Aggro

[Rogean]

Discussion regarding aggro changes.

[Rogean]

PEQ Copy

Originally Posted by atole

This sounds like this was intended to increase melee aggro? I've been having a harder time holding aggro over casters on adds since this change.

Anyone else having problems?

Originally Posted by ghostriderdoom

Yes it seems that tanking things is harder again. I tanked Maniatic behemoth todaY. I had him solo on my warrior with Bloodfrenzy and epic 1.5 both equipped. Stun buff also on. I had him solo like i always do, 20 seconds of tanking i called for dps, and once they started to enter room MB then decided to go outside the room and attack clerics for no reason. He should have been under contol without a doubt. there was not alot of healing going on.

Also in ldon, seems that when u pull 2 mobs and get group of dps on 1 of them, taunt both with warr, that 1 of them still goes after cleric. And again i have same gear as mentioned above. If the changes were suppose to make melee generate more aggro then it is not working correctly at all. Seems to be the exact opposite. the code seemed to work just fine before imo whatever thats worth. I use ldon as an example because thats what i have been doing recently. Everyone in guild says same thing about melee changes thats mobs are responding less now.

Originally Posted by Aliboo

If you're having problems with that kind of gear, I'm in big trouble with my very ill-equipped warrior. I thought aggro was worked out a while back. I wonder what prompted the change.

-----

[Rogean]

I think the problem is less due to the changes and more due to the fact that anything pulling aggro is probably generating more aggro than it should.

EQEmu has very sloppy aggro code, most formulas are assumed and incorrect, and its spread all over the code so its hard to find. We need to get the correct values for it instead of just hack fixing it.

The first change I made resulted in more overall melee aggro, due to every swing generating the same aggro. However the second change lowered melee aggro because the formula we were using before was probably double the aggro it is now.

When comparing aggro theres a lot of things to keep in mind other than simply melee swings. Like aggro procs, aggro focuses, etc.. and none of these may even be working correctly.

If your having problems with tanks holding aggro, please let me know what class is over-aggroing, and doing exactly what to cause it (What spells they are casting, etc).

[Skomag]

Originally Posted by Rogean

IIf your having problems with tanks holding aggro, please let me know what class is over-aggroing, and doing exactly what to cause it (What spells they are casting, etc).

Heals are drawing more aggro than a tank can counter. On MB in poi and LMM in hohb today clerics (cheal) were pulling aggro off the warrior tank (greenmachine)

[Rogean]

I'll take a look at the heal aggro and see if its using the correct formulas.

[testmur2]

I've noticed that Earthcaller proc (Slow) made my ranger more likely to draw aggro from my warrior, even with a couple taunts and weapon procs into the fight. Jolt procs have helped her with that aggro issue and the rogue has no problem at all but she's using her hide skill as soon as it's up.

Aggro from Complete Healings is way too high, almost every fight I'm having to pull the npcs off my cleric, I need to get her some AAs to reduce aggro and do more testing.

As far as I remember the PoP era, every class had its issues with aggro management and good tanks were known for their ability to keep the bosses on them for the whole fight. Besides the healing aggro, which seems way off, I don't think we are too far from what aggro generation was like on live.

My 2cp =)

[pfyon]

I'm not sure if any of the developers have seen this thread before, but it might give you some insight as to how aggro worked on live (at least, when the tests were performed). We could use this knowledge to help create a more accurate formula for aggro on eqemu.

Background: The wizards at graffe's decided to do some tests on aggro using concussion to determine how the spell worked. Concussion was used since a) it was of interest to wizards at the time and b) it provided a known aggro modifier and was completely under the testers' control.

[pfyon]

Stupid edit time limit.

Here's another thread which tests aggro decay and other theories (including taunt)

[Rogean]

Thats good for damage stuff, but still looking for solid facts for heal aggro.

[Rogean]

Alright, I found a bug in the code that was causing aggro reduction focuses to actually increase aggro instead of reducing it. Going to fix this and look for any other mistakes.

Also still looking for the real heal formula if anyone has a link to info on it.

[Dibalamin]

One thing to remember about CH, on live it generated approximately 400 hate, it was equal to it's mana cost, all other heals were a 1 point for 2 points healed if I remember correctly. Damage was 1 to 1 ratio.

[Rogean]

I wonder if all heal spells are just based on manacost for aggro?

[Shin Noir]

hmm purely relying on mana cost to determine healing aggro seems incorrect, because if that was the case what is the point of utilizing Heal Over Times? I utilized them because they would not generate aggro in a large burst, some HoTs costed a significant amount of mana, but did not aggro until it healed for a certain amount. The delayed healing spurts gave the tank adequite time to deal the damage or generate hate to keep the aggro off the cleric, where a complete heal or burst heal especially at pull would get it on the healer.

As is I think taunt is too good btw, but I don't think it's worth complaining about especially since I lack any formula to fix it. I think it primarily has to do with the mob lvl vs. your taunt level, I remember being a SK or War on live and trying to taunt a mob that was red to me and having aggro issues. If the mob was yellow or below I could taunt fine, you just had to be careful with high con mobs.

This is just the button btw, not procs or other forms of hate generating tools. I think that the stun line is also not generating enough aggro, as it typically pissed off a mob a LOT when you would cast it as a paladin or cleric or enchanter. Also mez would generate a lot of hate.

I've never once worried about breaking mez on my warrior because 1 press of taunt is guarenteed aggro off the mezzer on EQEMU... Which isn't right, on live you had especially earlier levels a lot of worry about the transition of breaking mez.. This is why later mezzes gained the memblur component to help assist in this tricky situation.

[Shin Noir]

Thinking about taunt button more, I think this was primarily an issue as SK/PAL though, warriors had a higher level skill cap of taunt and also earlier levels exclusively relied on taunt button to keep a mob at check. Perhaps look at the skill cap difference of an SK/PAL vs War, and make a % chance of failure based on War's skill level vs. his base level as a "80% gauge formula", and decrease it for each level above the war he tries to taunt, and same with SK/PAL they would have a handicap built into it.. They should be relying on their terror lines and stuns more often.

So a level 30 Warrior with taunt maxxed taunts a lvl 1-32 mob, he has a 80% chance of successful hatelist modification. A level 30 SK/PAL vs. 1-28 mob has the same 80% chance, but 29+ it is reduced a percentage each level. This would make you have to rely more on casting spells to maintain hate, which I personally think is more realistic to live.

This is without looking at the source for aggro, perhaps this is already implement in some form. ^^;

I think Divine Arbitration and the 1.5/2.0 shield clicky are also not generating enough hate, though I could be wrong at them generating any.

[Rogean]

==09/29/2009==
Rogean: Direct Heal and Rune Aggro are now based off the mana cost of the spell.
Rogean: HoT Aggro is now based off the per-tic heal amount, when the spell lands.
Rogean: Fixed several issues with focus effects. Beneficial Hate Reduction Focus should now work.
Rogean: Fixed AA Spell Casting Subtlety not applying to heal aggro. Also changed to correct values.

____________