Can't you simply Judge Seal of Righteousness? It's just a straight up nuke so that's pretty clean.
Judgements will always apply a debuff, either Justice, Light or Wisdom, depending on which specific Judgement button you press. The damage dealt by this button press is determined by which Seal is activated at the time.
Yes, you CAN apply Judgement of Righteousness' damage by having Seal of Righteousness active when you cast Judgement, but casting Judgement itself will still apply a debuff.
In any case, a Holy Paladin would be using Seal of Light or Seal of Wisdom anyway, since those can be glyphed to increase healing done while active or reduce healing spell mana cost while active, respectively.
Debuffs that aren't mentioned above:
Wounding Poison (better than Instant as Combat)
Aimed Shot (Blizzard is trying to get this into MM Hunters' single target rotation)
Gargoyle for DKs is no debuff, but Blood DKs have Heart Strike as an additional debuff.
I believe gargoyle actually has a debuff, to trace which target its attacking. It's something like "Gargoyle is attacking!" or whatever. It did last time I played my DK at least. Youc an add to blood for a part of the time Mark of Blood, depending on spec, and unholy is definitely the biggest offender, but Unholy Blight might not be always used, at least against single target bosses, I believe the math has shown it's not that great if there's not 2 targets.
Still, will run into issues with certain raid setups, especially since a lot of the debuffs that do the same thing and as such could be mutually exclusive also add something else different for every class(the melee haste reduction is the best example, with infected wounds adding a snare and frost fever adding a disease count for the DK). Then again, it took them a long long time to increase debuff duration, so not sure they're going to do it again until they upgrade most of the hardware again.
This is wrong. Icy Touch applies a debuff which reduces melee haste, as well.
I am not sure about this, for starters at no time in beta do I remeber doing an IT and the mob having two debuffs. Additionally in the game files we have,
Frost Fever Disease
Deals Frost damage over 12 sec.
Reduces melee and ranged attack speed.
This leads me to believe that if you IT and a mob is immune to diseases it gets the IT debuff, and if it is not immune, Frost Fever overwrites the IT debuff and you are left with only a FF debuff, which would be consistent with what I remember from beta, but I am not certain. Regardless it will be a shared debuff with clap, or what not so it is obviously something you add in the tank section not count against every single dk.
Originally Posted by Pyros
Still, will run into issues with certain raid setups, especially since a lot of the debuffs that do the same thing and as such could be mutually exclusive also add something else different for every class(the melee haste reduction is the best example, with infected wounds adding a snare and frost fever adding a disease count for the DK). Then again, it took them a long long time to increase debuff duration, so not sure they're going to do it again until they upgrade most of the hardware again.
Unholy blight does not add a disease anymore, and thus no debuff. Gargoyle/Mark of blood are going to be low uptime debuffs, so I don't think they should be lumped together with permanent debuffs.
Last edited by Flamingcloud : 11/11/08 at 12:56 PM.
Unholy blight does not add a disease anymore, and thus no debuff. Gargoyle/Mark of blood are going to be low uptime debuffs, so I don't think they should be lumped together with permanent debuffs.
That depends on how the overwriting code is set up. Let's say you've got a boss that already has 40 debuffs up. When the DK then casts Gargoyle, does it knock off a previous debuff to make room, or just not get applied at all? If it does knock off an existing debuff, does it remove the debuff that was applied to the boss earliest, or the debuff with the lowest duration remaining? Contingent upon those answers, Gargoyle/Mark of Blood could be just as serious debuff munchers as 100% uptime spells.
That depends on how the overwriting code is set up. Let's say you've got a boss that already has 40 debuffs up. When the DK then casts Gargoyle, does it knock off a previous debuff to make room, or just not get applied at all? If it does knock off an existing debuff, does it remove the debuff that was applied to the boss earliest, or the debuff with the lowest duration remaining? Contingent upon those answers, Gargoyle/Mark of Blood could be just as serious debuff munchers as 100% uptime spells.
Well you need to be running at 35ish debuffs to account for 7ish part time debuffs I would say. Overritting code is defintely not designed to not apply the new buff when you are at 40. It always knocks off an old debuff, originally I thought it knocked off the oldest debuff, but I found that not to be the case, I believe it has a random component, or a slot component(like debuffs go into a fixed array of size 40 and when all is full the debuff in slot 0 is removed, and then the next time a debuff has to go the one in slot 1 is removed which is not necessarily the oldest one)
Last time I saw the math on deadly/rupture(combat anyways) it was a minor loss (like under 100-150 dps in sunwell) to use instant/eviscerate instead, that could be different now, I am not sure.
Everyone always claims their dps would be crippled and yet for most classes they are just completely bsing you. Still, in the end the best way is to just not run affliction locks, and don't run more spriests than are needed for replenishment until a better solution comes.(assuming you are at the cap)
The math you saw was pre-3.0, they've made rogues very dependent on poisons now. Depending on spec, Deadly and Rupture can currently account for at least 10-25% of total damage. With my hit and crit both being lower at level 80 than they are now this will only increase the value of poisons and Rupture.
Assassination rogues will always use Instant/Deadly as they must maintain a poison on the mob to use Mutilate/Envenom effectively (talenting for both). Dropping Deadly isn't an option as Envenom requires it even if someone else is putting up a poison for Mutilate. Envenom is also how Slice and Dice is refreshed through Cut to the Chase (Eviscerate does ~40% of a full Envenoms damage). You're right that dropping Rupture from an Assassination rogues rotation will result in a DPS loss of around 150-200 DPS; it's the only 'reasonable' deviation from ideal though.
Combat rogues will prefer to run Deadly/Wound as it makes the most sense for untalented poisons. All combat rogues will take the Blood Poisoning talent as well for the AP boost (not just the debuff). Rupture is used because it bypasses armor. If the boss can be 'dazed' (like Mind Flay the debuff may actually be there) that's potentially five debuff slots for a Combat Rogue (Deadly/Wound/Blood Poisoning/Rupture/Daze). Substituting Wound with Instant would be the only one that wouldn't cripple DPS.
The math you saw was pre-3.0, they've made rogues very dependent on poisons now. Depending on spec, Deadly and Rupture can currently account for at least 10-25% of total damage. With my hit and crit both being lower at level 80 than they are now this will only increase the value of poisons and Rupture.
Assassination rogues will always use Instant/Deadly as they must maintain a poison on the mob to use Mutilate/Envenom effectively (talenting for both). Dropping Deadly isn't an option as Envenom requires it even if someone else is putting up a poison for Mutilate. Envenom is also how Slice and Dice is refreshed through Cut to the Chase (Eviscerate does ~40% of a full Envenoms damage). You're right that dropping Rupture from an Assassination rogues rotation will result in a DPS loss of around 150-200 DPS; it's the only 'reasonable' deviation from ideal though.
Combat rogues will prefer to run Deadly/Wound as it makes the most sense for untalented poisons. All combat rogues will take the Blood Poisoning talent as well for the AP boost (not just the debuff). Rupture is used because it bypasses armor. If the boss can be 'dazed' (like Mind Flay the debuff may actually be there) that's potentially five debuff slots for a Combat Rogue (Deadly/Wound/Blood Poisoning/Rupture/Daze). Substituting Wound with Instant would be the only one that wouldn't cripple DPS.
Even if deadly + rupture accounted for 25% of your dps you aren't losing 25% of your dps switching to evis + instant. If evis + instant hit only 80% as hard, you would only have a 5% total dps loss by switching, which is less than what most classes would lose by removing two of their debuffs. Ability X could make up 90% of your damage, but that doesn't mean switching to ability Y thats does 1% less damage will cripple your dps. Mind Flay debuff is there because all channeled spells have either a visible or invisible debuff attaching them, daze won't be there if the boss can't be dazed.
The argument about lower crit, is very valid however, and lower armor pen (widening the gap from rupture to evisc) The math does need to be ran again I am sure, but I would be surprised if rupture and deadly were not in the worst 5-10 debuffs.
The argument about lower crit, is very valid however, and lower armor pen (widening the gap from rupture to evisc) The math does need to be ran again I am sure, but I would be surprised if rupture and deadly were not in the worst 5-10 debuffs.
Very possible for rupture, but very doubtful for Deadly. The thing to keep in mind about Deadly is that it enables Envenom which 1) Ignores armor, 2 ) Does more damage than an Eviscerate even before the armor ignore is taken into account, and 3) significantly increases Instant Poison DPS for 6 seconds following it's use.
I'd be surprised if an Evisc could match 50% of an envenoms damage contribution when you factor all three in, along with the baseline damage of Deadly Poison.
Very possible for rupture, but very doubtful for Deadly. The thing to keep in mind about Deadly is that it enables Envenom which 1) Ignores armor, 2 ) Does more damage than an Eviscerate even before the armor ignore is taken into account, and 3) significantly increases Instant Poison DPS for 6 seconds following it's use.
I'd be surprised if an Evisc could match 50% of an envenoms damage contribution when you factor all three in, along with the baseline damage of Deadly Poison.
Deadly also enables Mutilate - a Mutilate with no poison debuff on the mob is 33% weaker. Mutilate is ~20% of your total damage. So even leaving aside the question of DP/IP and Evis/Envenom, losing DP is a straight up loss of 6.7% of your total damage. Adding everything in, then the DP debuff slot is going to be worth well over 10% of a rogue's total DPS.
Deadly also enables Mutilate - a Mutilate with no poison debuff on the mob is 33% weaker. Mutilate is ~20% of your total damage. So even leaving aside the question of DP/IP and Evis/Envenom, losing DP is a straight up loss of 6.7% of your total damage. Adding everything in, then the DP debuff slot is going to be worth well over 10% of a rogue's total DPS.
Several other debuffs qualify as 'poisoned' for Mutilate purposes, so that chunk of it, while certainly relevant, is less of a concern for any single rogues personal DPS.
Originally Posted by Koraa
We would like to increase the number of debuff slots, yes, and we most likely will. Unfortunately, this is very complicated to do tech wise. It greatly increases the amount of bandwidth for the server, for example, and also for the client (you).
While I obviously don't know how they've written their game engine, I'd be a bit surprised if the performance issues for increasing debuff slots were as clear-cut as Koraa suggests. For example, bandwidth between the server and the client could actually go down, since there'd be fewer debuffs being bumped off prematurely, meaning fewer updates to the client, though there would presumably be more "x did y to z" messages; which one would be more significant depends on their relative frequency.
On the server, the need to scan a longer list could easily be offset by the reduced need to run the "bump a debuff because the list is full" routine. Presumably their code is smart enough to not scan the list past the last active debuff. The only guaranteed resource concern is the slightly larger amount of memory needed for the extra slots, and that should be inconsequential on the client side, though it could possibly add up to a lot on the server.
I am fairly certain that on a set interval you are updated with the 40 debuffs and 40 buffs of every character in range that you have detected(ie you don't receive info on stealthed/invisible players). The only way to make this take the same bandwidth while still increasing the debuff size that I can see is to change it to 50/30 debuffs/buffs or 45/35 I don't think people would be upset if they could only have 30 or 35 buffs.
The amount of calculations on the server side for buffs/debuffs getting knocked off is completely trivial(O(1)) compared to the passing of information between the client/server(O(n)).
I am fairly certain that on a set interval you are updated with the 40 debuffs and 40 buffs of every character in range that you have detected(ie you don't receive info on stealthed/invisible players). The only way to make this take the same bandwidth while still increasing the debuff size that I can see is to change it to 50/30 debuffs/buffs or 45/35 I don't think people would be upset if they could only have 30 or 35 buffs.
This is an interesting idea but we'd have to make a list of all the invisible buffs in addition to the visible ones to make sure such a change in truly harmless.
I recall in the Naxx days on say Loatheb when you're buffed to the gills with Alchemy, World Buffs, etc. many weird things would occur due to invisible buffs being pushed off. For example, Warrior Stances are just buffs that aren't displayed, and our MT actually had his Defensive Stance pushed off a couple of times, naturally a less than desirable outcome.
Now of course we have less trouble with Alchemy and World Buffs, but many more raid-wide buffs, so I believe this could still be an issue.
This is an interesting idea but we'd have to make a list of all the invisible buffs in addition to the visible ones to make sure such a change in truly harmless.
I recall in the Naxx days on say Loatheb when you're buffed to the gills with Alchemy, World Buffs, etc. many weird things would occur due to invisible buffs being pushed off. For example, Warrior Stances are just buffs that aren't displayed, and our MT actually had his Defensive Stance pushed off a couple of times, naturally a less than desirable outcome.
Now of course we have less trouble with Alchemy and World Buffs, but many more raid-wide buffs, so I believe this could still be an issue.
Yes but that was back in the day of what, 16 debuffs? 30 Might be realistically possible for some classes in weird situations, but I don't really think 35 is possible unless you are doing dumb things like putting detect invis/underwater breathing up. Then again I am not even 100% sure it is 40 server side, it might be 40/30 already, and there is no way they can go to 45/25. Regardless if I thought of this solution in 30 seconds I am sure blizzard already thought of it and decided there was a better solution to the problem. Or maybe even decided it wasn't a big enough problem to bother fixing.
I'm think we may have had more than 16 buff/debuff slots then, but not sure. I remember there was another problem where people were unable to click off some useless buffs because the default UI only displays the first 16.
Its very possible for that number to hit 40+. Divine Protection, Guardian spirit plus some more totems easily accomplishes it. Clicky trinkets make it even worse. Lowering the buff limit isn't an acceptable solution IMO.
I am fairly certain that on a set interval you are updated with the 40 debuffs and 40 buffs of every character in range that you have detected(ie you don't receive info on stealthed/invisible players). The only way to make this take the same bandwidth while still increasing the debuff size that I can see is to change it to 50/30 debuffs/buffs or 45/35 I don't think people would be upset if they could only have 30 or 35 buffs.
That's interesting. I'd think it would be more efficient to only do a dump like that when something comes into range, and just send the deltas while it remains in range, but maybe I'm overlooking something. I wonder if it's possible to reasonably test for either case and if knowing that it's one or the other would have any practical implications for tuning connection performance.
The amount of calculations on the server side for buffs/debuffs getting knocked off is completely trivial(O(1)) compared to the passing of information between the client/server(O(n)).
How do you figure it'd be O(1)? We know they have a priority system, even if last I heard they hadn't added values for TBC ranks of anything, which means there's an extra cost in there somewhere, and if they optimize for replacement, it'll likely cost them in the more common case of just updating the list.
Do we know if they can distinguish between raid mobs and other mobs? If, for example, raid mobs could take up to 10 buffs and 60 debuffs, non-raid mobs could take up to 5 buffs and 20 debuffs, and players could take up to 40 buffs and 20 debuffs, it'd give them some extra room for optimization.
That's interesting. I'd think it would be more efficient to only do a dump like that when something comes into range, and just send the deltas while it remains in range, but maybe I'm overlooking something.
I'd assume they want to be robust, even if some communication packets are lost. In that kind of environment you want to report the current state, not just changes from the previous state.
I just posted this http://elitistjerks.com/977935-post7.html in the various simple Q/A threads, and I hope we'll get some more solid numbers for more classes on this matter.
Try to keep it in an orderly fashion when you post debuffs, or things will get very hard to keep track off.
I'd assume they want to be robust, even if some communication packets are lost. In that kind of environment you want to report the current state, not just changes from the previous state.
The game uses TCP for updating stuff like that, meaning that no communication packets can get lost (or come in the wrong order). If something goes wrong you'll simply lose the connection or get lag.
This is clearly visible when you get a typical lag spike where the game seems to pause (TCP packets from the server are getting clogged up in a buffer somewhere on the network), and then suddenly everything appears to happen at once (the network buffer is getting emptied very fast when the network problem is over).
On topic, Blizzard will surely attempt to handle as much on the server as possible. Even rather complex operations on the server are infinitesimal compared to the time it takes to send even the simplest information to the client.
It dawned on me the other night that Blood Death Knights using Dancing Rune Weapon over Gargoyle will potentially double the number of diseases they place upon the mob. This could push other debuffs off the boss, and also make Dancing Rune Weapon less desirable.
What I lack in intelligence I make up for in verbosity.
There is a blue post saying that there is no debuff cap anymore.
We changed how the game handles debuffs (ie negative state effects on targets) on creatures with Wrath of the Lich King. The old hard cap of 40 debuffs on a target no longer effectively exists. You can now apply way more debuffs to a target without them dropping off before their duration expires. The default WoW UI will not normally display all these debuffs, but they really are still there!
11/11/08, 12:17 PM
Similar Threads
