Antiarc

It's sorta a pain to set up filtering for each fight, though. When does the 110/130% not play nice, though, out of curiosity?

The only way we have to determine aggro is through targeting. When a mob uses an RSTS, he targets someone besides the tank, but the tank still has aggro. There is no programmatic way to know that the tank still has aggro, since all we know is "The mob is not beating on the tank right now." For all intents and purposes, the RSTS target momentarily has aggro.

Also, what if I have 120% of the tank's threat? I may be top threat, but not have aggro. Dangerous? Sure. Correct to assume that I'm the basis for the aggro gain bar? Absolutely not - suddenly, the aggro bar shows at 156% of the tank's threat - 26% past the point that another ranged would actually pull aggro off of the tank, and a full 46% past the pull point for melee, which makes the bar not only worthless, but dangerously misleading.

Well, you can turn the aggro gain bar off, if you like. I think the option's in there somewhere.

If you're running KTM, you don't need TTK - TTK is just a light interface for sending KTM commands to KTM users. The actual Threat -> KTM stuff is done in Threat now. TTK is basically just used for setting KTM master targets and such without actually having to run KTM.

 

Hygeia

That could be the problem

 

Thezilch

Off the top of my head, any fight in SSC would act in this fashion, where a ranged (at no more than 105% threat) had a chance at gaining aggro from our MT. Ever since having this happen repeatedly on encounters like Hydross or Tidewalker (who really wants to re-pull these, when ranged can control themselves), we've since not played with fire above the MT. Because of that, I can't speak about it happening on other encounters, but it is directly tied to RSTS abilities (guess), which most all encounters now have.

Also, the 77% (76.9) is correct. Just shot the post out too fast at work. That would be the 130% gain mark on the ranged player, with respect to the MT (incapacitated). The MT can very well be subject to the 130% rule, when a ranged gains aggro under an incapacitate.

 

Antiarc

I'd love to see some formal testing on this, because I'm just nearly dead certain the "non-preferred aggro targets" (ie, not the current threat king, but selected due to incaps) don't set a new threat threshold. Our tank had a hell of a time stance-dancing Nightbane before they slowed down his fear, but she'd get aggro back as soon as the fear broke, pretty much every time, even though a shadow priest that was right in her backpocket threatwise would end up with aggro while she was feared.

And why would 77% be dangerous - why not 91%? The tank isn't going to try to regain aggro at range, after all.

 

◊ Praetorian

Well it's a difference between running back to the tank and the tank having to run there and start hitting it.

An incapacitate/fear/etc. is just an effect that makes the mob ignore that player, but I think the new player the mob chooses is a "real" aggro target from that point on. This should be very simple to test.

Get a paladin and a ranged DPS.

Paladin generates 1k aggro on a mob.
Other player generates 1.2k aggro on a mob, at range.

Paladin bubbles. Mob runs to other player, who does nothing. Wait for bubble to wear off or click it off. What happens?

 

Thezilch

Well, we've always observed this on fear targets. More recently, we had ranged targets above the line on Kael'thas's Sanguinar. Because the fear can come just before 30 seconds, said ranged were dead. Generally, we prefer our tank to be able to regain aggro through a battleshout/what-have-you establishing 130%+ gain.

 

Roana

I don't think that would work?

Example: Mage1 has 120% of the tank's threat and highest aggro; the moment Mage2 reaches 108.33% of Mage1's threat, she'll pull aggro, not at 130% of Mage1's threat.

 

Roana

Actually, that's what I frequently do. Say, the Big Bad Wolf. I miss a fear with Berserker Rage; since we keep ranged DPS at a fair distance from the wolf, I can just use Battle Shout/Commanding Shout the moment the fear breaks to call the wolf back to me -- assuming no DPSer has broken the 76.9% threat barrier.

It's pretty situational, though.

 

Antiarc

Huh, good to know.

I'm considering adding an Incap bar, for sure, provided that tests establish that incaps do indeed set a new threat threshold. I'm just wondering if it'd best be useful at 76 or 91%.

 

Whiteknight

I guess there's a design philosophy question here for your mod. In my opinion the mod should provide the most accurate information possible and make no attempt to 'force' skillful play.

If there's a fight where paying attention to the 130% aggro gain bar is a liability, it's better to have the bar and tell people to ignore it than it is to have the mod try do fancy stuff like key off 120% off the top threat list or remove the bar altogether.

Basically, my ask is - keep the 130% aggro bar working off the known theory and let me and my raiders choose when to pay attention to it and when to look for the MT threat number instead. Most raids are already trained to do this - on aggro switch fights folk are tracking below the second or third on the hate list because the players know the fight.

I'd rather the mod not attempt to play for me :P

 

Eyana

I discovered a bug today:

Omen seems to check on raid join who is runnning omen and who is not. if someone switches from omen to ktm, while he is in a raid, and relogs, omen still thinks he runs omen and doesn't catch ktm data for that player at all. this works for ktm -> omen as well.
to fix the problem all omen running players in the raid have to relog/reloadui.
you need to check more often what version(if any at all) a player is running. or at least give us a command to force a check.

 

Nich

While not quite that test, I've often found that after eating a fear while tanking, I need to make some/any proactive move on the mob to regain aggro - almost as if it's forgotten about me, and I need to remind it I'm there.

I can't say how that varies with different relative threat levels, tho'.

 

Rasputin

This was my experience on my paladin as well. Until you generate some more threat after bubbling, you do not exist to the mob. A simple Seal or Blessing should do the trick though, shouldn't it?

 

• Chicken

That should work yes, I know I usually simply reapply Seal of Righteousness once if I need to Divine Shield out of some debuff. There is at least some threat generated by casting the Seal itself though (50 or so).

 

Marroc

Can't we just calculate it based purely on threat? Assuming the numbers are accurate, which they are getting very VERY close to, wouldn't Agro gain always be 110%, or 130% more than the person with the most threat? Then you'd just have to ensure that your threat changes are accurate, so for fights with temporary agro drops or gains (mulgar charge, netherspite red beam, etc.) your boss mod just modifies the threat correctly.

That is to say completely ignore who the game is telling you has agro, and let the mod figure it out. If done right, it should come out with the same answer anyway, but it will remove the 'agro bar bouncing' as the bar will no longer be generated from the mobs target.

Of course for this to work you would need to assign people as tanks so the mod knows who's threat is okay to base the agro gain bar off of.


Another thing you can do is 'delay' the agro bar showing up. By this I mean put in a threshold to basically say "you're not the current agro holder unless the mob targets you for X seconds"

Honestly, I think option two would be the easiest to implement, and with a little tweaking could be very accurate, it'd just be a little delayed (which is usually fine for most fights).

edit: I can't spell >_<

 

• Chicken

I think you're overcomplicating the possible solution to this "problem", an easier and more elegant solution would be to add a "Minimum time before Aggro Gain bar changes after target switch" option, set it to 2 seconds and you'll generally only ever get the Aggro Gain bar to show the appropriate information outside of a few rare encounters in which the bar is most likely useless in any case.

Edit: This was already said above, just read this as "I think the Delay idea is the best and most easily implemented!".

 

Marroc

you missed the second half of my post :P

"Another thing you can do is 'delay' the agro bar showing up. By this I mean put in a threshold to basically say "you're not the current agro holder unless the mob targets you for X seconds""

 

• Chicken

I think I did yes. That'll teach me to read properly.

 

Elhana

Yes, every time someone in raid join/goes online...
But precious cpu cycles > you, such a waste. Why do you want ever to turn off this wonderful threat meter and go with KTM? D </sarcasm>

 

Antiarc

No, it won't.

Tank has aggro with 1,000 threat.
Mage is DPSing and has 1,200 threat (120% of Tank).
Mage does not have aggro, as Mage is 10% below 130% of Tank's threat.

Calculating aggro gain as 130% of the highest threat would show the aggro gain at 156% of Tank's threat, which is 26% over the point that Mage will pull aggro, leading Mage to overnuke and actually pull aggro without ever passing the Aggro Gain bar.

That's what should be happening now, but I need to ensure it's working. The logic is basically "If the mob targets someone with less threat than the previous target, wait 2.5 sec before assigning them as the new aggro holder; else, switch instantly."

Eyana, I'll add an option to manually check versions and resynch using/not using Omen status. Right now, it's only done when someone joins the party (or more accurately, when the party size increases), so the dirty way to force a using/not using resynch is to kick someone from the raid and re-invite them. That's hardly an obvious solution, though.

 

Marroc

Thats why you'd mark 'tanks' and only base agro off the highest 'tank' in the list, but again the second option is most likely much cleaner and more stable.

 

Antiarc

That'd a) require user intervention to mark tanks, and b) would still break when someone marked as a tank is sitting in the aggro holder's buffer zone.

Edit: And c) it wouldn't show an aggro gain bar when anyone that isn't a tank gets aggro - say, to let a tank know what he needs to hit to regain it.

 

Marroc

True. All the more reason to go with Option 2. Clearly 2 > 1 in this case :P

edit:
Sorry about the math pun... 8hr work shift + hopped up on caffeine to make it through the day >_<

 

Agren

There was a change in this behavior with one of the patches after Ahn'Qiraj was released. Previously, you didn't have to do anything to recover aggro after coming out of an incapacitate. I remember getting Enveloped on Ossirian with a poor OT, they had dragged him to the next pylon, I came out of Envelop, and Ossirian made a beeline for me because I had more than 130% of the OT's threat. After the change went in, you were only evaluated for whether you should have aggro when you did something that generated threat. Priests noticed the difference when using Fade as well.

 

Marling

Interesting happenings in Gruuls Lair tonight: Notice the warlock, Wuillie:



followed shortly by,



His TPS just seemed a little....off....during the fight. Couldnt SS properly, but it spiked up to 8k tps, leveled out at ~4k for a bit and then returned to a more normal 300tps. Didn't have the problem with anyone else - just an outdated omen/threat issue?

http://wow.tachyonsix.com.nyud.net:8...dge!v$1!f$.png