Originally posted by solarbear88 I am not sure this will work out. Every intelligent mob will just go right for and annihilate the squishy target and in case he doesn't die they will just kite in circles. Till mob dead.
No idea if kiting will be a viable option in EQN. Mobs could simply not chase players mindlessly (hoping AI doesn't follow a player in a circle) or they could CC players if they try to run.
As far as mobs going after the squishies, this is where team work comes in. Mobs would go after the DPS in the trinity system as well.
The tank + casters having avoidance abilities are usually what happens. If a Mage is casting away and a Mob decides they don't like it, the Warrior with the shield better stop it from charging.
This could lead to a ping pong effect, but the Mobs could CC the Warrior and then go after the caster who would have to use movement/defense to stay alive and hope their Cleric buddy can get some heals in before the Mob CCs them as well.
A Mob could cast Silence on a caster or cast a buff to deflect magic for a short period or any number of things to make the DPS not so effective while they deal with the guys stabbing them in the face.
Really, the amount of complexity is only limited by dev imagination and how much they program in and allow to happen.
Then again, a Mob may find the the guy knocking them down to be a bigger threat then the guy hurling fireballs and stay on the meat shield the whole time. We don't know how it will work out, but there at least is room for ideas beyond Tank and Spank.
Smart AI has been around for ages.. It's nothing new.. AI is nothing more then programming the actions of the mob.. Most games choose to go "simple" where as other games allow "complex" or advanced AI.. This is just like playing strategy games.. Civ 5? C&C?.. Players are given options to go simple or go "god" mode if they wish.. Players will ALWAYS find the best way to expose the weakness in AI codes and use it to defeat it.. Then what? Make the code so hard that 80% of your customer base says "SCREW THIS SHIT".. and moves on.. People talk about playing a GAME for fun, but if you code the AI too high, then it's no longer fun to the masses, it's becomes more like a job..
In conclusion.. The smart AI "HAS" to remain stupid for the masses.. So why bother? Think about it..
Smart AI has been around for ages.. It's nothing new.. AI is nothing more then programming the actions of the mob.. Most games choose to go "simple" where as other games allow "complex" or advanced AI.. This is just like playing strategy games.. Civ 5? C&C?.. Players are given options to go simple or go "god" mode if they wish.. Players will ALWAYS find the best way to expose the weakness in AI codes and use it to defeat it.. Then what? Make the code so hard that 80% of your customer base says "SCREW THIS SHIT".. and moves on.. People talk about playing a GAME for fun, but if you code the AI too high, then it's no longer fun to the masses, it's becomes more like a job..
In conclusion.. The smart AI "HAS" to remain stupid for the masses.. So why bother? Think about it..
Or they could have simple to "god" mode mobs and cater to everyone? With a huge world to explore, it is reasonable to believe that some/most players will not see everything nor beat it all.
WoW's high end raids were only completed by a very small percentage of players (out of millions), yet they still had millions playing. That has since changed, but it was effective for many years. Add in numerous new things to do in the game beyond just mob grinding and raiding and people will stick around for a while.
Or just cater to only one part of the community and make everything the exact same difficulty, seems to have worked for many games...
Originally posted by solarbear88 I am not sure this will work out. Every intelligent mob will just go right for and annihilate the squishy target and in case he doesn't die they will just kite in circles. Till mob dead.
How do they know which is the squishy target?
What if they really hate Elves and your frontline guy is an Elf?
What if the Orcs you are attacking recognise 'Slinky Orcbane' the tank and ignore the others simply to kill their sworn enemy?
What is the mobs you are fighting are really stupid and really only attack the closest enemy?
What if the mobs are intelligent enough that they realise soon after targeting the squishy character that there is a bigger threat from the bug with the big longbow and so choose to swap target?
The complexity of the fights with rely on the time taken to create a good set of options for each mob to recreate their 'thought processes'. Really, time is the only reason it won't improve things.
Smart AI has been around for ages.. It's nothing new.. AI is nothing more then programming the actions of the mob.. Most games choose to go "simple" where as other games allow "complex" or advanced AI.. This is just like playing strategy games.. Civ 5? C&C?.. Players are given options to go simple or go "god" mode if they wish.. Players will ALWAYS find the best way to expose the weakness in AI codes and use it to defeat it.. Then what? Make the code so hard that 80% of your customer base says "SCREW THIS SHIT".. and moves on.. People talk about playing a GAME for fun, but if you code the AI too high, then it's no longer fun to the masses, it's becomes more like a job..
In conclusion.. The smart AI "HAS" to remain stupid for the masses.. So why bother? Think about it..
Or they could have simple to "god" mode mobs and cater to everyone? With a huge world to explore, it is reasonable to believe that some/most players will not see everything nor beat it all.
WoW's high end raids were only completed by a very small percentage of players (out of millions), yet they still had millions playing. MILLIONS played it because the general open world AI was was stupid.. The hard raids were instanced.. HUGE difference.. That has since changed, but it was effective for many years. Add in numerous new things to do in the game beyond just mob grinding and raiding and people will stick around for a while.This remains to be seen.. None of us have any insider knowledge on what non-combat game play will be..
Or just cater to only one part of the community and make everything the exact same difficulty, seems to have worked for many games... But you are asking them to "cater" to your desire of game play.. Double Standards? I have yet to hear anything from SOE about EQN cater to everyone.. EQN is going niche as well..
If EQN wished to attract the masses it MUST be programmed for the lowest denominator.. PURE and SIMPLE and never open for discussion.. If the new smart AI offends too many customers, it will result in a very small niche game, or it will be tweaked down to dummy mode.. You mention "gode" mode mobs. UMmm.. EQ did that with dragons, and GW2 does that with open world bosses and both ended up being ZERGED.. It is what it is.. As the old sales slogan goes.. EQN is trying to sell the "sizzle" , and we have yet to see the steak..... Get the idea?
The concepts may not be "new" but can you point me to a mmorpg that has them already or to the scale at which EQN may?
As with all things tech, I'm sure they are talked about and imagined long before anyone gets their hands on the finished product. Why you wouldn't be excited to see in action something you knew about 20 years ago is beyond me.
Every mmorpg uses the same methods to program the AI. "Threat generation" behavior in mmorpgs is programmed _absolutely the same way_, it's just that the mechanics of threat generations are added as rules to weighted behavior of monsters.
And it was done this way "because reasons" - because without additional rules and tweaking emergent AI quite often emerges either stupid and easily taken advantage of (say, a guy in a dress and with a magic wand in his hands gets "haste" from group-members and runs around screaming, while monster trying to attack him and everyone else just pounds the monster without any retaliation) or becomes boringly genius - say, the monster and all his minions focus their attacks on the healer no matter what, and it's a 100% wipe for a group, no way to overcome.
I see in action that "utility AI" in a lot of games and it's absolutely nothing special - it's how you implement rules what's important... and the descriptions given here are also nothing special: that's how every game programmer tries to implement his AI... but it has lots of limitations.
So, I'm not excited because what I read here is "we are going to make the same AI as everyone else have done everywhere, but ours will be totes better because reasons!" So no, I'm not excited until I see it.
If EQN wished to attract the masses it MUST be programmed for the lowest denominator.. PURE and SIMPLE and never open for discussion.. If the new smart AI offends too many customers, it will result in a very small niche game, or it will be tweaked down to dummy mode.. You mention "gode" mode mobs. UMmm.. EQ did that with dragons, and GW2 does that with open world bosses and both ended up being ZERGED.. It is what it is.. As the old sales slogan goes.. EQN is trying to sell the "sizzle" , and we have yet to see the steak..... Get the idea?
I don't think there is any doubt they are going for easy mode game. Current game scripted mobs had a set of actions to do, and always did the same actions with the same intensity. The only difference between this and "intelligent AI" I see is that the new mob can make mistakes.
I can already see billions of ways to abuse that system. Like taking a weak character into party or naked and fast character, to make the difference in armor so big, that the monster AI will keep choosing him every few secons as a new way of taunting. It is that tag system. So easy to create your own environment, in which monster will keep choosing the same actions because of the differences in stats you throw at him. Unless they make monster retarded enough to not take such differences into consideration, then it is even easier.
Originally posted by Enrif watch at 2:00 and at 4:15 and stop assuming stuff they didn't said.
It is just a bunch of promises. We are trying to discuss about it from technical side, and see how it is really possible to achieve. If everything developers said about their game in hype phase was real then every game released so far would look completely different... It is like taking seriously election promises.
Wow... is it always like this around here? Some of you folks amuse me, some inspire me, some sadden me. And some of you... well... I don't know what to think.
Regardless, thanks for all the interest -- even the speculation and skepticism. Wish I could tell you more.
Soon.
President & Lead Designer of Intrinsic Algorithm, Author of Behavioral Mathematics for Game AI, Game AI consultant, GDC AI Summit advisor, co-founder of AIGPG | IntrinsicAlgorithm.com
I'm sure devs could think up even more sophisticated systems to keep everything in check. Unlike GW2 (at least when I played) where you only had to do a small amount of damage to get credit among a huge zerg of players.
SOE, EA-Mythic, and others have made past efforts regarding how to measure your input for the quality of loot
Wow... is it always like this around here? Some of you folks amuse me, some inspire me, some sadden me. And some of you... well... I don't know what to think.
Regardless, thanks for all the interest -- even the speculation and skepticism. Wish I could tell you more.
Soon.
Yeah it is like this pretty much all the time.
Some people see possibilities, some lack what it takes to see them.
Thankfully the short-sighted people are not making games, they just clog up these forums from time to time.
Originally posted by Enrif watch at 2:00 and at 4:15 and stop assuming stuff they didn't said.
It is just a bunch of promises. We are trying to discuss about it from technical side, and see how it is really possible to achieve. If everything developers said about their game in hype phase was real then every game released so far would look completely different... It is like taking seriously election promises.
If given the choice between believing a group of proven industry vets or a bunch of negative ill-informed forum dwellers then I am very excited for the future of this game and all future games.
Still waiting to see this in action over time. The fact that all of this has to be written in code and things will still have triggers lead me to believe you'll still have scripted fights on a basic level. It's still an algorithm, which is inherently "step-by-step" thus being predictable.
I'd be more worried about their reasoning for all this. They honestly think tanks, healers, and support don't have fun and would rather be DPSing.
Still waiting to see this in action over time. The fact that all of this has to be written in code and things will still have triggers lead me to believe you'll still have scripted fights on a basic level. It's still an algorithm, which is inherently "step-by-step" thus being predictable.
I'd be more worried about their reasoning for all this. They honestly think tanks, healers, and support don't have fun and would rather be DPSing.
Exactly.. People can keep calling it whatever they want.. CODE is CODE.. The fight will either be scripted by code, or it will be random.. If it is scripted by code then players will learn that code and use it against the mobs.. So we are right back to square one...... If the fights are random, then smart AI is just smoke and mirrors.. Toss this in with all the twitch non tab targeting.. I would be greatly surprised to see medium size audience after a year..
Still waiting to see this in action over time. The fact that all of this has to be written in code and things will still have triggers lead me to believe you'll still have scripted fights on a basic level. It's still an algorithm, which is inherently "step-by-step" thus being predictable.
I'd be more worried about their reasoning for all this. They honestly think tanks, healers, and support don't have fun and would rather be DPSing.
I believe its the opposite. DPS alone is not a role. Everyone gets to do that in addition to other things at the same time.
you pretty much have to do it that way these days because so few play the dedicated roles.
Still waiting to see this in action over time. The fact that all of this has to be written in code and things will still have triggers lead me to believe you'll still have scripted fights on a basic level. It's still an algorithm, which is inherently "step-by-step" thus being predictable.
I'd be more worried about their reasoning for all this. They honestly think tanks, healers, and support don't have fun and would rather be DPSing.
Exactly.. People can keep calling it whatever they want.. CODE is CODE.. The fight will either be scripted by code, or it will be random.. If it is scripted by code then players will learn that code and use it against the mobs.. So we are right back to square one...... If the fights are random, then smart AI is just smoke and mirrors.. Toss this in with all the twitch non tab targeting.. I would be greatly surprised to see medium size audience after a year..
expect there to be as much clothes as armor in this game. Expect there to be as much crafting and building as combat. Expect perfect world meets minecraft meets everquest meets wow.
I actually quite like twitch-based combat; the most fun I had in a fantasy-based mmorpg over last years was playing a double-knuckle warrior in Allods-online - no targeting, just arc of damage. Man, did I have fun! So, twitch-based is quite allright with me, provided it has no lag and smallest ping possible (like Allods for me).
And utility-based AI may be quite ingenious and fun too... But! And it's a big great but. If it's custom-tweaked for every encounter, with all those weights and rules. And that's a LOT of coding time and testing for every mob, for every thinkable combination of groups. We are speaking weeks for one single boss here, at least, for a group of highly-paid programmers. And storybricks will not help here, like they may help for macro-AI in the world interaction.
Whole "threat generation mechanics" were invented specifically to create a system which would work, more or less, for most mobs for most parties for most encounters, thus reducing the time needed for tweaking each and every mob, or at least boss, for each and every situation.
I don't see here anything that will make the AI in any way better than GW2 AI whatsoever. In fact, I'm betting the AI programmers will go the same way here (because it's the only actually reasonable way to go, based on where they started and where they want to go), and the AI will be very close in its behavior.
Smart AI has been around for ages.. It's nothing new.. AI is nothing more then programming the actions of the mob.. Most games choose to go "simple" where as other games allow "complex" or advanced AI.. This is just like playing strategy games.. Civ 5? C&C?.. Players are given options to go simple or go "god" mode if they wish.. Players will ALWAYS find the best way to expose the weakness in AI codes and use it to defeat it.. Then what? Make the code so hard that 80% of your customer base says "SCREW THIS SHIT".. and moves on.. People talk about playing a GAME for fun, but if you code the AI too high, then it's no longer fun to the masses, it's becomes more like a job..
In conclusion.. The smart AI "HAS" to remain stupid for the masses.. So why bother? Think about it..
Or they could have simple to "god" mode mobs and cater to everyone? With a huge world to explore, it is reasonable to believe that some/most players will not see everything nor beat it all.
WoW's high end raids were only completed by a very small percentage of players (out of millions), yet they still had millions playing. MILLIONS played it because the general open world AI was was stupid.. The hard raids were instanced.. HUGE difference..
I don't see why? EQN could still have mainly "dumb" AI in Tier 1-2 and then have Tier 3-4 be much harder, with a sprinkling of various degrees of difficulty in each (mobs that could one hit you in the Commonlands).
That has since changed, but it was effective for many years. Add in numerous new things to do in the game beyond just mob grinding and raiding and people will stick around for a while.This remains to be seen.. None of us have any insider knowledge on what non-combat game play will be..
True, but they have given plenty of hints to what will be in the game. To make an educated guess.
Or just cater to only one part of the community and make everything the exact same difficulty, seems to have worked for many games... But you are asking them to "cater" to your desire of game play.. Double Standards? I have yet to hear anything from SOE about EQN cater to everyone.. EQN is going niche as well..
Guess I forgot the /s tag... Almost all games cater to everyone, difficulty wise. They have said that the while the AI will be "smart" there will be a scale of "intelligence." A random Orc may just swing at you mindlessly, but the Orc Chief may have a whole bag of tricks to choose from. Tiers will vary in difficult which sounds like an attempt to cater to a large audience. No clue how EQN will be niche if it is trying to cater to a wide variety of play styles.
If EQN wished to attract the masses it MUST be programmed for the lowest denominator.. PURE and SIMPLE and never open for discussion.. If the new smart AI offends too many customers, it will result in a very small niche game, or it will be tweaked down to dummy mode.. You mention "gode" mode mobs. UMmm.. EQ did that with dragons, and GW2 does that with open world bosses and both ended up being ZERGED.. It is what it is.. As the old sales slogan goes.. EQN is trying to sell the "sizzle" , and we have yet to see the steak..... Get the idea?
Again, all they have to do and what they've basically said they are doing, is have a scale of AI "smartness" that will allow for anyone to find something to do in a massive world.
As I've said to other Zerg comments, simple solution, don't reward people for doing so. Each mob should have a suggested player number attached to it and if it is exceeded, give no reward or one greatly reduced. On top of having non-static spawns to discourage a zerg of 50-100+ people running around the countryside killing everything in the way.
Risk vs Reward, reward those that actually play "fairly" even when it comes to mob slaying.
While some high end mobs were much harder then the regular trash, usually it was linked to massive amounts of HPs and hitting like a train. Not many have good AI that makes it a challenge. Just a set pattern (sometimes it is pretty tricky to figure out) to learn and then have the endurance to burn the mob down. EQN's AI may offer a lot more possibly.
A large chunk of the game does need to cater to the "lowest denominator" in some form, but that doesn't mean it has to be content that the "highest denominator" doesn't want to experience either.
I might love Chess, that doesn't mean I have to hate Checkers.
The concepts may not be "new" but can you point me to a mmorpg that has them already or to the scale at which EQN may?
As with all things tech, I'm sure they are talked about and imagined long before anyone gets their hands on the finished product. Why you wouldn't be excited to see in action something you knew about 20 years ago is beyond me.
Every mmorpg uses the same methods to program the AI. "Threat generation" behavior in mmorpgs is programmed _absolutely the same way_, it's just that the mechanics of threat generations are added as rules to weighted behavior of monsters.
And it was done this way "because reasons" - because without additional rules and tweaking emergent AI quite often emerges either stupid and easily taken advantage of (say, a guy in a dress and with a magic wand in his hands gets "haste" from group-members and runs around screaming, while monster trying to attack him and everyone else just pounds the monster without any retaliation) or becomes boringly genius - say, the monster and all his minions focus their attacks on the healer no matter what, and it's a 100% wipe for a group, no way to overcome.
I see in action that "utility AI" in a lot of games and it's absolutely nothing special - it's how you implement rules what's important... and the descriptions given here are also nothing special: that's how every game programmer tries to implement his AI... but it has lots of limitations.
So, I'm not excited because what I read here is "we are going to make the same AI as everyone else have done everywhere, but ours will be totes better because reasons!" So no, I'm not excited until I see it.
Again can you point out some examples? I'm not doubting you, I just don't feel that I've played a mmorpg in the last 18 years that has what EQN is hinting at. It doesn't have to be ezmode kite until dead or hard mode, mobs kill everyone in 5 seconds because they can read your mind. There can be a balance and with enough "intelligence" programmed in along with an understanding of player expectations, I don't see why this won't be the case.
Again can you point out some examples? I'm not doubting you, I just don't feel that I've played a mmorpg in the last 18 years that has what EQN is hinting at. It doesn't have to be ezmode kite until dead or hard mode, mobs kill everyone in 5 seconds because they can read your mind. There can be a balance and with enough "intelligence" programmed in along with an understanding of player expectations, I don't see why this won't be the case.
Yeah I would like some examples given too.
I cannot think of an MMO where an NPC ups their prices because you are friendly with his wife and he is Jealous. Or combat where, because you have killed a load of the mobs friends and are 'known' to them that they will target you. Or where mobs will move because the place they are doesn't have enough travellers on the road to attack.
All these things, unless you can provide examples, are new.
Again can you point out some examples? I'm not doubting you, I just don't feel that I've played a mmorpg in the last 18 years that has what EQN is hinting at. It doesn't have to be ezmode kite until dead or hard mode, mobs kill everyone in 5 seconds because they can read your mind. There can be a balance and with enough "intelligence" programmed in along with an understanding of player expectations, I don't see why this won't be the case.
That's my point. I simply don't believe what EQN is hinting at. I don't believe that EQN AI would be in any way whatsoever different than any other AI - because while they are "hinting" and promising miracles and marvels, nothing that they actually show is in any way different than what was already done a thousand times.
If you want an example of "utility AI" implemented in recent games, you can check out Radiant AI, developed for Oblivion. You can also check the resulting hilarity (I think a lot of articles about that was titled "Radiant AI my a$$!") - even while it worked, and worked reasonably well, some emergent behavior was ranged from hilarious to ridiculous.
In any case, this AI haven't turned fights into miracles of sentience, "hinted at" by EQN devs.
If you want to see utility AI in mmorpg monsters - GW2 has it. If you want to say "yeah, but EQN is hinting at so much more" - they can hint at anything they want, but what we'll get is, at best, GW2 level AI, no more - simply because there are limitations on what can be done in reasonable time with reasonable manpower resources with current "utility AI" - and there is no indication whatsoever that EQN has anything esle in stores.
I cannot think of an MMO where an NPC ups their prices because you are friendly with his wife and he is Jealous. Or combat where, because you have killed a load of the mobs friends and are 'known' to them that they will target you. Or where mobs will move because the place they are doesn't have enough travellers on the road to attack.
All these things, unless you can provide examples, are new.
I'm talking about fight AI. What you describe is the world/story AI, which is where StoryBricks (supposedly) comes to play; what you describe is routinely done for any single-player RPG out there and is NOT done for MMORPGS because implementing it requires big tables of relationships for each and every player in the world, loaded from database servers at every interaction with the player character.
Database connections and processing are one of the Great Bottlenecks of any MMORPG; implementation of such relationships is routine (as shows the example of single player rpgs); pushing that implementation through Database Connection Bottleneck is next to impossible.
Well, we have, of course, the example of EvE, where at least player's reputation with factions is taken into account and is, obviously, influenced by player's actions, but if EQN is planning on implementing something more complex than that (which I don't believe in), I expect quite serious lags in database connections.
But I may be wrong here, and it is quite possible that EQN has the solution for database lag, for the bottleneck here is not an AI implementation (which is trivial, as, again, is shown in any single player rpb since before Fallout), but database connections.
Wow... is it always like this around here? Some of you folks amuse me, some inspire me, some sadden me. And some of you... well... I don't know what to think.
Regardless, thanks for all the interest -- even the speculation and skepticism. Wish I could tell you more.
Soon.
You'll notice your post pretty much got skipped over under the weight of arguments. THAT'S what it's like around here. Heh.
I hope you all do open up, soon, and a lot. PR schemes suck. I mean, what, a year or more of this crap based on the lack of info? The negativity will prosper, and there will be none to wage the battle for your side because they don't have ammo. And promises don't do it, hard info on mechanics do.
Comments
No idea if kiting will be a viable option in EQN. Mobs could simply not chase players mindlessly (hoping AI doesn't follow a player in a circle) or they could CC players if they try to run.
As far as mobs going after the squishies, this is where team work comes in. Mobs would go after the DPS in the trinity system as well.
The tank + casters having avoidance abilities are usually what happens. If a Mage is casting away and a Mob decides they don't like it, the Warrior with the shield better stop it from charging.
This could lead to a ping pong effect, but the Mobs could CC the Warrior and then go after the caster who would have to use movement/defense to stay alive and hope their Cleric buddy can get some heals in before the Mob CCs them as well.
A Mob could cast Silence on a caster or cast a buff to deflect magic for a short period or any number of things to make the DPS not so effective while they deal with the guys stabbing them in the face.
Really, the amount of complexity is only limited by dev imagination and how much they program in and allow to happen.
Then again, a Mob may find the the guy knocking them down to be a bigger threat then the guy hurling fireballs and stay on the meat shield the whole time. We don't know how it will work out, but there at least is room for ideas beyond Tank and Spank.
Smart AI has been around for ages.. It's nothing new.. AI is nothing more then programming the actions of the mob.. Most games choose to go "simple" where as other games allow "complex" or advanced AI.. This is just like playing strategy games.. Civ 5? C&C?.. Players are given options to go simple or go "god" mode if they wish.. Players will ALWAYS find the best way to expose the weakness in AI codes and use it to defeat it.. Then what? Make the code so hard that 80% of your customer base says "SCREW THIS SHIT".. and moves on.. People talk about playing a GAME for fun, but if you code the AI too high, then it's no longer fun to the masses, it's becomes more like a job..
In conclusion.. The smart AI "HAS" to remain stupid for the masses.. So why bother? Think about it..
Or they could have simple to "god" mode mobs and cater to everyone? With a huge world to explore, it is reasonable to believe that some/most players will not see everything nor beat it all.
WoW's high end raids were only completed by a very small percentage of players (out of millions), yet they still had millions playing. That has since changed, but it was effective for many years. Add in numerous new things to do in the game beyond just mob grinding and raiding and people will stick around for a while.
Or just cater to only one part of the community and make everything the exact same difficulty, seems to have worked for many games...
How do they know which is the squishy target?
What if they really hate Elves and your frontline guy is an Elf?
What if the Orcs you are attacking recognise 'Slinky Orcbane' the tank and ignore the others simply to kill their sworn enemy?
What is the mobs you are fighting are really stupid and really only attack the closest enemy?
What if the mobs are intelligent enough that they realise soon after targeting the squishy character that there is a bigger threat from the bug with the big longbow and so choose to swap target?
The complexity of the fights with rely on the time taken to create a good set of options for each mob to recreate their 'thought processes'. Really, time is the only reason it won't improve things.
I am afraid it will develop into dress wearing dude running in circles.... similar to GW2 encounters.
If EQN wished to attract the masses it MUST be programmed for the lowest denominator.. PURE and SIMPLE and never open for discussion.. If the new smart AI offends too many customers, it will result in a very small niche game, or it will be tweaked down to dummy mode.. You mention "gode" mode mobs. UMmm.. EQ did that with dragons, and GW2 does that with open world bosses and both ended up being ZERGED.. It is what it is.. As the old sales slogan goes.. EQN is trying to sell the "sizzle" , and we have yet to see the steak..... Get the idea?
Every mmorpg uses the same methods to program the AI. "Threat generation" behavior in mmorpgs is programmed _absolutely the same way_, it's just that the mechanics of threat generations are added as rules to weighted behavior of monsters.
And it was done this way "because reasons" - because without additional rules and tweaking emergent AI quite often emerges either stupid and easily taken advantage of (say, a guy in a dress and with a magic wand in his hands gets "haste" from group-members and runs around screaming, while monster trying to attack him and everyone else just pounds the monster without any retaliation) or becomes boringly genius - say, the monster and all his minions focus their attacks on the healer no matter what, and it's a 100% wipe for a group, no way to overcome.
I see in action that "utility AI" in a lot of games and it's absolutely nothing special - it's how you implement rules what's important... and the descriptions given here are also nothing special: that's how every game programmer tries to implement his AI... but it has lots of limitations.
So, I'm not excited because what I read here is "we are going to make the same AI as everyone else have done everywhere, but ours will be totes better because reasons!" So no, I'm not excited until I see it.
I don't think there is any doubt they are going for easy mode game. Current game scripted mobs had a set of actions to do, and always did the same actions with the same intensity. The only difference between this and "intelligent AI" I see is that the new mob can make mistakes.
I can already see billions of ways to abuse that system. Like taking a weak character into party or naked and fast character, to make the difference in armor so big, that the monster AI will keep choosing him every few secons as a new way of taunting. It is that tag system. So easy to create your own environment, in which monster will keep choosing the same actions because of the differences in stats you throw at him. Unless they make monster retarded enough to not take such differences into consideration, then it is even easier.
It is just a bunch of promises. We are trying to discuss about it from technical side, and see how it is really possible to achieve. If everything developers said about their game in hype phase was real then every game released so far would look completely different... It is like taking seriously election promises.
Wow... is it always like this around here? Some of you folks amuse me, some inspire me, some sadden me. And some of you... well... I don't know what to think.
Regardless, thanks for all the interest -- even the speculation and skepticism. Wish I could tell you more.
Soon.
President & Lead Designer of Intrinsic Algorithm,
Author of Behavioral Mathematics for Game AI,
Game AI consultant, GDC AI Summit advisor, co-founder of AIGPG | IntrinsicAlgorithm.com
SOE, EA-Mythic, and others have made past efforts regarding how to measure your input for the quality of loot
example: Public Quest loot in EQ2 DOV
http://eq2wire.com/2011/06/29/public-quest-rewards-the-healer-conspiracy/
http://eq2wire.com/2011/04/11/kwills-quill-punishing-the-bad-really-punishes-the-good/
EQ2 fan sites
Yeah it is like this pretty much all the time.
Some people see possibilities, some lack what it takes to see them.
Thankfully the short-sighted people are not making games, they just clog up these forums from time to time.
If given the choice between believing a group of proven industry vets or a bunch of negative ill-informed forum dwellers then I am very excited for the future of this game and all future games.
Still waiting to see this in action over time. The fact that all of this has to be written in code and things will still have triggers lead me to believe you'll still have scripted fights on a basic level. It's still an algorithm, which is inherently "step-by-step" thus being predictable.
I'd be more worried about their reasoning for all this. They honestly think tanks, healers, and support don't have fun and would rather be DPSing.
Exactly.. People can keep calling it whatever they want.. CODE is CODE.. The fight will either be scripted by code, or it will be random.. If it is scripted by code then players will learn that code and use it against the mobs.. So we are right back to square one...... If the fights are random, then smart AI is just smoke and mirrors.. Toss this in with all the twitch non tab targeting.. I would be greatly surprised to see medium size audience after a year..
I believe its the opposite. DPS alone is not a role. Everyone gets to do that in addition to other things at the same time.
you pretty much have to do it that way these days because so few play the dedicated roles.
expect there to be as much clothes as armor in this game. Expect there to be as much crafting and building as combat. Expect perfect world meets minecraft meets everquest meets wow.
I actually quite like twitch-based combat; the most fun I had in a fantasy-based mmorpg over last years was playing a double-knuckle warrior in Allods-online - no targeting, just arc of damage. Man, did I have fun! So, twitch-based is quite allright with me, provided it has no lag and smallest ping possible (like Allods for me).
And utility-based AI may be quite ingenious and fun too... But! And it's a big great but. If it's custom-tweaked for every encounter, with all those weights and rules. And that's a LOT of coding time and testing for every mob, for every thinkable combination of groups. We are speaking weeks for one single boss here, at least, for a group of highly-paid programmers. And storybricks will not help here, like they may help for macro-AI in the world interaction.
Whole "threat generation mechanics" were invented specifically to create a system which would work, more or less, for most mobs for most parties for most encounters, thus reducing the time needed for tweaking each and every mob, or at least boss, for each and every situation.
I don't see here anything that will make the AI in any way better than GW2 AI whatsoever. In fact, I'm betting the AI programmers will go the same way here (because it's the only actually reasonable way to go, based on where they started and where they want to go), and the AI will be very close in its behavior.
Again, all they have to do and what they've basically said they are doing, is have a scale of AI "smartness" that will allow for anyone to find something to do in a massive world.
As I've said to other Zerg comments, simple solution, don't reward people for doing so. Each mob should have a suggested player number attached to it and if it is exceeded, give no reward or one greatly reduced. On top of having non-static spawns to discourage a zerg of 50-100+ people running around the countryside killing everything in the way.
Risk vs Reward, reward those that actually play "fairly" even when it comes to mob slaying.
While some high end mobs were much harder then the regular trash, usually it was linked to massive amounts of HPs and hitting like a train. Not many have good AI that makes it a challenge. Just a set pattern (sometimes it is pretty tricky to figure out) to learn and then have the endurance to burn the mob down. EQN's AI may offer a lot more possibly.
A large chunk of the game does need to cater to the "lowest denominator" in some form, but that doesn't mean it has to be content that the "highest denominator" doesn't want to experience either.
I might love Chess, that doesn't mean I have to hate Checkers.
Again can you point out some examples? I'm not doubting you, I just don't feel that I've played a mmorpg in the last 18 years that has what EQN is hinting at. It doesn't have to be ezmode kite until dead or hard mode, mobs kill everyone in 5 seconds because they can read your mind. There can be a balance and with enough "intelligence" programmed in along with an understanding of player expectations, I don't see why this won't be the case.
Yeah I would like some examples given too.
I cannot think of an MMO where an NPC ups their prices because you are friendly with his wife and he is Jealous. Or combat where, because you have killed a load of the mobs friends and are 'known' to them that they will target you. Or where mobs will move because the place they are doesn't have enough travellers on the road to attack.
All these things, unless you can provide examples, are new.
That's my point. I simply don't believe what EQN is hinting at. I don't believe that EQN AI would be in any way whatsoever different than any other AI - because while they are "hinting" and promising miracles and marvels, nothing that they actually show is in any way different than what was already done a thousand times.
If you want an example of "utility AI" implemented in recent games, you can check out Radiant AI, developed for Oblivion. You can also check the resulting hilarity (I think a lot of articles about that was titled "Radiant AI my a$$!") - even while it worked, and worked reasonably well, some emergent behavior was ranged from hilarious to ridiculous.
In any case, this AI haven't turned fights into miracles of sentience, "hinted at" by EQN devs.
If you want to see utility AI in mmorpg monsters - GW2 has it. If you want to say "yeah, but EQN is hinting at so much more" - they can hint at anything they want, but what we'll get is, at best, GW2 level AI, no more - simply because there are limitations on what can be done in reasonable time with reasonable manpower resources with current "utility AI" - and there is no indication whatsoever that EQN has anything esle in stores.
I'm talking about fight AI. What you describe is the world/story AI, which is where StoryBricks (supposedly) comes to play; what you describe is routinely done for any single-player RPG out there and is NOT done for MMORPGS because implementing it requires big tables of relationships for each and every player in the world, loaded from database servers at every interaction with the player character.
Database connections and processing are one of the Great Bottlenecks of any MMORPG; implementation of such relationships is routine (as shows the example of single player rpgs); pushing that implementation through Database Connection Bottleneck is next to impossible.
Well, we have, of course, the example of EvE, where at least player's reputation with factions is taken into account and is, obviously, influenced by player's actions, but if EQN is planning on implementing something more complex than that (which I don't believe in), I expect quite serious lags in database connections.
But I may be wrong here, and it is quite possible that EQN has the solution for database lag, for the bottleneck here is not an AI implementation (which is trivial, as, again, is shown in any single player rpb since before Fallout), but database connections.
You'll notice your post pretty much got skipped over under the weight of arguments. THAT'S what it's like around here. Heh.
I hope you all do open up, soon, and a lot. PR schemes suck. I mean, what, a year or more of this crap based on the lack of info? The negativity will prosper, and there will be none to wage the battle for your side because they don't have ammo. And promises don't do it, hard info on mechanics do.
Once upon a time....