Norgulzar

I've noticed that Omen disregards all threat coefficients tied to Righteous Fury and Holy Shield; I've checked the damage against the threat generated on all abilities and it's always 1:1. Granted, a threat meter doesn't impact my threat, but it makes the dps rather nervous when they creep past what they see as my threat, which is closer to the halfway mark.

 

Antiarc

Yeah, something in my RF checking is screwy; I'm checking it now.

 

Roana

I'd guess that he used Curse of Doom -- threat spikes seem to have a tendency to make Omen's TPS calculations stick at high values for a bit.

Overall, TPS calculations seem to be off quite a bit; when I just auto-attack something in defensive stance, TPS is consistently lower than DPS * 1.495, and TPS then is also consistently lower than total threat divided by the duration of the fight. And it takes at least a good 15 seconds (judged by the Charge timer) for TPS from auto-attack to approach anything resembling even close to the actual value (Omen r43332, Threat-1.0 r43327).

 

Ngita

Perhaps this isnt the place for this, since it it is a vanish bug rather then omen, but this is the only place i have seen it mentioned. We have had issues with this the last 2 weeks.

Here are the relevant lines, from my log, while I do have both deadly poison and rupture ticks from the rogue in my log I have none while he was vanished.

7/11 23:12:10.635 Oscarvil performs Vanish on Oscarvil.
7/11 23:12:10.662 Oscarvil gains 15 Energy from Oscarvil's Combat Potency.
7/11 23:12:11.086 Oscarvil gains Stealth.
7/11 23:12:11.086 Oscarvil gains Vanish.
7/11 23:12:11.101 Oscarvil gains 15 Energy from Oscarvil's Combat Potency.
7/11 23:12:13.012 Oscarvil's Garrote - Silence fails. Magtheridon is immune.
7/11 23:12:13.434 Oscarvil gains 1 extra attack through Windfury Attack.
7/11 23:12:13.434 Oscarvil hits Magtheridon for 531.
7/11 23:12:13.477 Stealth fades from Oscarvil.
7/11 23:12:13.478 Vanish fades from Oscarvil.
7/11 23:12:13.478 Oscarvil gains Windfury Attack.

Due to the warning we had posted on our forums about it being caused by rupture and Deadly poison

7/11 23:17:16.927 Oscarvil's Envenom hits Magtheridon for 1483 Nature damage. He did a envenom to remove poison.
7/11 23:17:17.795 Oscarvil performs Vanish on Oscarvil.
7/11 23:17:18.904 Oscarvil gains 15 Energy from Oscarvil's Combat Potency.
7/11 23:17:18.360 Oscarvil gains Stealth.
7/11 23:17:18.360 Oscarvil gains Vanish.
7/11 23:17:19.463 Oscarvil's Garrote - Silence fails. Magtheridon is immune.
7/11 23:17:19.705 Oscarvil hits Magtheridon for 508.
7/11 23:17:20.951 Stealth fades from Oscarvil.
7/11 23:17:20.952 Vanish fades from Oscarvil.

80 seconds after his second vanish and 12 mins into the fight, he pulls aggro
7/11 23:18:43.476 Magtheridon hits Oscarvil for 18791. (crushing)

Our proposal next week is the rogues run away from Magtheridon and then vanish. Before they pull aggro of course.

Edit, Omen of course did show he was fine.

 

Anedris

We had that happen on our kill tonight (two rogues in the raid, both pulled aggro a minute or two after vanishing and got splattered). One of them talked a GM and was told that it's a known bug specific to Magtheridon. (No idea if moving away from him before vanishing would help... worth a try I guess.)

 

sp00n

That are some serious lag issues there.

Should this 18.9 maybe read as 17.9, or 16.9?
7/11 23:17:17.795 Oscarvil performs Vanish on Oscarvil.
7/11 23:17:18.904 Oscarvil gains 15 Energy from Oscarvil's Combat Potency.
7/11 23:17:18.360 Oscarvil gains Stealth.


Anyway, next time I vanish I'll check if .5 seconds is normal between 'performs vanish' and 'gains vanish'.

 

Tehehe

I noticed something odd today and somehow remembered to take a screenshot.

This was on a pull in SSC (Lurker trash), about 10 mobs in the pull. My target is a unique target in that pull (doesn't share the name with anything else). It is being tanked by the a druid named Pourpre (shown as having highest threat) and he had been tanking it for a bit (at least 15 seconds, the mob was at 49%). The mob does not have any secondary abilities that he would have de-targeted the tank for. The tank has a pretty updated Omen (he updates 1-2x a day), but I am not sure exactly what revision.

What happend was that even though Pourpre was tanking the whole time, the *Aggro Gain* bar was being displayed for someone else. Unfortunately I'm not sure exactly who Threat thought was tanking.



edit: I also just noticed Whysp's bar is slightly longer in width than Passenger even though he has less threat.

 

Marroc

It's more than likely calculating Shatter damage, as it is done by one player to another. If he was standing on top of someone you could get him doing 8k damage instantly, which could put his TPS through the roof.

 

Marroc

It won't. Tried it last week on my rogue. For some reason Mag just doesn't like 100% threat drops when you have dots on him...

 

Rallick

We did Hydross on Tuesday and after a wipe to excessively sloppy play where Omen worked flawlessly it went completely nuts. We haven't had the entire guild switch over to Omen yet, mainly because of the occasional anomaly like this. The people in the raid were exactly the same from the attempt where Omen worked and the attempt where it didn't. I should have posted this earlier as my memory isn't that great but now that I think about it I don't remember having any problems while he was poisoned with Zado as the tank, I can't be sure though.

The tank is nowhere to be seen in the first screen, looks like the % threat is in the hundreds of thousands. The only thing that was different between the two pulls to my knowledge was that a pally got too close to Hydross but our tank managed to pull (intercept) just in time. The bar below Hydross' is his target.

 

Marroc

lol... 284% threat... ya something isn't right there.

 

Solstice

I noticed the same behavior from KTM during certain encounters - specifically Kael'thas whereby clearing the meter and setting focus to the next add would lead to a certain hunter starting with an allegedly huge amount of threat. If the majority of your guild is still using KTM, or indeed anyone at all, it's quite likely that the source of the misinformation comes from KTM rather than Omen.

 

Marroc

It's not just him though... EVERYONE in that list has 100%+ threat, and that shouldn't be happening.

 

Xerophyte

Everyone above Chaostrider in the list will have 100%+ threat, so that's expected behaviour given that a large part of the raid had their threat bug out.

 

Marroc

Right, what I was getting at is it isn't the same issue as the KTM bug because it's effecting multiple people, and that it wasn't normal behavior because someone with 200%+ threat would pull agro, which clearly wasn't the case.

 

Antiarc

This is my suspicion as well. If Master Target wasn't set, then if the warlock were to shatter a few folks for, say, 9k damage in one shot, then his TPS would spike through the roof, then settle back down. I notice he's running KTM there, which is why I say that; without a master target, player<->player damage is counted towards that player's threat in KTM.

The Hydross thing there looks like a combo of KTM missing threat clears (hence that hunter waaaay at the top there), and something screwy with the tank's threat reporting. Do you have revision numbers for the tank available, by any chance?

 

Rallick

I'll try to find out what revision it is, hopefully he hasn't updated it. The weirdest part for me was that his threat was showing up just fine on the previous attempt.

 

Harem

Would it be possible to exempt boss mods related clearing of threat from the TPS reading? For example, on Leotheras, everyone has negative threat readings post whirlwind, which is just odd to look at mainly.

Also, last night I had two different issues on Vashj. One being the fire mage spontaneous threat at the start of the fight (IE entering combat with 60k~ threat for no reason), as well as the issue where if you pull out a bar from Omen in combat it sticks to your mouse cursor. Sorry I don't have version numbers, I believe I updated yesterday.

 

Marroc

The better solution to this is to floor TPS display at 0, rather than actually remove the calculations for a threat clear that DID happen.

Both are known bugs. No known solution to the first to my knowledge. For the second however, don't pull bars out as you engage in combat. It has to do with blizzard locking down frame movement during combat, and it can 'trap' things onto your mouse if you start dragging them before combat, enter combat, then let go of it.

 

Antiarc

I need to reset TPS sigmas on a threat clear; that's my oversight, and I'll get it fixed up.

I was hoping I'd fixed the mage bug yesterday. Apparently not, it'd seem. Bleh.

I haven't seen the bar-sticking-to-mouse bug in a while. Any steps to reproduce?

 

Marroc

If you fixed the mage issue recently he may not have the latest revision (or the mage may not)

Same as before to reproduce the frame issue :P

Start Drag -> Start Combat -> End Drag = Stuck frame

Don't think there is much you can do to fix it as it's got to do with blizzards limitation of frame manipulation while in combat.

 

Apate

That definitely does it for me, as it has happened twice by accident now. It's not truly "stuck" though, because next round of combat you can remove it. Other mods with similar problems required a restart in my experience.

See you, auntie.

 

Marroc

would you prefer I used the word 'semi-stuck' instead? :P

 

Llilyth

Strange. I find that Omen's TPS calculations are way off as well, but they are way too high. Has anyone noticed that Omen's TPS numbers are significantly larger than the TPS numbers in KTM? And according to what I've read on EJ about a tank's TPS and what our DPS members's DPS is, KTM's TPS numbers make more sense.

On Fathom-Guard Tidalvess (the Shaman add in the Fathom-Lord Karathress encounter) tonight, I was watching Omen carefully, and it reported that my TPS was hovering around 2200, and reaching 2500 at one point. I did not have Windfury.

I also noticed our warlock tank pull 2500 TPS on Grand Astromancer Capernian Caperinian (the Mage in the Kael'thas encounter).

I have the TPS update frequency set to the maximum (2 seconds). Does that make the TPS calculations less accurate? I thought it'd just mean the numbers don't update as often, but over the long run, TPS should be correct? Or am I misunderstanding something?

 

Antiarc

Well, the pullout bars aren't secure frames, so they shouldn't be affected by the combat lockdown. I suspect the drag focus just gets changed there, so I need to watch for whatever it is.

Edit: I did a fairly heavy bit of comparison of TPS numbers between KTM/Threat, and some simple math checks as well; they've checked out for me, but if others are consistently experiencing problems, I may need to re-evaluate it. They do tend to spike more easily than KTM's numbers, in my experience, due to the algorithm I use, but overall they are usually quite stable.

The display update frequency doesn't have any effect on the TPS algorithm, so no, it shouldn't affect it.