Forums » Suggestions
Bots are boringly easy once you know how to fly and shoot. I think it would be cool to have a new variant of each type: collector, assault, guardian. The new variant could use the same models, but have a different name and increased armor, firepower, speed, and maybe better aim. I know the bots where buffed years ago so they don't have superhuman aiming ability. These new variants could maybe show up in grey, so new pilots would not have to worry about them. I love to bot, but the bots are way too easy now. Anyway comments and ideas are appreciated.
If you find the bots too easy, perhaps you should step up to the next level: pvp
have you tried fighting the grey space bots?
I too have considered suggesting some tougher bots. While it may be a major part of the game, PVP is not the "be all and end all" of VO. I would suggest a harder variant of the skirmish missions, available only at high Combat license (10+ maybe).
Might be time to utilize those sick new models we got a preview of ages ago?
Might be time to utilize those sick new models we got a preview of ages ago?
Problem is, the AI in VO is not the best. It really does boil down to "am i attacking? yes/no - if im attacking get in range and point and shoot at it AND if incoming fire happens slow down and speed up to give the appearance of a dodge"
And that about sums up VO's AI's. No matter how much firepower or armour you give them they will always fall for the same old tricks, the best thing that happened to VO's AI was actually a result of bad AI programming changes and that was in 2005!
What the game needs is a better AI where the bots actually adapt to different setups and player tactics, but this is a LOT harder than it sounds, as Momerath found out. So yes, we do need better AI opponents, but just slapping on more of the same in tougher ships will not cut it. PvP really should not be the only challenge that VO has when it comes to combat.
[edit]
Oh, and I might add, this was back when AI's attacked in groups, not this pussy "1 at a time" behaviour they added for mobile newbs. I really wish they would revert to the old "gank" behaviour, at least for the tougher AI's like Prosus and Valent types.
And that about sums up VO's AI's. No matter how much firepower or armour you give them they will always fall for the same old tricks, the best thing that happened to VO's AI was actually a result of bad AI programming changes and that was in 2005!
What the game needs is a better AI where the bots actually adapt to different setups and player tactics, but this is a LOT harder than it sounds, as Momerath found out. So yes, we do need better AI opponents, but just slapping on more of the same in tougher ships will not cut it. PvP really should not be the only challenge that VO has when it comes to combat.
[edit]
Oh, and I might add, this was back when AI's attacked in groups, not this pussy "1 at a time" behaviour they added for mobile newbs. I really wish they would revert to the old "gank" behaviour, at least for the tougher AI's like Prosus and Valent types.
Yoda, it is my understanding they were nerfed many years ago due to them being super-deadly, never missing-all-dodging buggers. While I don't think we need THAT level of skill from them, a variant of tougher ones would be nice for those of us that can kill them willy-nilly.
Your understanding is wrong faille. Bots could never hit each other in the fashion you describe, but players v's them has always been weak. In 2002, I was on the top kill list for killing bots alone. The AI in VO has never been very good. Heck, in 2005 (if you read the thread above) we all got excited by a mistake in their programming.
Inc, please step in here about stuff you know nothing about!
a1k0n for president of VO!
Inc, please step in here about stuff you know nothing about!
a1k0n for president of VO!
A roving collection of assasin bots that takes revenge when the levi is killed ;)
See Luxen, that would not be a bad thing, but it also would with the current VO situation on VO's AI.
Any proper programmers wanna step forwards? Momerath tried, as did my god and deity a1k0n. Can you?
[edit]
Before you all say anything, watch what my god and saviour is doing now https://www.facebook.com/jalopnik/videos/10155562453520168/?hc_ref=ARSRJaQ1yE80y1_Wm-VBO3YBjTNGCPh9nBazKVdBrv2ChF6neXDj-TiyCJOg4a8J_g4&pnref=story
Any proper programmers wanna step forwards? Momerath tried, as did my god and deity a1k0n. Can you?
[edit]
Before you all say anything, watch what my god and saviour is doing now https://www.facebook.com/jalopnik/videos/10155562453520168/?hc_ref=ARSRJaQ1yE80y1_Wm-VBO3YBjTNGCPh9nBazKVdBrv2ChF6neXDj-TiyCJOg4a8J_g4&pnref=story
Have you played many mission arcs? Some ought to be fairly challenging, even for a seasoned pilot. Which nation do you belong to? I could recommend a mission to you that might test your skill.
+1. Greyspace bots need to be mean ganking bastards just like the rest of us.
Inc, please step in here about stuff you know nothing about!
Jesus, Yoda. I'm not the one arrogantly proclaiming lots of wrong information and misleading people on this thread.
Making bots that are harder to fight is pretty easy. Making them adaptive is a bit more complicated, but not really super-hard either. Ray and I have talked about it in the last month, related to some of my goals for the game.
From a historical standpoint, there was a problem in the very early days with bots just being way too challenging. This was before Yoda's time. Also, speaking of history, Momerath was involved with large-scale AI, like Kourier, which controls things like.. how many Hive Skirmishes are going on, and where they're located. I never seriously allocated him to trying to rewrite the low-level "combat" AI. Nor is there much need for that?
The low-level combat AI actually has a number of different options for positions and tactics, aggression and other factors. Mostly they haven't been used or changed in a long time. There's also a "gangup" factor that impacts the number of bots who may attack you at once, which is also seriously nerfed (it was not nerfed because of mobile, again Yoda is misinformed).
It is not inordinately complicated for us to roll a few new bot behaviours, and then have the bot change between them periodically based on success or failure, or random timing intervals, or other factors. This is not some crazy "deep learning neural net" stuff, this is a simple rule-based system, which drives almost 100% of all gameplay AI in all games everywhere.
Anyway, it's nice to hear that more people would like to see harder bots. I've been planning on that for awhile, probably to put more challenging variants in grayspace, and in the "corners" of systems, as well as more sporadic spawns related to pirates and random "Aces" and other situations.. which also ties into why we've been developing those "hulk" assets.
Jesus, Yoda. I'm not the one arrogantly proclaiming lots of wrong information and misleading people on this thread.
Making bots that are harder to fight is pretty easy. Making them adaptive is a bit more complicated, but not really super-hard either. Ray and I have talked about it in the last month, related to some of my goals for the game.
From a historical standpoint, there was a problem in the very early days with bots just being way too challenging. This was before Yoda's time. Also, speaking of history, Momerath was involved with large-scale AI, like Kourier, which controls things like.. how many Hive Skirmishes are going on, and where they're located. I never seriously allocated him to trying to rewrite the low-level "combat" AI. Nor is there much need for that?
The low-level combat AI actually has a number of different options for positions and tactics, aggression and other factors. Mostly they haven't been used or changed in a long time. There's also a "gangup" factor that impacts the number of bots who may attack you at once, which is also seriously nerfed (it was not nerfed because of mobile, again Yoda is misinformed).
It is not inordinately complicated for us to roll a few new bot behaviours, and then have the bot change between them periodically based on success or failure, or random timing intervals, or other factors. This is not some crazy "deep learning neural net" stuff, this is a simple rule-based system, which drives almost 100% of all gameplay AI in all games everywhere.
Anyway, it's nice to hear that more people would like to see harder bots. I've been planning on that for awhile, probably to put more challenging variants in grayspace, and in the "corners" of systems, as well as more sporadic spawns related to pirates and random "Aces" and other situations.. which also ties into why we've been developing those "hulk" assets.
Do limit the more advanced/harder changes to the higher (above 5) level players so that the bots don't beat up the newbs too badly. Perhaps add one higher level behavior per level above 5. Ganking should also be stepped up in a similar way.
Feel free to demonstrate how easy it is to create more combat worthy bots and place them in a convenient location such as Pelatus. Its a request that has been made repeatedly for years.
Feel free to demonstrate how easy it is to create more combat worthy bots and place them in a convenient location such as Pelatus. Its a request that has been made repeatedly for years.
If we add dangerous bots, they're likely to just.. be dangerous. Not exclusively dangerous to people at certain levels, that might be a little odd. But they will be located in more outlying regions.
I know the request has been made for years. But that goes for.. everything else, too?
Anyway, there's a new version of the "debris" sector in Latos that may be rolled out soon, I'm kickng around dropping some experimental bots in there if I get a chance. But, whether I have that chance in the immediate future is uncertain.. I need to write a newsletter and a dozen other things.
I know the request has been made for years. But that goes for.. everything else, too?
Anyway, there's a new version of the "debris" sector in Latos that may be rolled out soon, I'm kickng around dropping some experimental bots in there if I get a chance. But, whether I have that chance in the immediate future is uncertain.. I need to write a newsletter and a dozen other things.
Yoda you bring up another point: Make some sectors swarm you. I want a bot furball.
As for pvp, I pvp whenever I possibly can, but there are times i am on and there isw NOBODY on. So I bot.
As for pvp, I pvp whenever I possibly can, but there are times i am on and there isw NOBODY on. So I bot.
another possible idea: elite variants that have a 10% chance to spawn. merely take ordinary bots and jump their AI up. the higher teired bots become slaughterhouses, while the lowly collectors merely might hit once or twice more than normal on the wee little newbies. rewards with a increased lisc or drop percentage, maybe.
The biggest problem IMO is that bots have no aggro table. No matter how "dangerous" a bot is, the formula is always the same and it does not create any incentive for group play.
Example: Newbie wants to do some mining in a sector, he jumps in and one bot queues up on him. Another pilot jumps in and starts killing the bot. The bot continues to target player 1 no matter how much damage player 2 does. When player 1's bot is dead, a new bot queues up and the process begins anew.
Bots do have a proximity sensor, but it's essentially "Is the player in range? If so, send one to attack." The most meaningful change would be to give bots a smaller proximity and any bots in that proximity would start attacking the player within the proximity that has done the most damage to them. Players could still fight out away from the bots one at a time pulling a new one when the fight is done, or they could fight in the asteroids with 4-5 bots trying to attack them. This would also allow players to actually fill a role as tanks or protectors by drawing aggro off of players that want to mine without fighting.
I'm sure this would fundamentally change bot logic, and would probably break some things so I'll step back now and let Yoda comment on the feasibility and incarnate to tell him how wrong he is about his assumptions.
Example: Newbie wants to do some mining in a sector, he jumps in and one bot queues up on him. Another pilot jumps in and starts killing the bot. The bot continues to target player 1 no matter how much damage player 2 does. When player 1's bot is dead, a new bot queues up and the process begins anew.
Bots do have a proximity sensor, but it's essentially "Is the player in range? If so, send one to attack." The most meaningful change would be to give bots a smaller proximity and any bots in that proximity would start attacking the player within the proximity that has done the most damage to them. Players could still fight out away from the bots one at a time pulling a new one when the fight is done, or they could fight in the asteroids with 4-5 bots trying to attack them. This would also allow players to actually fill a role as tanks or protectors by drawing aggro off of players that want to mine without fighting.
I'm sure this would fundamentally change bot logic, and would probably break some things so I'll step back now and let Yoda comment on the feasibility and incarnate to tell him how wrong he is about his assumptions.
so I'll step back now and let Yoda comment on the feasibility and incarnate to tell him how wrong he is about his assumptions.
Heheh.
Savet, all that stuff is built-in, but we have to change some parameters, although there are some annoying and silly limitations that need to be reworked. The current gangup parameters are not great, and basically bots should always aggro on the immediate threat (not even "most damage", by the time that's high enough.. they'll explode before aggro'ing. So like if they're attacking someone who damaged them, but hasn't done any damage in 10 seconds, and a new hit comes in, they'll prefer that target, etc).
Aside from that, I have a lot of notes for more cohesive bot-killing goals.. like if you kill more than X in a sector, it increases aggression in remaining bots to Y, and perhaps spawns some higher-level bots. And if you clear an entire sector of bots, then Z may happen and you get some counter towards a badge. And maybe high-level bots that require significant sector-clearance (to "make appear") have some interesting drops. All of that would mesh pretty well with cooperative PvE.
Heheh.
Savet, all that stuff is built-in, but we have to change some parameters, although there are some annoying and silly limitations that need to be reworked. The current gangup parameters are not great, and basically bots should always aggro on the immediate threat (not even "most damage", by the time that's high enough.. they'll explode before aggro'ing. So like if they're attacking someone who damaged them, but hasn't done any damage in 10 seconds, and a new hit comes in, they'll prefer that target, etc).
Aside from that, I have a lot of notes for more cohesive bot-killing goals.. like if you kill more than X in a sector, it increases aggression in remaining bots to Y, and perhaps spawns some higher-level bots. And if you clear an entire sector of bots, then Z may happen and you get some counter towards a badge. And maybe high-level bots that require significant sector-clearance (to "make appear") have some interesting drops. All of that would mesh pretty well with cooperative PvE.
+1!
I would hope killing a lot of bots in an area would also bring in much more dangerous and aggressive bots to help defend.
It also occurs to me that the hive would in fact hunt down and destroy any easy target seen in thier grey space area.
(translate easy to read as" send enuff bots to get the job done ie. a hit squad" )
It also occurs to me that the hive would in fact hunt down and destroy any easy target seen in thier grey space area.
(translate easy to read as" send enuff bots to get the job done ie. a hit squad" )