    Jun 9, 2012
    Messages:
    149
    "No plan survives contact with the Enemy."

    The "Enemy" in this case is the players. You can expect that the players will always find weaknesses in the AI no matter how much work you put into it pre-launch. So I imagine there will be plenty of things to fix post launch.
     
  5. Velasti

    Velasti Warlord

    Joined:
    Sep 26, 2014
    Messages:
    103
    The AI affects my enjoyment of the game much more than a bunch of other stuff.

    Playing the AI doesn't feel like playing a real opponent. You don't have to scout and find out what the AI is up to.

    http://christophermpark.blogspot.co.nz/2010/05/designing-emergent-ai-part-6-tempo-and.html

    This blog kind of alludes to the idea of tempo, and in CiV, the AI is designed to give the player the tempo in the vast majority of situations.

    But the issue in PVP-designed AI is you have to figure out what the enemy is doing. That's the idea behind threat mapping and so on

    I disagree with Chris, in PVP RTS games, the best players scout to find out what the other is doing, and adapt their strategy based on what they find. The best RTS games are the ones that allow you to throw a monkey-wrench at the other player out of nowhere.

    The greater factor is probably that the developers are very good at developing games, and not actually highly skilled players.

    Thinking about Sins of a Solar Empire for a second - the way the AI (and hence the developers) play the game is a completely different way than the way the best players play the game.

    AI builds lots of siege frigates in their fleet. Building siege frigates is suggested in the tutorial also.

    Players spend the same resources on cheap damage and cheap health, and rely on capital ships solely to bomb planets. Because siege frigates aren't worth the opportunity cost.

    Developers aren't always the experts at their game.

    ~~~
    I think one thing to remember, for people serious about AI programming, is that there are more lucrative fields than the games industry.

    Not if the AI is emergent or a genetic algorithm.

    The "enemy" in this case is the developers who are failing to embrace innovation.

    you know how the AI always targets damaged units near cities?

    How much do you abuse that? I build sacrificial units to soak fire.

    Any kind of adaptive AI and that wouldn't be a problem. I target the unit that is posing the most threat to the city.


    Decision based AI are easy to fool.

    Granted I have no idea how the AI in civ actually works.. hmm.. maybe I'll ask in the moddng subforum
     
  6. Smith of Lie

    Smith of Lie Chieftain

    Joined:
    Jul 24, 2013
    Messages:
    77
    As someone who has absolutely no experience in programming and especially AI programming, past reading some pop-science articles, how likely is such thing to be created for a video game?

    From what I understand most game AIs are building a long checklist of IF X THEN Y and setting priorities. At least at the very basic level. How does emergent or genetic algorithm compares to this?

    I'm curious how much work compared to what is estimated to have gone into writing the BE AI would go into making a "learning one".

    I also do remember reading a news snippet about a university research group (at least I think it was one of those, can't remember the details as it was back in CiV vanilla days) having written an AI that played CiV as if it was player (id est. under the same ruleset) and was consistently beating the in game AI (I assume on Deity, but difficutly was not mentioned). One wonders how different it was from the game AI and if the methods used could help in creating less silly AI for an opponent.
     
  7. jaldaen

    jaldaen Prince

    Joined:
    Feb 19, 2006
    Messages:
    467
    As Miravlix said, AI is the last thing many developers work on in a game. Why? Because the rules aren't set until the game reaches beta and if you program the AI before the rules are set, then if you change any of those rules due to player feedback, you have to go back and change the AI, which means a lot of wasted work. Instead, it is much easier to get all the rules in place and then hammer out the AI at the end.

    The main question will be how well does the AI handle the rules at launch. We only have a few more days to find out.
     
  8. Pyramid Head

    Pyramid Head Chieftain

    Joined:
    Sep 25, 2010
    Messages:
    79
    It's Firaxis. "Graphics and DLCs are more imprtant than the code" is their motto

    Moderator Action: Please do not troll the forums. We welcome all opinions, when stated in a civil way with facts to back it up.
    Please read the forum rules: http://forums.civfanatics.com/showthread.php?t=422889
     
  9. Aeson

    Aeson orangesoda Retired Moderator

    Joined:
    Nov 16, 2001
    Messages:
    2,686
    Civ AI (in the past) has some emergent functionality. Most of the economic decision making relies on personalities and weightings and a random element. They aren't told to build X, Y, Z according to a script ... they are told to value things (military, science, culture, etc) and then may choose to build X based on what they value and whatever situational pressures are factoring in at the time.

    Of course that doesn't mean it will be any good.

    Tactically the AI is more reactive. If/then type of decisions at least in many cases.
     
  10. Atepa

    Atepa Chieftain

    Joined:
    Oct 7, 2014
    Messages:
    4
    Location:
    Ontario, Canada
    While some are built that way, others use a weighting system now days (I can't remember who at the moment though). E.g. for worker AI it would evaluate the city, the tiles around it, any bonuses those tiles do & can provide, then it would do that same logic for each other city and at the end of the calculation it would have a value for all of it's moves and randomly pick one of the top 5 or such.

    On a higher level the AI computers do the same thing and based off difficulty they choose from a smaller and smaller pool of 'good' answers. So a Settler difficulty AI may choose from the top 20 for an action. The Deity AI may choose from the top 5.


    Another factor in AI intelligence is time as well. The smarter the AI is, the longer it takes between turns, for them to finish up because the more calculations and factors it is considering and evaluating. IIRC the Civ AI only does some of its calculations during the player's turns, most is done in the off-cycle which means they have to be quick about it so that players aren't complaining about the AI taking 5-10 minutes to run through their cycles.
     
  11. Ryika

    Ryika Lazy Wannabe Artista

    Joined:
    Aug 30, 2013
    Messages:
    9,395
    That's heavily dependent on optimization though. If the AI didn't move every single unit during peace time and instead knew how to use "scouts" to secure its borders, then I don't think that would really make much difference in the amount of processing needed, but it would surely make the AI seem less stupid - 3/4 of their units would become "passive", while 1/4 does more efficient work.

    I'm still waiting for the Technological Singularity to occur though. Machines will create games with AIs that are just unbeatable and... well, then probably kill all humans for being worthless bags of meat. But still, enjoying the 18,35315 Seconds of fun before annihilation would totally be worth it.
     
  12. Velasti

    Velasti Warlord

    Joined:
    Sep 26, 2014
    Messages:
    103
    I read that article TODAY :)

    Such a thing has already been created for some video games. Michael Robbins used the concept of a genetic algorithm for supcom2.

    Ternary vs Binary

    What you have to realise is a humans runs not in binary, but in ternary.

    The sky is blue.

    One day, you'll wake up, you'll see that the sky is green.

    The fact that the sky is blue is irrelevant. You can see for yourself that the sky is green.

    In ternary, there are 3 values - true, false, and neither true nor false.

    The human brain is constantly able to reassess its known facts. It is also able to synthesise information from many different sources to create new information.


    IF X THEN Y


    Let's take an example from the article I linked - let's say you're attacking my base in Age of Empires.

    You attack my base. I have a weak point in my wall - specifically, the wall is not compete.

    So you attack through the hole in my wall.

    My entire army is defending the hole in my wall. Nothing else is defended.

    You lose. In the future, you'll know not to attack the weak spots in my wall. You'll make your own hole.

    The computer in this situation however, is hardcoded with "If X, then Y".

    So once you've figured out that the computer will always attack the weakspot of your wall, you can always defend it. This is an exploit.

    Genetic Algorithm/Historical Information
    A genetic algorithm works by natural selection. It creates a genepool. Those who are successful succeed. Those that aren't, don't. The successful genes are passed to the next generation.

    So in the above scenario, natural selection would favour AI's that didn't attack the hole in the wall. And eliminate the rest.

    You see the problem with that approach? It's the same problem that is present in real-life biology.

    Walking through the weakspot of the wall is actually a really good decision, if I don't realise the weakspot is there

    The other problem is that you have to run training sessions for the AI. This would be excellent if it was done by every player world wide, but often the AI is just pitted against other versions of itself.

    This gives the AI the ability to learn from the player's style. Imagine an AI that could learn how to play the game from the best players.

    Emergent AI

    Emergent AI is unpredictable. It doesn't always do what is best, or what is optimal. It's creative and does its own thing.

    He means an emergent AI in the sense of an unpredictable AI. Essentially, each individual has its own intelligence, and does what is best for itself but takes into account what he larger group is doing. It has a subcommander so that it doesn't act completely randomly. .

    With this approach the decisions that the AI could make are ranked, then it applies fuzzy logic to make its decision (if a decision is predictable, it's not the best decision - think targeting in CiV - always targeting damaged units)

    The difficulty from the game comes because of the fuzzy logic. The easy AI's are more likely to make mistakes that novice players might make.

    There's an argument that unpredictable, sub-optimal decisions are better in the long run than predictable, optimal decisions. It prevents the AI from being baited or trapped.


    Basically, we're playing a strategy game, right?

    Strategy is exploration, not exploitation. There is no such thing as "optimal" strategy. Any kind of "optimal" strategy demonstrates a balance discrepancy.

    Of course, in strategy games there tends to be a favoured strategy, which creates favoured ways to counter it. But there is a very real difference between favoured strategy and optimal strategy.

    There is no best technique. Decision trees can be used to fix problems that the learning AI is having. Emergence has an unpredictability that could be fatal. Genetic Algorithms can form a predictability that could be fatal.

    https://www.youtube.com/watch?v=WXd6CQRTNek - this kind of illustrates the point of exploration and exploitation.

    The biggest issues with the AI is the tactical AI. Although some behaviour to learn how the player manages to keep up with the AI on Deity could be useful
     
  13. Velasti

    Velasti Warlord

    Joined:
    Sep 26, 2014
    Messages:
    103
    I would like a way to increase the allowable turn length for the AI. Having to wait for the other player to make turns would probably be good for me.


    Time is probably a rather minor factor. You know how many cycles there are in a minute? Those calculations can be run on their own thread. Again, refer to the emergent AI of AI War, which is based on large amounts of calculations.

    There is almost definitely more time taken up by the animations than would be taken up by the AI thinking.
     
  14. DarkestOnion

    DarkestOnion Prince

    Joined:
    Jul 27, 2014
    Messages:
    533
    I think it is a bit of a stretch to ask a game developer to develop a PvE AI which plays at the same potential level as another human being in PvP. Of course, we all want a challenge...but we might want to make more reasonable/acheivable AI requests too.
     
  15. GeoModder

    GeoModder Deity

    Joined:
    Nov 29, 2005
    Messages:
    7,512
    OP seems to only start threads about AI behaviour...
     
  16. SupremacyKing

    SupremacyKing Prince

    Joined:
    Jun 19, 2014
    Messages:
    564
    I think that is very harsh and unfair. "junk" implies no value whatsoever. I think thousands of gamers will find plenty of value in CivBE even if they never play MP. There is certainly a lot to give value to CivBE from the new affinities, the quests, the building choices, the tech web, and the covert ops etc.
     
  17. Strategist83

    Strategist83 King

    Joined:
    Dec 30, 2005
    Messages:
    625
    The 1,000,000$ question is why you're on this veritable crusade to apologize for what flaws the game has (and apparently quite willing to do so in a rude manner).

    Everybody here is excited for BE. That doesn't mean the game can't be criticized - and should be. OP brought up some entirely relevant points. This is an established studio, professional programmers working on a supposedly A+ title that will sell for millions of dollars (have you considered how many people will be purchasing the game and therefore how much money is involved? Civilization games aren't even in the same league as other games here so it is only reasonable to expect a better product). Yet, it is somehow acceptable if the game ships with the exact same well known fallacies that plagued Civ V's AI? We aren't even talking about truly good AI, OP only asked for it not to commit the same obvious mistakes once again. Which it does - right in your face. It isn't even subtle.

    Mad's ARC LP has me concerned, as well. In episode 10 he attacks the Slavic Apollo AI, and... it isn't even a battle, it's a slaughter. And then... the Brazilians bombard the Slavic capital down to no hitpoints, then abandon their attack(sic!) and leave the human player to capture it. Afterwards, Mad launches an assault on Franco-Iberia which likewise puts up absolutely no resistance. Mad's gameplay showcases a general contempt for the AI's ability (neglecting military etc) which is how you would play Civ V also when realizing the AI doesn't represent any actual threat at all.

    If Mad's LP videos are representative of the final product then we can expect the genre-leading franchise to have made ZERO progress in the AI field. But hey! There's new graphics and music! And the tech is now a web rather than a tree! Now available for pre-order on Steam for only 50€... Praise Firaxis! Nobody dare raise a critical voice! :rolleyes:
     
  18. Zet

    Zet Warlord

    Joined:
    Jun 23, 2012
    Messages:
    212
    @Strategist83 that is a good post. The AI exhibits the same problems as in Civ V.

    I've had the suspicion that this is a symptom of 1UPT.
     
  19. TheGrumpyBuddha

    TheGrumpyBuddha King

    Joined:
    Aug 8, 2013
    Messages:
    822
    Anyone know whether they fixed the White Peace problem? One that got added, somehow, it wasn't an issue in the earlier versions of Civ V.
     
  20. Lord Shadow

    Lord Shadow General

    Joined:
    Oct 14, 2005
    Messages:
    2,001
    Gender:
    Male
    Location:
    Buenos Aires
    It all boils down to this expectation...

    And I can't think of any complex strategy game in which the AI plays like a "real opponent". You keep bringing up examples in blogs which generally talk about real-time strategy examples, which are simpler and also have the advantage of allowing the AI to constantly compute. People already complain about the length of turns by the late game. RTS AI opponents also generally cheat as well in the higher levels, or in AI War's particular case, they start in an overwhelmingly superior position and the game just gradually removes restrictions. Intelligence-wise, they aren't particularly advanced.

    Really, thinking about it, AI War isn't a valid example in any way, given how it's an asymmetrical strategy game in which the AI doesn't play at all like a human. Also following completely different rules.

    I'm not saying it's an unworthy goal to aim for exceptional AI, not at all, but rather that it's unrealistic to demand it from any complex game such as Civilization at this point in time. Functional? Of course, but any standard AI will have holes for humans to exploit, and one can mainly keep plugging them and in check. AI is just one of many areas a good game needs to develop, and pulling it off more than just well requires a seriously disproportionate amount of resources.


    PS: Don't bring up GalCiv2 because that's still a simpler game than any Civ, being completely devoid of geography.
     

Share This Page