A.I. - The Good, The Bad, The Intelligent - Printable Version +- Universal Gaming (https://universalgaming.net) +-- Forum: Gaming Galaxy (https://universalgaming.net/forumdisplay.php?fid=1) +--- Forum: General Gaming (https://universalgaming.net/forumdisplay.php?fid=2) +--- Thread: A.I. - The Good, The Bad, The Intelligent (/showthread.php?tid=165) |
RE: A.I. - The Good, The Bad, The Intelligent - lp0 on fire - Jul 17th, 2018 (Jul 17th, 2018, 02:21 AM)Nightingale Wrote: It was and it was also was/is toxic as all heck, and I like my vidya gaemz with some trash talking bullshit. Also you just did make 300 now RE: A.I. - The Good, The Bad, The Intelligent - Nightingale - Jul 17th, 2018 (Jul 17th, 2018, 02:25 AM)percussive maintenance Wrote: But how's the AI on LoL? xD RE: A.I. - The Good, The Bad, The Intelligent - lp0 on fire - Jul 17th, 2018 I dont even know if LoL has AI as I don't play MOBAs RE: A.I. - The Good, The Bad, The Intelligent - Nightingale - Jul 17th, 2018 (Jul 17th, 2018, 02:49 AM)percussive maintenance Wrote: Yeah, neither do I. DOTA was going to have a voicepack by none other than Ross Scott, but they rejected it. That killed the genre for me. RE: A.I. - The Good, The Bad, The Intelligent - lp0 on fire - Jul 17th, 2018 (Jul 17th, 2018, 03:03 AM)Nightingale Wrote:Since we are dealing with Vavle here and making a game does this surprise you? RE: A.I. - The Good, The Bad, The Intelligent - Nightingale - Jul 17th, 2018 (Jul 17th, 2018, 03:41 AM)percussive maintenance Wrote:It bugs me, it's what it does. RE: A.I. - The Good, The Bad, The Intelligent - lp0 on fire - Jul 17th, 2018 I asked does it surprise you! Not if it bugs you! RE: A.I. - The Good, The Bad, The Intelligent - Nightingale - Jul 17th, 2018 (Jul 17th, 2018, 05:03 AM)percussive maintenance Wrote: I'm such a rebel RE: A.I. - The Good, The Bad, The Intelligent - Maniakkid25 - Jul 17th, 2018 I thought there was an option to play against bots in most MOBAs. Meh, I don't play them either, so I don't really know. (Jul 16th, 2018, 11:12 PM)Moonface Wrote: This is a common thing in RPGs, especially MMOs, and it's known as the enemy's "aggro sphere". The idea behind the aggro sphere is that it gives you time to plan the situation while still being in real-time, as well as being able to run away from a fight if you're outmatched. Regulating aggro, including aggro spheres, is one of the most important skills in an MMO, after all. That said, this is common to RPGs, so what the hell it's doing in a survival horror game like The Evil Within 2, I don't even know. As far as discussing the topic further goes, I've already referred to Mark Brown's episode on what makes good AI, but I'mma link it again because it's that interesting. Meanwhile, I'll mention two things that I'm sure I've mentioned before: The Combine Gunship and the AI in FEAR. I love these two examples because they're examples in how intelligence can be coded in, whether by accident or on purpose. The Combine Gunship trying to shoot down a rocket was a bug in the code, being coded to fire at the most significant threat. Note that it doesn't specify the player, just the most significant threat. The code then took the ball from there, and decided that the rocket that was now going to damage it became a more significant threat than the player, and focused fire on that. After playtesting and finding the "bug", because of playtesters positive reactions, they kept it and even refined it into what it is in the game. FEAR's Replica Soldiers (that's the name of the main enemy in the game), meanwhile, are an example of emergent behavior as far as AI coding goes. Emergent behavior, or Emergence in general, is what happens when complex and sophisticated structures or concepts follow from much simpler rules. A snowflake is a good example of it, because it's just water freezing into a crystal, but because of the rules of chemistry, it creates beautiful and complex patterns and shapes. Based on what I understood from their paper (Which you can also hear in presentation form), FEAR's AI was more intentional than that, but what I do know is that the soldiers are hard-coded with far more basic ideas: Move in groups, provide covering fire, take cover, and move relative to the player (yes, they can also jump through windows and stuff like that, but that's not the point). These parameters lead to FEAR having what is still considered some of the best AI in a video game, 13 years after the game's release. I actually bought the game myself on this alone, just because I wanted to see for myself how this AI works in practice. They didn't disappoint. Right from the start, they'll keep you on your toes, and it's amazing to think that a game from that long ago has AI where even a small squad of soldiers will always be a legitimate worry if you let them operate. As far as the bad goes...well, I'mma look at Gen 1 Pokemon for that. Yes, we all know it's a mess of a game, but one of the interesting things is that the AI is coded to always use attacks that are super-effective when possible, and to not use moves that are not very effective (i.e. does half damage), even if the move doesn't do damage. This can lead to...interesting side-effects, especially a notorious softlock with Lorelei (though you basically have to do it on purpose for it to happen) where if you Rage with a fighting or poison-type pokemon, she will infinitely use Rest with her Dewgong, causing an infinite battle. I call it notorious because Pokemon Yellow was specifically hard-coded to check if the pokemon you're fighting is Lorelei's Dewgong, and will change her AI if it is, which is a kludge if I've ever heard of one. RE: A.I. - The Good, The Bad, The Intelligent - Moonface - Jul 18th, 2018 (Jul 17th, 2018, 08:00 PM)Maniakkid25 Wrote:I can understand it being in RPG games, because it would make it ridiculous if you could be infinitely chased. Also I'm not sure if it classifies as A.I. or not, but rubberbanding in any racing game is terrible and has no valid reason for inclusion. RE: A.I. - The Good, The Bad, The Intelligent - Maniakkid25 - Jul 18th, 2018 Well, that at least has the spark of a good idea behind it: it's trying to keep the player challenged. You want to allow the player to win, but you also don't want them breezing through the game. Admittedly, the implementation is clunky at its best and absolutely annoyingly fake difficulty at its worst, but there's an honestly good idea behind it. RE: A.I. - The Good, The Bad, The Intelligent - Nightingale - Jul 19th, 2018 Quote:Also I'm not sure if it classifies as A.I. or not, but rubberbanding in any racing game is terrible and has no valid reason for inclusion. Man, you are gonna love Polaris Snocross, then The AI has the "stretchiest rubberband I have ever seen". Go two miles an hour, everyone else does xD I think the perfect AI to a racing game is that that challenges -but not obliterates- you. Like that of NFS III: Hot Pursuit. You ain't gonna beat the game in one sitting, but you will beat it... eventually. RE: A.I. - The Good, The Bad, The Intelligent - Moonface - Jul 19th, 2018 (Jul 18th, 2018, 08:57 PM)Maniakkid25 Wrote:True, it does have a good idea for it. I just rarely see it executed well. Motorstorm's AI can beat you even when you do the perfect race, and ruined the game for me. Higher difficulties should have the AI catch up to you easier, but they're not going to randomly overtake you unless you do something that warrants it such as slowing down. If you're doing nothing wrong and still get overtaken by rubberbanding, then the rubberbanding isn't good. Also, Sheva has the worst companion AI of any AAA co-op game that I've experienced. However I dare to ask if anyone knows of co-op AI in a game that is worse than her. RE: A.I. - The Good, The Bad, The Intelligent - Maniakkid25 - Jul 20th, 2018 Operation Raccoon City. At least Sheva understands "fire bad" and TRIES to not burn to death! Also, seriously, if you hate the AI in RE5, play the remaster. Her AI is seriously so much better. For one thing, she immediately reacts to threats against you, meaning you don't have to shout at her that you're about to die, heal me you dumb bitch! She'll recognize it and already have healed you (assuming you two are standing together) by the time you can even call to her. There are also other, much smaller changes that lead to me willingly saying that I'd be almost willing to take her on a Pro difficulty run. ...ALMOST. The only thing that's stopping me from doing is is I DO NOT WANT TO GRIND THAT S*** AGAIN! RE: A.I. - The Good, The Bad, The Intelligent - Moonface - Jul 22nd, 2024 (Jul 20th, 2018, 05:26 AM)Maniakkid25 Wrote:Amusingly late but only because I was reading everything in here before bringing this thread back from the dead first; this makes it sound like you have done a Professional difficulty run in RE5 at some point and if you did that with the AI and not a human companion I commend you, because I remember trying to do that difficulty with @Karo and some of the stuff I learned about in that mode (such as practically zero time to pick up a downed partner) made me reach the conclusion that doing it with an AI companion is subjecting yourself to self harm. Anyway, I saw a Reddit thread last night talking about AI in games and a comment in there mentioned Dead Rising, and god, I never had that game cross my mind when this thread was originally made. I've only played DR2 and I hated the survivor AI in that, but apparently it's significantly worse in DR1 which I can't even imagine. I wonder if Dead Rising has the most canonically useless survivors in any game universe, because absolutely none of them would have a chance in hell of surviving for even an hour let alone making it to the safehouse without the player character babysitting them. x.x |