Aller au contenu

Photo

Aggro mechanics


  • Veuillez vous connecter pour répondre
121 réponses à ce sujet

#76
Deerber

Deerber
  • Members
  • 16 824 messages

dr_random wrote...

It is said that the drone looks through cloak, this would explain why I the Prime sucks so much in Shadow vs Geth solos.


Uhm, I always thought the drone seeing through cloak is just a rumour spread by those annoying "cloack doesn't turn me invisible" guys...

#77
Computron2000

Computron2000
  • Members
  • 4 983 messages

Deerber wrote...

dr_random wrote...

It is said that the drone looks through cloak, this would explain why I the Prime sucks so much in Shadow vs Geth solos.


Uhm, I always thought the drone seeing through cloak is just a rumour spread by those annoying "cloack doesn't turn me invisible" guys...


Its not a rumor. It really does see though cloak as experienced many times. This is why a prime with his drone floating off far away acts like everyone else and shoots your last known location, while with his drone close by (as in the pre-fixed drone times), your cloak is not useful. You can test it yourself by getting rid of everyone except 1 prime on platinum wave 1 then drag the drone away from the prime

#78
Deerber

Deerber
  • Members
  • 16 824 messages

Computron2000 wrote...

Deerber wrote...

dr_random wrote...

It is said that the drone looks through cloak, this would explain why I the Prime sucks so much in Shadow vs Geth solos.


Uhm, I always thought the drone seeing through cloak is just a rumour spread by those annoying "cloack doesn't turn me invisible" guys...


Its not a rumor. It really does see though cloak as experienced many times. This is why a prime with his drone floating off far away acts like everyone else and shoots your last known location, while with his drone close by (as in the pre-fixed drone times), your cloak is not useful. You can test it yourself by getting rid of everyone except 1 prime on platinum wave 1 then drag the drone away from the prime


Uhm... I'm sorry, but I'll need some more evidence than that to believe it. And no, I'm not gonna solo a platinum wave of Geth just to know that :sick:

But I'm gonna give it the benefit of doubt, if nothing else. Thanks for the info :)

#79
Catastrophy

Catastrophy
  • Members
  • 8 452 messages

Deerber wrote...

dr_random wrote...

It is said that the drone looks through cloak, this would explain why I the Prime sucks so much in Shadow vs Geth solos.


Uhm, I always thought the drone seeing through cloak is just a rumour spread by those annoying "cloack doesn't turn me invisible" guys...


That's why I wrote "it is said" and use conditional - though I notice Primes ARE being the bane of my Shadow, the sneakiest of all the characters I play.
I have Atlantes and rocket troopers miss me when I cloak, Hunters firing the wrong way, Centurions firing all about in a panic, but the Prime and Drone a steady sources of pain and punishment. Especially that &%§$$! drone, it even staggers out of HM.

#80
Deerber

Deerber
  • Members
  • 16 824 messages

dr_random wrote...

Deerber wrote...

dr_random wrote...

It is said that the drone looks through cloak, this would explain why I the Prime sucks so much in Shadow vs Geth solos.


Uhm, I always thought the drone seeing through cloak is just a rumour spread by those annoying "cloack doesn't turn me invisible" guys...


That's why I wrote "it is said" and use conditional - though I notice Primes ARE being the bane of my Shadow, the sneakiest of all the characters I play.
I have Atlantes and rocket troopers miss me when I cloak, Hunters firing the wrong way, Centurions firing all about in a panic, but the Prime and Drone a steady sources of pain and punishment. Especially that &%§$$! drone, it even staggers out of HM.


HM? Well anyway, I haven't been playing the shadow in a loooong time. Maybe I should take her for a ride... :lol:

#81
UnknownMercenary

UnknownMercenary
  • Members
  • 5 547 messages

Deerber wrote...
HM? Well anyway, I haven't been playing the shadow in a loooong time. Maybe I should take her for a ride... :lol:


Heavy melee? Also, you should. Decapitating enemies! :ph34r:

#82
Deerber

Deerber
  • Members
  • 16 824 messages

UnknownMercenary wrote...

Deerber wrote...
HM? Well anyway, I haven't been playing the shadow in a loooong time. Maybe I should take her for a ride... :lol:


Heavy melee? Also, you should. Decapitating enemies! :ph34r:


Oh yeah, right... I was pretty sure it was me being stupid :lol:

Yeah, I used to love her when I started playing gold, then I discovered so many other kits I almost forgot her... Gonna dish her out one of these days. She's gonna need a respec, though.

#83
Sidney

Sidney
  • Members
  • 5 032 messages
I've never had drone not see me in TC when I solo. When I'm in a group I can't recall getting hit by drones but solo they were unerring in their shooting...and not just shooting down the wire or the last place I was but just plain hitting a cloaked target they should have had zero way of "seeing".

#84
Deerber

Deerber
  • Members
  • 16 824 messages

Sidney wrote...

I've never had drone not see me in TC when I solo. When I'm in a group I can't recall getting hit by drones but solo they were unerring in their shooting...and not just shooting down the wire or the last place I was but just plain hitting a cloaked target they should have had zero way of "seeing".


The drone's attack is a tracking one, like the Atlas' rocket and Banshees' warp. Even if you cloak, it will still hit you. That doesn't mean that it can see through cloak.

Atlases can't see through cloak, at least on this I think we can agree. Yet, if they fire their missile you can be cloaked all you like - you've gotta take cover anyway, cause it's a tracking attack.

Geth Primes' shots also track, but to a much lesser extent.

#85
Hausner85

Hausner85
  • Members
  • 544 messages
I think that being under attack also makes them switch aggro. But I'm not sure of that. Have to test it...

Also great work:)

#86
ThatGuyThatPlaysThisGame

ThatGuyThatPlaysThisGame
  • Members
  • 1 660 messages
Lol the cake called "aggro mechanics" is a lie.

#87
ThatGuyThatPlaysThisGame

ThatGuyThatPlaysThisGame
  • Members
  • 1 660 messages

Deerber wrote...

Sidney wrote...

I've never had drone not see me in TC when I solo. When I'm in a group I can't recall getting hit by drones but solo they were unerring in their shooting...and not just shooting down the wire or the last place I was but just plain hitting a cloaked target they should have had zero way of "seeing".


The drone's attack is a tracking one, like the Atlas' rocket and Banshees' warp. Even if you cloak, it will still hit you. That doesn't mean that it can see through cloak.

Atlases can't see through cloak, at least on this I think we can agree. Yet, if they fire their missile you can be cloaked all you like - you've gotta take cover anyway, cause it's a tracking attack.

Geth Primes' shots also track, but to a much lesser extent.

Always do 90 degree turns around corners to hit me :whistle:

#88
dumael

dumael
  • Members
  • 474 messages
Yet another great thread Cyonan.

#89
Oggy666

Oggy666
  • Members
  • 383 messages

Computron2000 wrote...

I didn't know how injured a target was factored into the aggro consideration. Good to know

Edit: Checked. The berserk limit is based on the number of enemies still alive. Its 2/4/6/10 for bronze/silver/gold/platinum, so the last spawn when the budget is exhausted will charge when there are 6 left on the map in gold


Isn't it maybe 10 for bronze and so on?  I remember when I was grabs-farming on white behind the old desk,
this berserk mode made it difficult to grab all the things that were coming.

#90
Jeremiah12LGeek

Jeremiah12LGeek
  • Members
  • 23 728 messages

Jeremiah12LGeek wrote...

Cyonan wrote...


That's one of the things I still would like to test, but is insanely hard to test properly due to how many variables that can interfere with it.


The only idea I can come up with is using four different classes on mic on hazard ghost. Whenever there are rain waves, all four players can try and simultaneously expose themselves.

There'd be some repetition involved, but it's possible AI tendencies to shoot lower health targets could be detected.


Had another thought. Testing it would work best on bronze, as that would give the longest window to observe AI behavour before succumbing to injuries. Lower damage would make it easier to detect if enemies switch targets due to injury.

I wonder if testing would need to be repeated for different factions/units.

Are the mechanics consistent for all unit types, or does a Prime target differently from a Dragoon? I could see a Prime taking injury into account more than other units, as they (maybe it's just me) seem to really prioritize finishing off wounded targets.

#91
Metandra

Metandra
  • Members
  • 788 messages
Bumping for an awesome thread.

On topic: I'd think that dragoons would focus more on the "distance to target" aspect, since CQC is where they're most dangerous. A lot of times, they seem to wander until they see something and then latch onto it like a leech. I think it would be fun to help test these ideas and theories out. A more informed player-base can only be a good thing, right?

#92
Blind2Society

Blind2Society
  • Members
  • 7 576 messages
Aggro mechanics? They all look for Blind2Society first. If he is dead, they shoot him a couple more times for good measure then shoot at someone else.

#93
.458

.458
  • Members
  • 2 113 messages
Don't forget that all abilities and all devices used create noise. Those sounds you hear when you deactivate a device or activate one are also heard by the enemy. The very act of enabling TC makes a noise. Some units have better hearing than others, and when they do hear or see something, they tend to tell units near them about what they heard or saw. Standing directly behind an enemy that so far does not know you are there will fail by activating TC. You can have duration spec TC, but units near an objective device will still know you are there because of those noises.

The game could use "silent walking" gear and weapon supressors.

#94
Jeremiah12LGeek

Jeremiah12LGeek
  • Members
  • 23 728 messages
 So I played a few games and kept my eyes open, looking for any patterns I could see regarding aggro/targetting.

There were limits on what I could test, because I was playing PuGs (unless I had been willing to sacrifice uknowing teammates for science :whistle:.)

   There's a phenomenon I've observed a few times (I'm sure we all have) where enemies ignore me and focus on a teammate despite obvious distance/LoS advantages to targetting me.

   After a few games, the phenomenon seemed to be observable about twice a game, on average (maybe a little less). I'll add at this point that I played all games on Silver as host, and using non-fitness specced Volus Adept and Asari Vanguard. So I was never using Tactical Cloak, and both characters have relatively low total health, if not specced for fitness. There was one game where I used a Vorcha Soldier, where I observed this happen several times, but his high health may have had something to do with that. It would be too hard to say.

   It started to seem as though, for most units, distance and LoS don't get much more priority than hearing does. And this would also lend a lot to the idea that the highest priority is given to the current target. Once they target you, it may take some kind of active change to the factors determining the target for a switch.

   At one point (as AV on FB White) I was able to follow a lone cannibal for about 15 seconds while this was happening. I would periodically use a hip shot to its back with my Carnifex, spacing my shots out because I wanted to keep it going for as long as possible.

   The cannibal made a turn, climbed a ladder, made another turn, and crossed most of the distance from FB White's LZ to the high ground perch before I killed it with a light melee. It was shot 4 or 5 times, but did not react to any of those shots.

   I started out at about 10 feet away, but gradually closed distance as I followed. By the end, I was following it about four feet away, and I shot it at least once from that distance. The cannibal never switched targets, or even turned to face me.

   I'm estimating it would have gone through 5 or 6 targetting cycles during my brief "experiment".

   Two of my teammates in that game (and several others) were using destroyer and batarian, and I didn't know how they specced them, but I'm guessing they had the highest health on the team.

   The fourth player had a Geth Engineer, and I think that may be relevent. I didn't see the GE go down more often than the other players, but we were dying at a higher than average rate (partly because I was dividing my efforts between killing enemies and just watching what they would do). Regardless of how the GE was specced, he would have had the lowest health on the team.

    I was never able to work out which of the three classes (other than myself) was being targetted the most.  I was able to follow enemies several times and I was either unnoticed, or ignored. I don't remember being appreciably wounded on any of those occasions, but I didn't do a good job of keeping track of that particular variable.

   I definitely think that once a target is chosen, regardless of the reason, iit takes a significant weight of factors to change it.

   I'll see if I can get a few guys from BSN to try a hazard ghost experiment later tonight.

Modifié par Jeremiah12LGeek, 14 janvier 2013 - 05:40 .


#95
Cyonan

Cyonan
  • Members
  • 19 352 messages

Jeremiah12LGeek wrote...

 So I played a few games and kept my eyes open, looking for any patterns I could see regarding aggro/targetting.

There were limits on what I could test, because I was playing PuGs (unless I had been willing to sacrifice uknowing teammates for science :whistle:.)

   There's a phenomenon I've observed a few times (I'm sure we all have) where enemies ignore me and focus on a teammate despite obvious distance/LoS advantages to targetting me.

   After a few games, the phenomenon seemed to be observable about twice a game, on average (maybe a little less). I'll add at this point that I played all games on Silver as host, and using non-fitness specced Volus Adept and Asari Vanguard. So I was never using Tactical Cloak, and both characters have relatively low total health, if not specced for fitness. There was one game where I used a Vorcha Soldier, where I observed this happen several times, but his high health may have had something to do with that. It would be too hard to say.

   It started to seem as though, for most units, distance and LoS don't get much more priority than hearing does. And this would also lend a lot to the idea that the highest priority is given to the current target. Once they target you, it may take some kind of active change to the factors determining the target for a switch.

   At one point (as AV on FB White) I was able to follow a lone cannibal for about 15 seconds while this was happening. I would periodically use a hip shot to its back with my Carnifex, spacing my shots out because I wanted to keep it going for as long as possible.

   The cannibal made a turn, climbed a ladder, made another turn, and crossed most of the distance from FB White's LZ to the high ground perch before I killed it with a light melee. It was shot 4 or 5 times, but did not react to any of those shots.

   I started out at about 10 feet away, but gradually closed distance as I followed. By the end, I was following it about four feet away, and I shot it at least once from that distance. The cannibal never switched targets, or even turned to face me.

   I'm estimating it would have gone through 5 or 6 targetting cycles during my brief "experiment".

   Two of my teammates in that game (and several others) were using destroyer and batarian, and I didn't know how they specced them, but I'm guessing they had the highest health on the team.

   The fourth player had a Geth Engineer, and I think that may be relevent. I didn't see the GE go down more often than the other players, but we were dying at a higher than average rate (partly because I was dividing my efforts between killing enemies and just watching what they would do). Regardless of how the GE was specced, he would have had the lowest health on the team.

    I was never able to work out which of the three classes (other than myself) was being targetted the most.  I was able to follow enemies several times and I was either unnoticed, or ignored. I don't remember being appreciably wounded on any of those occasions, but I didn't do a good job of keeping track of that particular variable.

   I definitely think that once a target is chosen, regardless of the reason, iit takes a significant weight of factors to change it.

   I'll see if I can get a few guys from BSN to try a hazard ghost experiment later tonight.


I've noticed that they really like sticking to the person they're currently targeting, even if another target becomes easier. The list is certainly weighted in favour of attacking their current target again.

Every once in a while the AI will also do something odd and target a player that wasn't the logical choice they should have made.

That's just sort of standard behaviour for AI that isn't scripted to do very specific things in the majority of games, though. I see it happen all the time, and not just in Mass Effect.

#96
ValorOfArms777

ValorOfArms777
  • Members
  • 3 089 messages
it's not really an AI Cyonan it's a VI as it does not even "learn"

#97
SilentStep79

SilentStep79
  • Members
  • 3 701 messages
i just hope no salarians were harmed during these experiments...

#98
Jeremiah12LGeek

Jeremiah12LGeek
  • Members
  • 23 728 messages

Cyonan wrote...

I've noticed that they really like sticking to the person they're currently targeting, even if another target becomes easier. The list is certainly weighted in favour of attacking their current target again.

Every once in a while the AI will also do something odd and target a player that wasn't the logical choice they should have made.

That's just sort of standard behaviour for AI that isn't scripted to do very specific things in the majority of games, though. I see it happen all the time, and not just in Mass Effect.


   I seemed to notice the phenomenon more with basic mooks, as well. I want to try having a look again, because I think Marauders may have a sligthly better set of priority definitions. I don't think a Marauder would have ignored me for as long as the Cannibal did.

#99
Learn To Love Yourself

Learn To Love Yourself
  • Members
  • 7 373 messages

Kalas321 wrote...

LegionofRannoch wrote...

again,

Cyonan is my hero. Posted Image



#100
Xaijin

Xaijin
  • Members
  • 5 348 messages

Ziegrif wrote...

Don't enemies also ''hear'' targets footsteps?
How does that factor into all this if at all?


The sound operates in a certain radius, if it's triggered within that radius and the enemy hears you and is only moving or in cover, they will re-evaluate their targeting. If the enemy is already engaged it depends on what the active target is doing and if the enemy is actively firing/attacking, which is how you can somehow surprise people as a Krogan Vanguard.

The AI also tends to repeatedly attack person with highest framerates, esp if they are hosts, as the update for targeting come out much quicker and they don't have to wait on delta timing resolves to get to the next active state, which can make maps like London and Condor a giant pain in the ass if you win the glowing reticle lottery.

Modifié par Xaijin, 14 janvier 2013 - 08:08 .