Aggro from different pet skills

Solandri - Heavens Tear · April 2009

I decided to finally get some semi-definitive answers to this. I sent my pet to do one attack with each of the different skills, then stopped it. Then I hit the mob with my sword until I drew aggro off my pet. At that point I flew up out of range, and added up all the damage I'd done. Unfortunately this took a lot more time than I expected (I need to find more cooperative test subjects) so I only got 5-6 trials per skill. But the results were pretty consistent, with the sword damage being the biggest source of error (each hit was about 350-650 damage, yielding an average uncertainty of about +/- 250). I'll spare you the math on the uncertainties, I'll just note that the samples were small enough I used basic error propagation instead of standard deviations.

First the base case. I sent the pet in to do one attack with no skills. As expected, the moment my melee damage exceeded the pet's melee damage, the mob turned to attack me. I'll save spell damage for a later date (and encourage others to try it for themselves).

Crystalline Magmite - Bash 2
Damage per hit: 1503
Average damage to draw aggro: 5051 (+/- 5%)
Damage equivalent of Bash 2: 3548 (+/- 7%)
Damage equivalent is damage to draw aggro minus the damage dealt, leaving only aggro from the skill.

Shadou Cub - Bash 1
Damage per hit: 966
Average damage to draw aggro: 3917 (+/- 6%)
Damage equivalent of Bash 1: 2951 (+/- 8%)

Shadou Cub - Bash 2
Damage per hit: 1057
Average damage to draw aggro: 4505 (+/- 6%)
Damage equivalent of Bash 2: 3448 (+/- 7%)

Shadou Cub - Bash 3
Damage per hit: 1147
Average damage to draw aggro: 5284 (+/- 5%)
Damage equivalent of Bash 3: 4137 (+/- 6%)

Shadou Cub - Bash 4
Damge per hit: 1237
Average damage to draw aggro: 6051 (+/- 4%)
Damage equivalent of Bash 4: 4814 (+/- 5%)

Shadou Cub - Flesh Ream 1
Damage per hit: 726
Average damage to draw aggro: 3510 (+/- 7%)
Damage equivalent of Flesh Ream 1: 2784 (+/- 9%)

Frogling - Toxic Mist 1
Damage per hit: 579
Average damage to draw aggro: 3466 (+/- 7%)
Damage equivalent of Toxic Mist 1: 2887 (+/- 9%)

Crystalline Magmite - Roar 4
Damage per hit: 0
Average damage to draw aggro: 4972 (+/- 5%)
Damage equivalent of Roar 4: 4972 (+/- 5%)

Notes:

There were a few outliers which I threw out. I assumed they were due to me not stopping my attacks in time. But it's possible that there are other factors I didn't consider which caused those outliers (e.g. drawing aggro with much lower or higher damage).

Critical melee hits didn't seem to affect the results any. I had about a dozen scattered through my testing, including one test run which had three. Its final results were the same as the runs without any crits.

Bash 2 on the cub and magmite seem to add the same amount of aggro. That is, there higher pet attack doesn't seem to augment the taunt of the skill. The only difference is in the damage component.

The samples are too small to say for sure, but it kinda looks like Bash 1 adds aggro equivalent to 3000 damage, and each upgrade adds 600 more.

Flesh Ream 1 appears to add the same amount of aggro as Bash 1 (within my margin of error). Of course Flesh Ream has a 15 sec cooldown vs. Bash's 8 sec, so Bash clearly has an advantage.

Toxic Mist 1 appears to add the same amount of aggro as Bash 1. This will need more testing, but if true it means the elemental attacks may be as effective as Bash at holding aggro.

The Frogling was level 56, vs. level 66 for the magmite and cub. So assuming Bash 1 and Toxic Mist 1 are equivalent, it would appear pet level does not affect the taunt value of the skills.

"Helpful" people would randomly buff me or "save me" from the mob which was "killing" me or my pet. I guess this is a good thing that players are so altruistic. But nevertheless in this case it slowed things down. b:chuckle

Surfer_Rosa - Sanctuary · April 2009

Thank you for sharing, this is pretty interesting and could be very useful as well. I would have to ask how much do you think level affects your figures? Where one to expect the aggro effect to proportionately increase with a pet's level, this would eventually signify a reduction in a pet's ability to hold aggro as a higher level venomancer's damage output does increase at a higer ratio than mean pet damage. Is any given skill's ability to mantain hate related to this? Also, more definitive testing may be acomplished by a veno with a Phoenix, as Claw might allow you a way of testing how much of an effect higher damage will provide all other factors being equal. Keep up the good work.
Edit; I don't think i made it clear i'm also refering to aditional aggro generated by regular attacks as being a factor which should influence overall threat level. Also, how would you test hate generated by Roar? I don't think is so much a skill meant to mantain aggro but rather to grab it.

Flyingpix - Heavens Tear · April 2009

I don't know but something seems a little off with the results to me. As an arcane build the melee damage that I can draw with swords is well pathetic while my spells are my source of damage. I can routinely hit for 1400+ on same level mobs and never draw agro when my crab has hit first no matter how much I spam my skills. That's with my crab using Bash level 3 hitting for around 800 with an average dps of 550.

Solandri - Heavens Tear · April 2009

That's a good question. I'm making the implicit assumption that one point of pet melee damage always generates the same amount of aggro. It's possible that depending on the actual damage of each hit, the amount of aggro varies (e.g. a single hit for 2000 points has different aggro than 5 hits for 400).

Unfortunately I don't have any way to test that. I was already pushing the limits of my pets' tanking ability. I healed my pets in a couple of the earlier trials and the heals definitely changed the results (the healing adds aggro, though I'm not sure how yet). So I had to collect all the above data without healing the pet. If I try to decrease my melee damage, it will take more time and my pet will die before I finish. And I can't increase my melee damage significantly. Well, maybe fox form, but that's only about 2x. It might be possible to test it with two people - one person having the pet attack and keeping it healed, and the other person doing higher damage attacks until s/he pulls aggro.

There was one outlier which drew aggro off the pet at about 500 points lower damage than the others (Frogling test, ~3000 damage instead of ~3500). I read off and added up those damage numbers 4 times just to be sure because I couldn't understand how that was possible. If the damage was higher then I probably didn't stop my attacks in time and there's an extra hit's worth of damage in my total. But I can't think of any explanation why I would draw aggro with significantly less damage than the average. There was nothing unusual about my damage in that run - no criticals, no unusually low or high damage. In the end I ended up throwing that result out, but I still have no explanation for it. So it's very possible there are other factors at work here.

Also note that the above numbers are biased towards the high end. If I ran a thousand trials, the final "Damage equivalent" figure I got would be the actual aggro of the skill + half my average damage per hit. That's because to draw aggro I always need to exceed the threshold. And on average I'm going to exceed it by half my average damage.

Flyingpix - Heavens Tear wrote: »

I don't know but something seems a little off with the results to me. As an arcane build the melee damage that I can draw with swords is well pathetic while my spells are my source of damage. I can routinely hit for 1400+ on same level mobs and never draw agro when my crab has hit first no matter how much I spam my skills. That's with my crab using Bash level 3 hitting for around 800 with an average dps of 550.

I encourage others to try testing it with spell damage. It very well may behave differently than melee damage. I don't know. That's the whole point of this - to find out. And if your spells are hitting for 1400 with an average cooldown + channel of 2.5 seconds, then your DPS is only 560. So your pet with the added taunt from Bash should easily be able to keep ahead of you.

aazadan · April 2009

Interesting. I think you're making some assumptions though, particularly when you throw out data. If you have the numbers of the attacks it's more likely that you're missing a component of hate than something messing up.

I think it's obvious that damage itself isn't the sole provider of aggro, if that were the case the tank would always be the highest dps person in a squad (or as a veno, your pet would always be atleast 51% of your dps if you want it to tank) and that isn't the case.

Have you considered it working off of potential damage or swings? For example, each pet swing could be assigned a certain aggro value, bash could as well, and so could other skills. If that were the case then looking at damage won't get you anywhere except for the situations where you and the pet always hit for the same damage. If you still have all of your data, try looking at it and seeing if aggro changed at x attacks each time.

I've also seen that at various points a pets aggro will spontaneously improve which would lead more towards hate per swing increasing than anything else as a small damage boost wouldn't be large. An example here is leveling my petite sawfly against foxwing supremes. At level 42 it had to use bash, toxic mist, and have some lead time on melee in order to let me nuke, but at 43 I could nuke from the start provided I used both skills, and at 44 the aggro is exactly the same as at 43. The dps over that time has only gone up by about 6% so it seems pretty safe to say the dps boost isn't what increased the aggro.

Anyways, I would be interested in seeing the swing data if you have it (or any other raw data), I mostly reverse engineered the aggro code of other MMO's I've played in the past so I'm kind of interested in taking a shot at this. And one idea I had that may work out for you, is have your pet use tough prior to aggro that will get you a few more seconds or if that doesn't work, a second veno to heal the pet.

Edit: Just for a little more anecdotal evidence about damage/aggro not equaling each other, I've seen most stuff I pull aggro from my pet on while soloing being magic resistant creatures as I tend to nuke more then to compensate (plus dots). If aggro were directly tied to damage my aggro would be lowest on these fights as I have the least dps on them. Similarly my pet would do noticeably worse on fights with Increased Defense as the pet does less damage, and thus the least aggro yet they're no different than things without attributes.

OMGLAZERZ - Heavens Tear · April 2009

You might consider running your tests in an instance to avoid outside interference.

IceJazmin - Heavens Tear · April 2009

Did you get the chance to test the pet skill ROAR? I have it on some of my low damage pets and I am very pleased how well it works.

Roar does no damage but it is the best skill at holding/getting aggro on the pet.

senovit · April 2009

Wonderful. Thanks for doing this, people have always wondered about pet skill aggro. I'm somewhat surprised at flesh ream drawing more aggro than an elemental bash though.

Solandri - Heavens Tear · April 2009

Just to answer some questions:

Except for that one strange case of me drawing aggro with less than expected damage, all the outliers I threw out were cases where I knew I might have made an extra hit. Once I got a feel for how many hits it took to draw aggro, I slowed down before that point and would do one swing at a time and wait to see if I got aggro. Occasionally I didn't stop my attack fast enough and did two swings. If this happened and I drew aggro, I threw the result out since I didn't know if it was the first or second hit which gave me aggro.

The number of swings to draw aggro was in almost all cases within +/- 1. I can't really vary my damage per swing without a second person to help by healing the pet.

I don't have the raw swing data because soon after I started, I realized the mob I'd picked had a defense-boosting buff. I was careful to work around that (made sure the buff was up for my hits), but realized this mob wasn't a good choice for thorough testing. I just finished off the tests I posted to satisfy my curiosity.

I didn't do this in an instance because then I would have to waste time killing the mob after I finished each test. Outdoors, I simply fly up out of range, lose aggro, then drop down and do the next test on the same mob (I have very good hp regen gear b:chuckle).

Yes I tested Roar. Look through my first post, you'll see the results from Roar in there. It seems to be worse at holding aggro than Flesh Ream or Bash due to doing no damage. I'll test Pounce (stun) next time too.

aazadan · April 2009

What were you using to test it on? And did you consider using a weak creature (less damage dealt, less hp) for easier testing? When you say the number of swings was in almost all cases +/- 1, do you simply mean on bash, or do you mean flesh ream, toxic mist, and roar also all fell within that +/ 1?

I would be happy to help with this, but being on Dreamweaver and you on HT I suppose that's not really an option. There's also one other thing to consider with instances, there's no guarantee that elites/bosses follow the same aggro rules as regular creatures, and since the vast majority of instance creatures are elites, that introduces another potential variable.

One last question, where you list damage per hit, that's what your pet actually hit for with the attack right? Not the damage stat of the pet? Could you give those stats if so?

IceJazmin - Heavens Tear · April 2009

Solandri - Heavens Tear wrote: »

Just to answer some questions:
...
Yes I tested Roar. Look through my first post, you'll see the results from Roar in there. It seems to be worse at holding aggro than Flesh Ream or Bash due to doing no damage. I'll test Pounce (stun) next time too.

I think next we need to test which skill is better at getting aggro back to the pet after someone steal the aggro... Roar is the fastest skill at getting aggro back, you will always get the aggro back after you use Roar, while it may take a few hits with the other skills to get the aggro back.

Thanks for your testing, soooo, Roar to get aggro back, bash to keep aggro. It is good to know

OMGLAZERZ - Heavens Tear · April 2009

aazadan wrote: »

There's also one other thing to consider with instances, there's no guarantee that elites/bosses follow the same aggro rules as regular creatures, and since the vast majority of instance creatures are elites, that introduces another potential variable.

The amount of aggro built is determined by the skill/ability, not by the mob being attacked.

Solandri - Heavens Tear · April 2009

aazadan wrote: »

What were you using to test it on? And did you consider using a weak creature (less damage dealt, less hp) for easier testing? When you say the number of swings was in almost all cases +/- 1, do you simply mean on bash, or do you mean flesh ream, toxic mist, and roar also all fell within that +/ 1?

I tested on those pink mobs with extra arms above their heads in the Swamp because that's what I was fighting at the time. The problem with using a weaker mob is that it increases my damage, causing greater uncertainty. What we need is a mob which does little damage, but has high defense and hp.

The pet made just one attack. The +/-1 is the variation in the number of attacks I made before drawing aggro. As each skill and pet did a different amount of aggro, each required a different number of attacks for me to draw aggro. But for each skill and each pet, the variation on number of attacks was almost always +/- 1 (a few criticals and lucky/unlucky streaks changed it, but not by much).

One last question, where you list damage per hit, that's what your pet actually hit for with the attack right? Not the damage stat of the pet? Could you give those stats if so?

Yup, the damage my pet did to the mob in their one attack (so I can separate out damage aggro from skill aggro).

Golem (L66) attack score is 1894
Cub (L66) attack score is 1102
Frogling (L56)attack score is 942

aazadan · April 2009

Solandri - Heavens Tear wrote: »

I tested on those pink mobs with extra arms above their heads in the Swamp because that's what I was fighting at the time. The problem with using a weaker mob is that it increases my damage, causing greater uncertainty. What we need is a mob which does little damage, but has high defense and hp.

The pet made just one attack. The +/-1 is the variation in the number of attacks I made before drawing aggro. As each skill and pet did a different amount of aggro, each required a different number of attacks for me to draw aggro. But for each skill and each pet, the variation on number of attacks was almost always +/- 1 (a few criticals and lucky/unlucky streaks changed it, but not by much).

Yup, the damage my pet did to the mob in their one attack (so I can separate out damage aggro from skill aggro).

Golem (L66) attack score is 1894
Cub (L66) attack score is 1102
Frogling (L56)attack score is 942

What about an air mob then? I know there's a few level 60 foxwings around tusktown that are low enough for a land tank to grab it (particularly if you have a friend grab initial aggro and pull it to you for the pet).

Solandri - Heavens Tear · May 2009

Addendum to the original tests

Cub Pounce 1 (does no damage)
I needed 2889 avg (2776 min) melee damage to pull aggro.

Kowlin Pierce 1 (does no damage)
I needed 3483 avg (3351 min) melee damage to pull aggro.

Kowlin Howl 1 (does no damage)
I needed 2925 avg (2664 min) melee damage to pull aggro.

Basically, Pounce and Howl give as much non-damage aggro per application as Bash 1. At this point I'd say it's likely that all the pet attack/debuff skills generate the same amount of aggro. Bash (and the elemental attacks) just have an advantage in that their cooldown is only 8 sec.

The Pierce result is interesting in that I needed about 21%-26% more damage to pull aggro than the other skills. Pierce 1 increases melee damage by 24%. This suggests that number of hits matters for aggro, not just damage dealt. Or that the aggro calculation is done before the damage multiplier for Pierce (and presumably other skills like Amplify Damage and Ironwood Scarab) are applied. If the latter is true, this means casting Ironwood does not help your pet keep aggro. Likewise, it's possible that against an Increased Defense mob, your pet's ability to hold aggro is not reduced.

Aggro from different pet skills

Comments

Categories