PDA

View Full Version : 3.2 and DK DW'ing



lyd
06-19-2009, 06:53 AM
from the patch notes for 3.2:







Threat of Thassarian: New 3-point talent. When dual-wielding, your Death Strikes, Obliterates, Plague Strikes, Blood Strikes and Frost Strikes have a 30/60/100% chance to also deal damage with your off-hand weapon. Off-hand strikes are roughly one half the effect of the original strike.






as a dedicated DW'ing tank, i am in love with this change. what do you guys think? will anyone be trying out 1h dps specs again?

Mukluk
06-19-2009, 07:42 AM
Short of algalon I've already tanked everything in uld 10 and 25, would love to try dw'ing myself but my concern is ToT doesn't affect rune strike and many fights like 3min hodir im grasping for threat as it is.

Satorri
06-19-2009, 08:53 AM
Mmmm, it does look like a fun talent, I wish it were a little more embracing of supporting the other trees, but I'll take what I can get.

Mukluk, RS doesn't actually need it, per se. RS does benefit from the faster swing speed. In theory you'll be able to use a slightly higher % of your RS's, and that compensates some for the smaller value of each. It's hard to do the numbers on though, I've been trying for a while.

(Small tip on Hodir that might help, my melee dps'ers are on strict orders that when they get the Storm Power buff they are to come hit me with it as well! As Blood that's no small buff, Frost would benefit nicely as well I'm sure, and Unholy would do better than nothing)

Molohk
06-19-2009, 09:00 AM
I'm a bit worried about the ammount of talent points it takes to make DW viable. I hope they balanced the dual-weapon strikes to compensate for the loss of other valuable talents, but I guess we can only wait for PTR test results to start showing up.

Mukluk
06-19-2009, 09:56 AM
Also (just thinking out loud here) wonder if dw tanking worked, if it would be better to dw dps style weapons or tank style (warrior/pal 1hander) weapons.

Chamenas
06-19-2009, 10:02 AM
The only reason "dps" weapons work with two-handers is because of SSG. When dual wielding, you no longer have SSG and thus lose 25 defense (not rating, 25 DEFENSE), thus it's imperative that you find weapons with defense rating, unless you are somehow 25 points above uncrittable, aka 565.

I have some thoughts on two talents that could make dual-wield tanking more interesting, but I'm not a numbers person so I don't have the numbers balanced. If you're interested and can crunch numbers, send me a message.

Mukluk
06-19-2009, 10:18 AM
The only reason "dps" weapons work with two-handers is because of SSG. When dual wielding, you no longer have SSG and thus lose 25 defense (not rating, 25 DEFENSE), thus it's imperative that you find weapons with defense rating, unless you are somehow 25 points above uncrittable, aka 565.


I don't agree, I'm over 540 (uncrit) defense and i use the +4% parry rune, and my 2hander is the furious glad sword, no def on it, although when I got it I gave up some def gems and use the resilience. Not to say one way or the other on dps 1handers vs tank 1handers, just saying ssg isnt required. I think its more of a crutch until you get the gear to run without it.

Chamenas
06-19-2009, 10:22 AM
I don't agree, I'm over 540 (uncrit) defense and i use the +4% parry rune, and my 2hander is the furious glad sword, no def on it, although when I got it I gave up some def gems and use the resilience. Not to say one way or the other on dps 1handers vs tank 1handers, just saying ssg isnt required. I think its more of a crutch until you get the gear to run without it.

I'm in fairly decent gear and I'm still not 25 points over 540. In order to get there I would need to sacrifice hit, stamina and expertise. imo, it's not worth it to get dps weapons for that slight advantage when defense should be the most important stat until you hit 540. There's also no reason to sacrifice stamina, hit or expertise as a tank just to get weapons that deal slightly more dps for threat.

Satorri
06-19-2009, 01:21 PM
It is not hard to get the defense required from gear, SSG is just a handy alternative that lets you add stuff elsewhere instead of defense. It is not a make or break for dps weapons. That said I think one of the major values of dual wield tanking is that you can use tanking weapons. It gives you the chance to stack extra survival stats in.

Dual wield tanking is possible, and this will make it far more workable than it has been in months, but I'd wager you still won't see any more than the people who do it already.

Chamenas
06-19-2009, 01:38 PM
See, my thoughts were... I'd like to see some talents directed at tanking with two weapons. If we... took out less popular talents (I had two in mind, but PvPers would kill me) and freed up... 4 talent points we could add two 2-talent point talents.

Weapon Block
1/2
When wielding two weapons you have a 5% chance, with additional chance to block dependent on your defense to block an incoming attack, the chance to block increases as your block value increases.
2/2
When wielding two weapons you have a 10% chance, with additional chance to block dependent on your defense to block an incoming attack, the chance to block increases as your block value increases.

Improved Parry
1/2
When you are hit by an incoming melee attack, if you are wielding two weapons you receive a buff that increases your chance to parry by 3% for 12 seconds or until you successfully parry an incoming attack. This effect stacks up to 3 times.
2/2
When you are hit by an incoming melee attack, if you are wielding two weapons you receive a buff that increases your chance to parry by 6% for 12 seconds or until you successfully parry an incoming attack. This effect stacks up to 3 times.

Help with numbers and wording would be appreciated as they currently stand as overpowered at the moment. What I hope to convey is the basic idea, some defensive talents centered around dual wielding that would make it an equivalent if not better option for frost tanks to two-handers, making sure to address the issues that DKs have while tanking.

Edgewalker
06-20-2009, 09:54 PM
DW tanking is already very possible, it just pulls lower TPS at a higher cost with a more strict gear dependency.
If you are having trouble with 3 minute Hodir, go frost and make sure you get the first storm power.
Singed makes all the difference.

The problem I have Satorri is that theoretically you would have more RSes with DW from faster speed, but especially on slow hitting bosses, avoidance (even at 30/25) doesnt allow for as many as you might think.

Satorri
06-21-2009, 05:23 AM
Indeed, the numbers are hard to bear out, and there's something I never took the time to figure out which could make a big difference: if you use a slow main hand and fast offhand, do your offhand swings proc a hit based on main hand damage, or offhand damage. Obviously the two would vary the damage quite a bit.

The effect of speed and number of incoming swings, along with your avoidance level, and the random nature of when you get avoids makes a world of difference. I, not unfrequently, avoid two swings in a row within the span of my weapon swing timer, so I know I miss RS's just for that. That's particularly the case when I'm popping my on-use avoidance trinkets.

It still gets sticky of course because using a weapon that is twice as fast and hits half as hard still very well may not catch twice as many procs.

Darksend
06-21-2009, 08:41 AM
if you use a slow main hand and fast offhand, do your offhand swings proc a hit based on main hand damage, or offhand damage. Obviously the two would vary the damage quite a bit.


As i have no idea what procs you could be talking about as i have no idea about DKs I can only offer this little bit of info

The top rogue dps spec at some points in BT was either a main hand fist or mace with an offhand sword with both mace and sword spec.

This was because offhand sword hits proced main hand mace swings.

If it is still like that i have no idea but its there for what its worth

Chamenas
06-21-2009, 09:14 AM
DW tanking is already very possible, it just pulls lower TPS at a higher cost with a more strict gear dependency..

That's the thing. Just because something "can" be done doesn't mean it is preferable or even reasonable. I won't argue that it can be done, but it's not worth what would be lost in order to switch over to it from two-hand. That's why I'm hoping that perhaps some other things might come into play which would make up for what is lost in the crossover. I'd love to DW tank, but right now I'd be sacrificing a lot of stats to do so.

Satorri
06-22-2009, 06:09 AM
That is somewhat helpful Darksend, though I knew that (for my rogue =D).

The procs I'm speaking about are this:
When a DK dodges/parries an incoming attack it makes Rune Strike available, a la Revenge for Warriors. It is an on next swing move that replaces your auto-attack with a 50% more damage, 50% more threat strike.

As I write that I think I figured out most likely how it works. If it is truly next swing and not next 'main-hand' swing (since our swings don't proc it, incoming ones do), then that may be a good reason not to have a too-fast off-hand as it would more often eat the procs with smaller hits, maybe faster than we want. Like Edge said, it doesn't proc THAT often, so maybe just 2 2.5 sec swings is enough to get maximal effect out of RS procs next to a big slow 3.5.

Now if we can just get Blizz to offer more nice slow 1-hand tank weapons! (And let DK's use fist weapons! I love the graphics on the ulduar ones)

lyd
06-22-2009, 07:17 AM
That's the thing. Just because something "can" be done doesn't mean it is preferable or even reasonable. I won't argue that it can be done, but it's not worth what would be lost in order to switch over to it from two-hand. That's why I'm hoping that perhaps some other things might come into play which would make up for what is lost in the crossover. I'd love to DW tank, but right now I'd be sacrificing a lot of stats to do so.

it's not as bad as you think. generally, it works out so that you net a little less threat and a little more avoidance, just due to the way your gear tends to settle out. i have yet to be denied a spot in a raid due to any of my stats being found wanting compared to other tanks. also, the new DW talent in 3.2 is looking like it will bring DW threat in line with 2h, but that remains to be seen.

i currently rock @ 550 defense w/ relic proc, 40k hp raid-buffed, ~46% self-buffed avoidance (counting +3% miss from talents), and 41 expertise (-10.25% to parry/dodge). i use +2% parry on both weapons. i sacrificed 1% dodge so i could grab rune tap (totally worth it). I haven't stepped in ulduar 25 yet, but i've been tanking ulduar 10 with no issues. last night i actually found myself much more survivable than the warrior tank who was running with 2k more hp. i have yet to see any evidence that my style choice has created a liability for my healers, or an untenable gap in performance between me and my 2h-wielding peers.

Edgewalker
06-22-2009, 11:17 AM
Now if we can just get Blizz to offer more nice slow 1-hand tank weapons! (And let DK's use fist weapons! I love the graphics on the ulduar ones)

I rocked out Vulmir (the 2.6 speed 232 MH mace from Thorim 25) MH with a Sorthalis (1.6 speed 239 OH mace from XT Hard Mode 25) OH, both with Cinderglacier. Using one DPS weapon and one tank weapon OH is possible, but again it limits your gear choices for defense cap. I still believe that on a fast hitting boss, DW tanking will be the higher single target threat if glyhed for RS with the Tier Set bonus especially :)

@ Lyd -
Not capping expertise is still quite bad, parry chains even with -9 or 10% are significantly more plausible dual wielding. I think one of the larger problems with dual wielding right now is the lack of a good Runeforge for mitigation or avoidance on a one hander, 2% parry is nice, but 4% parry total doesn't have quite the same value or flexibility 25 Defense and 2% Stamina does.

lyd
06-22-2009, 03:23 PM
@Edgewalker: the parry-chain issue has been math'd to death in other threads. suffice to say that it is sorely exagerrated, and at higher levels of expertise, a negligible mechanic, even when compared to 2h.

the following comparison may seem silly, but bear with me: compared to a molecule, a grain of sand is enormous. however, compared to the rest of the world, both are insignificantly small. This is how i would describe the difference between 2h and DW in terms of parry-chains. as your expertise rises, the chance of a parry-chain becomes so insignificantly small that the gap between 2h and DW ceases to be a large concern.


don't take this math at face value, since i'm horrible at math, but consider the following:

wowwiki puts lv 83 boss parry chance at 14%. at -10.25%, that means that my attacks have a 3.75% chance to be parried. the chance of a boss parrying two of my attacks in a row is equal to (.0375*.0375)= .00140625 or .14%. so right now, there is a .14% that a boss will parry 2 of my attacks in a row. if we assumed an optimal set-up with 2 slow weapons, you'd have swing times of 2.50, or 20 swings a minute. Over the course of a 10-minute fight, i will swing 200 times with both weapons, so 400 times total. The chance of 2 parry-hasted attacks occurring consecutively is .14%, or 1/714. assuming an average fight lasts ten minutes, roughly once every other boss, i will experience 2 consecutive parry-hasted attacks. ok, good.

Now let's consider the fact that i have @ 47% raid-buffed avoidance. so even if the unlikely case of 2 consecutive parry-hasted attacks occurs, there is a 47% chance that one or the other parry-hasted attack will be avoided entirely. this reduces the chance of actually suffering two consecutive parry-hasted attacks even further. so with 47% avoidance, there is a 53% chance that a parry-hasted attack will actually hit. given that there is a .14% chance that a boss will parry two of my attacks in a row, .14%*.53% = .0742%. so there is a .0742% chance that a boss will parry 2 of my attacks consecutively AND that both attacks will not be avoided. ok, great.

but what about mitigation cooldowns? IBF is up 18 seconds at a time. unbreakable armor is up 20 seconds at a time. my trinkets last 15 seconds. assuming i use these cooldowns independent of each other, that means that 53 seconds out of every minute, i have some sort of additional mitigation present, or roughly 88% of the time. So there is roughly an 88% chance that, should the boss succeed in landing two consecutive parry-hasted attacks, that they will land during a time when i have additional mitigation.


BUT WAIT, we forgot about runestrike. runestrike can't be parried, and takes the place of a main-hand white attack. I don't have any WWS parses, but even if we assume that only half of my main-hand weapon attacks become Runestrikes, i've cut the number of parry-haste opportunities by 25%. without even trying to run the numbers, you can see how the parry-haste mechanic becomes infinitessimally small at high levels of expertise. Now my math here isn't exactly PhD level, but once all the numbers are run, the chance of ACTUALLY experiencing 2 consecutive parry-hasted attacks from a boss at high levels of expertise is outrageously small, to the point of being, for practical purposes, a negligible concern.


if this doesn't convince you, consider anecdotally the fact that there are currently DW tanks in Ulduar 25. I personally have tanked through all pre-ulduar content without any issues, and about half-way through ulduar 10. As far as i can tell, i have never died to a parry thrash mechanic. I know that's just me telling you my experience, but take it for what it's worth.

Satorri
06-23-2009, 05:57 AM
Which reminds me, Lyd, I haven't asked in a while and I'm sure your gear has gotten better since we last spoke about it. Can you look into some wws reports and figure out the proportion of regular melee swings to RS's you're getting (just the count of each will do, make sure you include the count of crits)? That info along with your avoidance would be helpful, I want to get some empirical data to roughly gauge the dual wielding use of RS procs. If you want to just PM me a WWS report from a good raid night, with info from which fights you tank I can do the busy work.

lyd
06-23-2009, 06:06 AM
i'd be happy too. we should be in Ulduar-10 again this week, i'll try to remember to run the wws client this time ><

Edgewalker
06-23-2009, 07:01 PM
The point isn't the mathematical possibility of it happening, it's that you are FORCED into a heavier expertise set to keep the chance low at virtually no gain.
That was a wall of text for no reason :\

Mukluk
06-23-2009, 09:07 PM
New 3.2 notes

Scent of blood 10 second cooldown has been removed.

This talent in a dw tank spec would add noticeable increased runic power generation now. More to think about.

lyd
06-24-2009, 08:30 AM
The point isn't the mathematical possibility of it happening, it's that you are FORCED into a heavier expertise set to keep the chance low at virtually no gain.
That was a wall of text for no reason :\

sorry, i wasn't trying to clobber you over the head. You are of course, correct that DW requires high expertise. Personally, i've found that alot of high-level tanking gear comes with expertise already, so it kind of takes care of itself. you did, however, make the point that even with high expertise, being below the cap has inherent danger. i was merely trying to show that once you surpass the soft cap, the danger becomes so increasingly low that it's not worth talking about.

i'll see if i can get a WWS report up within the week so people can look it over. I think seeing real numbers will help give some coherency to all the anecdotal evidence and armchair theorycraft.


New 3.2 notes

Scent of blood 10 second cooldown has been removed.

This talent in a dw tank spec would add noticeable increased runic power generation now. More to think about.

i saw this. i think they are going to have to combine threat of thassarian and nerves of cold steel into one talent, or at least make ToTh a 2-point talent instead of 3. i'm finding myself having to choose between 3 points in anticipation, 3 points in scent of blood, or 3 points in acclimation. higher dodge makes scent of blood more effective, but i would not want to lose the magic resistance afforded by acclimation.

i hope it gets fleshed out on the PTR. scent of blood is now a great DW tanking talent, but without consolidating frost talents a little, it will be very difficult to incorporate it effectively into a frost-heavy DW build without losing some avoidance/magic mitigation.

Satorri
06-24-2009, 12:05 PM
Well, you don't NEED more expertise, and in a lot of fights you won't notice the difference for survival purposes.

I'll try to be super brief with the math, but the point is that expertise goes further for dual wielding than a 2h.

Let's say, using conservative estimates, that it breaks down like this:
2 DK's, same gear, etc.

DK A:
2h with 3.5 sec swing time, 30% RS replacement (30% of normal swings are RS's), and 0% miss chance on auto-attacks

DK B:
2x 1h, Slow(2.5)/Fast(1.5), 40% RS replacement, and 16% miss chance on auto-attacks

Both will have 50% avoidance, and we'll average the number of avoids at 15 per minute. For reality sake, since with an incoming attack speed of 2.0 the 2h tank will have 1.75 swings incoming per outgoing we'll say the 2h can make use of 8 of them (53%), while the dual wielder will actually get out 2 swings per incoming, so we'll give it the benefit of the doubt and say he gets to use only 13 of them (87%).

DK A will throw 17 swings per minute, though 8 will be RS, so that's only 9 parriable swings per minute.

DK B will throw 24+40 swings per minute, though 13 of them will be RS's, so that's 51 parriable swings per minute, factoring miss chance that drops to 43.

If the boss has a 14% chance to parry, then DK A will get 1.26 parries per minute, while DK B will get 6.02 parries per minute. If the DK's each have enough to soft-cap (6.5%), the boss only has a 7.5% chance to parry. So, DK A will now average 0.675 parries per minute and DK B will have 3.23. The dual wielding DK sees several fewer swings, as opposed to the 2h'er who already has a very low count. Complicate that further with the fact that the parry only hastes the boss if they're in the first 60-80% of their swing, which means the random lining up required drops the probability further.

What this doesn't account for, which many people miss, is that both DKs may be using parriable -strikes (though dual wielders will favor using non-strike moves more because they don't lose damage from using fast weapons with smaller damage ranges). Even if they both used the same strike based sequence, that would be a much larger proportion for the 2h DK. A rough idea, if the DK is using an average of 4 -strikes per rune set (say, BS/HS, ScS/OB/DS, and accounting for every other set applying diseases), that's 4 per 10 sec, or 24 per minute that are as parriable as auto-attacks.

So factoring that in and adding to the values above:
DK A has 9+24 parriable strikes, or 4.62 parries/min w/o Exp and 2.48 w/
DK B has 43+24 parriable strikes, or 9.38 parries/min w/o Exp and 5.03 w/

If we factor that 30% of those parries will have little recognizable hasting:
DK A = 1.74 parries/min
DK B = 3.52 parries/min

And remember that half of those increased swings are going to be avoided anyway (actually increasing the number of RS's and therefore the % of RS conversion to non-parriable swings), and you start to see that the difference between the two, with the same set of gear/stats etc, is actually just 1-2 extra swings taken per minute.

You are right, of course, that dual wielding increases your possible incoming damage when you are tanking *on bosses that can parry-haste*, but it misses the fundamental appreciation of just how small that actual value is. If you are die-hard, it's not worth the increase, that's fine, but there have long been people trying to sell others that it's a much bigger deal than it is.

Damnit, that was a lot longer than I wanted. Oops.

My issue with dual wielding tanking has long been that I can't find a good reason to do it or something that feels good for doing it that I can play up. Many people who do it, do it because they like the idea, or they just want to show that it can be done. I'm looking forward to the new talent since it will actually make dual wielding feel like its own thing again, like you're not just trading %'s on damage sources around.

Inaara
06-24-2009, 12:46 PM
You are right, of course, that dual wielding increases your possible incoming damage when you are tanking *on bosses that can parry-haste*, but it misses the fundamental appreciation of just how small that actual value is. If you are die-hard, it's not worth the increase, that's fine, but there have long been people trying to sell others that it's a much bigger deal than it is.

This is really what it all comes down to. Using basic estimates you will see that DW tanking will offer more avoidance while 2H tanking will offer more health. Comparing what I consider to be the "best" raid accessible tanking weapons we have Worldcarver - Item - World of Warcraft (http://www.wowhead.com/?item=45165) and Sorthalis, Hammer of the Watchers - Item - World of Warcraft (http://www.wowhead.com/?item=45442).

Ignoring weapon damage you will recieve the following stats from either setup.
Worldcarver - Item - World of Warcraft (http://www.wowhead.com/?item=45165)
137 Strength
136 Stamina
2.13% Haste
10.61 Expertise
25 Defense
2% Stamina

2x Sorthalis, Hammer of the Watchers - Item - World of Warcraft (http://www.wowhead.com/?item=45442)
80 Strength
152 Stamina (With 12 stamina from gemming bonus)
11.8 Defense
5.46% Parry (After Rune of Swordbreaking - Spell - World of Warcraft (http://www.wowhead.com/?spell=54446))
11.7 Expertise (After gemming 16 Expertise Rating in each)
As a human I would actually recieve 14.7 Expertise, 4 more expertise than with Worldcarver.

Factoring in Threat of Thassarian, DW looks more than viable, to me it actually looks preferred.

Edgewalker
06-24-2009, 05:50 PM
If you are die-hard, it's not worth the increase, that's fine, but there have long been people trying to sell others that it's a much bigger deal than it is.

Damnit, that was a lot longer than I wanted. Oops.

My issue with dual wielding tanking has long been that I can't find a good reason to do it or something that feels good for doing it that I can play up. Many people who do it, do it because they like the idea, or they just want to show that it can be done. I'm looking forward to the new talent since it will actually make dual wielding feel like its own thing again, like you're not just trading %'s on damage sources around.

I guess that's my problem. I have been tanking this game too long and seen the most minute of possibilities happen too much to allow for the possibility without a substantial benefit. Hell, I had a world 4th Felmyst turn into a 7th from a parry gib on 4 of my attacks parried in a row, the chance of which should have been mathematically negligible.

I also believe dual wield tanking will still be a lower base TPS Inara.

And Lyd, from a brief browsing of new tanking gear, Expertise seems to be dwindling as a stat. :\

Milq
06-24-2009, 09:44 PM
I made my second spec DW tanking for the fun of it. Seeing how I never dps anyway... useless to keep my 2nd spec dps.

Used to the big numbers blood provides it feels that DW is really weak damage and threat wise. Of course I only have a Broken Promise and a Red Sword but still, at least BP is slow so my specials hit bigger. I'm going to scare my guild and try it in U25 too but I am really afraid of the threat problems in advance...

In the daily heroic (VH) dps was stealing mobs left right and middle. I really put out pathetic threat that might be because of the rotation I'm not yet used to plus my general inexperience with frost, but still compared to the solid threat I manage with most 2h setups this heroic was way below standards.

Satorri
06-25-2009, 08:19 AM
With a smart Frost spec in 3.2 I'd be stunned if Dual Wielding did less damage.

The major problem before was that you lost 33% or so off your instant strike *if* you had a slow mainhand, more if you had a fast one. While you got larger white damage and could boost related mechanics, it didn't make up for that loss elsewhere with most specs/methods.

Now, with the new double hit strikes, using Inaara's helpful numbers:
If we have matching AP's at 4200 (for easy math)

World Carver instant weapon damage = 1091
Sorthalis instant weapon damage = 781

Using Oblit as an example:
Dmg = 80% weapon dmg + 467.2 x 12.5% per disease

We'll just say both diseases up, and disregard other talents (since they'll scale both the same way on %'s).

Pre-3.2:
Worldcarver OB = 2010
Sorthalis OB = 1638
(18.5% less)

With ToT:
Worldcarver OB = 2010
Sorthalis OB = 1638+819 = 2457
(22.2% more, and that's not accounting for Nerves of Cold Steel increasing off-hand damage)

It seems they're not trying to balance Frost's dual wield vs 2h, they're making Frost a dual wielding spec, not that you can't use 2h's but now Dual Wielding will have an edge. I'm really curious to see what that does to the prevelance of Frost tanking, how the community perception of dual wield tanking shifts.

DKTankaholic
06-25-2009, 09:16 AM
whats the difference in hit rating from a 2H and DW?

Inaara
06-25-2009, 09:18 AM
With the nerfs to Blood's EH, the extra 3% avoidance and 2% decrease in damage that Frost offers is looking more attractive. Add in the DW factor and OHMAGAWD, secksi times..... I mean look at my avatar! Imagine those were two Sorthalis! And I'm not wearing T7.5 anymore!

Feanorr
06-25-2009, 04:16 PM
With the nerfs to Blood's EH, the extra 3% avoidance and 2% decrease in damage that Frost offers is looking more attractive. Add in the DW factor and OHMAGAWD, secksi times..... I mean look at my avatar! Imagine those were two Sorthalis! And I'm not wearing T7.5 anymore!

Yea, I will definetly give DW frost a try because of the same reasons; plus, strangely I got some better 1H weapon waiting in my bank than my 2H weapon so it look like a sign ^^

Edgewalker
06-25-2009, 04:40 PM
I would be more apt to give frost tanking a try if cooldowns weren't such a vital part of end game tanking. Guile of Gorefiend / Unbreakable Armor just aren't up to the same level as a 30 second Vamp Blood / WoTN, especially when you factor in the new set bonuses and the increased CD on IBF to begin with.
Still waiting on something more before I am impressed.

Tarigar
06-25-2009, 05:17 PM
The only reason "dps" weapons work with two-handers is because of SSG. When dual wielding, you no longer have SSG and thus lose 25 defense (not rating, 25 DEFENSE), thus it's imperative that you find weapons with defense rating, unless you are somehow 25 points above uncrittable, aka 565.

I have some thoughts on two talents that could make dual-wield tanking more interesting, but I'm not a numbers person so I don't have the numbers balanced. If you're interested and can crunch numbers, send me a message.

I have a DK who is an avoidance junkie and is at 597 def

Edgewalker
06-25-2009, 05:19 PM
I have a DK who is an avoidance junkie and is at 597 def

I can't think of a way to not have terrible other stats and make that happen :(

Satorri
06-26-2009, 09:20 AM
I could do it very easily with two defense trinkets, but defense is not the most effective way to get avoidance as a DK. Not bad, but no where near what I could do with 2 dodge/parry trinkets (then pop them both and laugh at 60% dodge+parry).

Splug
06-26-2009, 01:31 PM
A few quick notes after tinkering on the PTR last night when my character finally copied over (evidently, the premades which copy quickly don't... have... onehanders. You have four two handed weapons though!).

* Both swings from Frost Strike crit on Killing Machine.
* XT-002 sigil was nerfed slightly; 114 damage on frost strike, but it applies to both hands, and seemingly after the offhand penalty. Thus it's a minor nerf but overall similar.
* The additional offhand strikes don't seem to be avoidable. I had a low sample size and limited mods, so I'm uncertain on that. It seems if the mainhand is avoided, the offhand does not fire at all, and if the mainhand connects, the offhand does as well.
* Death Strike only returns health once, even if both hands hit (there go my plans for the strangest overpowered tanking spec ever...)
* DW frost strike damage seems comparable on the PTR to its current 2h state on live, obliterate damage seems slightly higher.

Unrelated side note: the Scent of Blood change is a tooltip correction, the internal cooldown was already removed on live. The tooltip is still wrong on the PTR; it returns 10 RP per swing, but the tooltip lists it as 5.

-Splug

Feanorr
06-26-2009, 06:32 PM
I would be more apt to give frost tanking a try if cooldowns weren't such a vital part of end game tanking. Guile of Gorefiend / Unbreakable Armor just aren't up to the same level as a 30 second Vamp Blood / WoTN, especially when you factor in the new set bonuses and the increased CD on IBF to begin with.
Still waiting on something more before I am impressed.

I agree than UA is still pretty weak; when you think about it, it's like having a shield block for 20 sec every 2min. With the improvement in shield block value, it will probably be even weaker than that.

But, the blood tree, Vampiric blood included, work great with a high HP pool; I am not sure how it will work with our new HP (i.e. with less HP)

Satorri
06-27-2009, 06:53 AM
Thanks for the scouting Splug, that'll be useful info.

One small thing though, is it registering each strike as two separate ones, or a single strike. I assumed that it would be a combined strike just taking in the damage as a composite of the two weapons (off-hand damage penalty applied), and thus only a single strike to be avoided/miss, proc effects, etc.

Splug
06-27-2009, 08:57 AM
Thanks for the scouting Splug, that'll be useful info.

One small thing though, is it registering each strike as two separate ones, or a single strike. I assumed that it would be a combined strike just taking in the damage as a composite of the two weapons (off-hand damage penalty applied), and thus only a single strike to be avoided/miss, proc effects, etc.The combat log is listing two seperate attacks if the first one connects, and a single miss (dodge, parry, etc) if the mainhand doesn't. Again, I had a small sample size with limited mod availability (and have been backlogged at the office enough to be here on a Saturday morning) so I haven't followed up on it yet. I'd assume the mechanics are similar to Mutilate, but due to the specific talent implementation it may not be. Is anyone familiar with assassination rogues?

-Splug

Satorri
06-28-2009, 06:18 AM
I am actually, my rogue has been Assassination for a couple years, ha ha.

I figured Mutilate would be a rough way to code it. Muti for a rogue is a single move, and like you described, it's an all hit or not, but it is two separate portions of the hit.

ToT is a talent that changes multiple strikes. I assumed it would simply change the code of the strikes listed to composite its damage from main hand and offhand weapon damages, since like you mentioned, two hits on something like Death Strike would add a bit of complexity for programming to make sure the healing only triggers on one of the two hits. But, if that's how Blizz did it, it's good to know. That does mean that you won't see hit scales like a 2h DK, you'd just have twice as many and smaller.

I keep forgetting to set up my Frost off-spec to play with it, I've been having fun with premades of my classes that are still leveling in the 40s, of flash forwards on my under-geared alts to see what I can do with them in t8.5.

Jaeden
06-29-2009, 01:12 PM
My apologies if I actually missed it while reviewing the thread, but do we have any current accepted dw builds, dps and tank? EJ is being mum on it until they crunch their numbers.

Feanorr
06-30-2009, 05:36 AM
Does the assumption that slow 1H for main hand will still be true in 3.2?

I dont see why it wouldnt but then, is there any slow tanking weapon? I could find some slow 1H DPS weapon (most of them from pvp though) but none for tanking; did I miss some?

Anyway from your maths, Satorri, even with 2 fast weapon, the dps will be better than with a 2H, so it's not a big deal.

About DPS, what will be the best with 3.2: Slow/fast or Slow/slow? From what I could read, ToT use offhand weapon damage so a slow weapon will be better for our strikes, if I am not mistaken. But will it overvalue the advantage of a fast one?

That a lot of questions :p Thanks for any answers ^^

Satorri
06-30-2009, 05:55 AM
It *should* be slow/slow will be the new preference, since it, like mutilate, is an instant weapon damage strike on the off-hand as well.

For tanking that means we're doubly sad with no slow tanking weapons, but like you said Fean, even a fast/fast combo will do more damage than a 2h on those strikes (*with equivalent level weapons).

Edgewalker
07-01-2009, 06:59 AM
One handed PvP weapons are about as well itemized as you will get for tanking 1 handers.

Satorri
07-01-2009, 09:55 AM
Mmmm, I'm sure you're right, for getting the best threat/survival balance. I hadn't thought about that. Pity it takes such an additional investment in getting an arena rating high enough to buy the ones that will be most competitive with Ulduar gear.

And, to boot, in 3.2 you won't be able to do it with just a 2s team.

Edgewalker
07-02-2009, 11:03 AM
Even the 1850 rating ones are pretty much top of the line for tanking. Unless you can hit defense cap without resilience, you are forced to use fast weapons like Sorthalis or Shiver. I only need one defense weapon with the Resi Shoulder enchant, so I will probably be using a PvE DPS slow MH and Sorthalis OH.

Pwnanapuddin
07-02-2009, 11:22 AM
I know for death knights, our runeforge abilities for 2% parry are seen as the main and only option for DWing. However recently I have been testing using Blade Ward on at least one of the weapons, and I am thinking about using it on both.

While only using it on my primary weapon I can keep it up about 30-40% of the time, granting the same bonus, as well as a chance to stack it at least 2x from what I have noticed. Each stack adds 200 parry rating, and the resulting damage from the ward on our next parry also stacks from 600-800 up to 1200-1600.

As of now, I am still somewhat under geared for how I would like this to look, DWing Titanguards instead of Titan(P)/Legacy of Thunder(OH). I do have the LoT already tho, just need the higher Def gear to make it happen.


Feel free to tear me apart if this is a dumb idea.

Satorri
07-03-2009, 05:45 AM
Well, the question is (assuming you don't care about the price) whether Blade Ward offers a distinctly better buff. It's easiest to first look at the parry % it provides.

2% parry per weapon, non-diminishable, is as static as it gets.

200 parry rating, diminishable, is worth what, about 2.4% parry after returns? If it is up as you say, we'll average 35%, so it's functionally worth 0.84% on average, though as opposed to the 2% it is up for *slightly higher* amounts in short bursts.

The damage after parries may be a slight threat buff, but I think even with 100% uptime it wouldn't compare to Swordshattering for survival value. Just by the numbers. And the higher sustained parry amount may easily make up the threat from parry-counters with additional margins on RS damage (we have a thread going about that! Parry + DK = Win (http://www.tankspot.com/forums/f14/51913-death-knight-parry-win.html)).

Chamenas
07-03-2009, 06:12 AM
To go along with what Satorri is saying... I'm pretty sure that the runes from runeforges are supposed to be the best "enchants" out there for DeathKnights, whatever the role. If an enchant out there is beating them, I'm sure it is not intentional and once discovered it could easily be nerfed.

Pwnanapuddin
07-03-2009, 06:51 AM
Thank you for the good thoughts and breakdowns. One of the major reasons I am wanting to test this is due to the fact that one sword with this buff did perform so well. And since I am unwilling to throw this into the frey in ulduar, most of my testing has been against the "Highlords Nemesis" in DK Home. Therefore, I am never given a chance to parry an attack to wipe the buff while it is up at 1 or 2 stacks (I think 2 is the max).

Luckily cost to me to cover this is great, but my GL has the enchant and is giving me a substantial discount. So I will probably test this on a second weapon to see how it goes. Once I have my DEF pumped a little more to my liking, I will start using my sigil of deflection in conjunction. I really was not planning to unload this build until the point that the new DW talent comes out. I will admit I am a to the core, through and through frosty DK, and the new skill just gives me a reason to give DW a shot.

NexsusTheFrostDeathKnight
07-12-2009, 05:03 AM
A bump.

Unfortunately thanks to a B.net account bug and WoW's forums full of elitists I haven't been able to find any new info on DW tanking in 3.2.

Has anyone been able to test it further if the new talent is able to compensate even if it doesn't affect rune strike? My reasoning is that with some careful talent placing there will be enough runic power for frost strike to compensate and as mentioned virulence (I think it's called >.>) will further it.

Satorri
07-13-2009, 06:56 AM
It's far more than making up for the lost strike damage, your strikes will now do more damage than a 2h, unless they mess with the values somehow. Rune Strike isn't included in the effect for a reason. As a dual wielder you will actually be able to use RS's more often already, so that compensates somewhat for the slightly smaller size of each.

Make no mistake, there is no deficiency in dual wield damage or threat under this system, if anything it's going to be too good so as to threaten the 'hardcore' min/maxers who oppose the evils of dual wielding to tank while simultaneously HAVING to take the biggest damage/threat dealing setup they can. I can't wait to see where they fall on the outcome. Maybe it will result in Frost being unsuitable for tanking in their eyes, ha ha ha.

Ollin
07-13-2009, 10:47 AM
The combat log is listing two seperate attacks if the first one connects, and a single miss (dodge, parry, etc) if the mainhand doesn't. Again, I had a small sample size with limited mod availability (and have been backlogged at the office enough to be here on a Saturday morning) so I haven't followed up on it yet. I'd assume the mechanics are similar to Mutilate, but due to the specific talent implementation it may not be. Is anyone familiar with assassination rogues?

-Splug

It sounds like it works in exactly the same manner as Stormstrike.

When you hit the strike it checks to see if the main hand, and only the main hand hits. The off hand strike is a 100% main hand proc that fires if and only if the main hand attack deals damage.

Edit:

I do have to wonder though, even without access to ToT, how much could an Unholy DW tank abuse the internal CD removal on Sent of Blood and the change to Unholy Blight to pump his TPS over the course of a long single-target fight?

Personally I don't think it would pan out, but it would be an entertaining experiment.

Reading your description this sounds like a typical blizz shortcut, although I can't fault them for it. It works with SS well enough.

NexsusTheFrostDeathKnight
07-13-2009, 05:15 PM
I can't wait to see where they fall on the outcome. Maybe it will result in Frost being unsuitable for tanking in their eyes, ha ha ha.


This is my greatest fear :( the elitists claiming that even with the new system Dw tanking will fail, even now posts on the wow forums arebeing shot down by trolls all over the claims of rune strike alone, along with the pathetic 'parry gib' myth.

If it can do it's job though I'll definately be changing to DW tanking.

Pwnanapuddin
07-13-2009, 06:17 PM
I gave it a run through on 10 Uld last on public not test.... so even without ToT, I was DW tanking effectively on both Thorim and Freya as frost. I can only see it getting better.

I also found our that blade ward stacks to 4 max now...... I had parry oozing out my ass.

NexsusTheFrostDeathKnight
07-13-2009, 07:09 PM
I gave it a run through on 10 Uld last on public not test.... so even without ToT, I was DW tanking effectively on both Thorim and Freya as frost. I can only see it getting better.

I also found our that blade ward stacks to 4 max now...... I had parry oozing out my ass.

You must be in the greatest guild of all time for them to "allow" you to dw tank.

So 2 blade wards look like they can replace the runeforge enchants?

Janana
07-14-2009, 01:17 AM
You must be in the greatest guild of all time for them to "allow" you to dw tank.

So 2 blade wards look like they can replace the runeforge enchants?
I was thinking about 1 swordbreaking and 1 bladeward, putting the bladeward in the off hand as a side spec to see if i liked it.

lyd
07-14-2009, 05:40 AM
This is my greatest fear :( the elitists claiming that even with the new system Dw tanking will fail, even now posts on the wow forums arebeing shot down by trolls all over the claims of rune strike alone, along with the pathetic 'parry gib' myth.


the wow forums are probably the worst place to get information about this game. EJ is fine, tankspot is good, anywhere you can post numbers and have people give constructive feedback instead of simply say "TLDR."

the wow forums are good for hilarious bouts of trolling and that's about it, honestly. even legitimate questions are answered with "NUB LOL" and the like. really just a terrible place all around, but entertaining to peruse occasionally. sometimes i get sucked in by their jack@ssery, but i try my best to keep a safe distance.

Satorri
07-14-2009, 07:11 AM
The funny thing is that 2h Frost isn't getting hurt by any means, and if anything it's getting a little buff gaining 2 talent points back from BotN. But if dual wielding does more damage people will start saying it's the "only legitimate way to play frost." It's foolish but it's popular rhetoric.

Dual wielding somehow losing out on Rune Strike is just clamor from people who don't understand the mechanics that well. It may or may not do better, but it won't be some grievous loss, that depends on avoidance and weapon choice. Maybe another bout of napkin math is in order.

Emi
07-14-2009, 07:37 AM
From a personal PoV this doesnt change much if i choose to continue with my 2H when 3.2 hits. Things standing as they are my build would be : Talent Calculator - World of Warcraft (http://ptr.wowhead.com/?talent#j0EGZhxx0AbIof0buzAo0x) . BotN moving to 3pts now forces me to pick up Deathchill and 2/2 on MC which means i wouldnt actually "gain" anything from those "extra" 2 points.

Now if i go DW then Talent Calculator - World of Warcraft (http://ptr.wowhead.com/?talent#j0EbZhxe0AbI0c0fuzAo0x) is what i'd be going for. Again, apart from the DW talent pick ups/2H drops and no points in MC the tree is basically the same.

Satorri
07-14-2009, 08:13 AM
Heh, it doesn't *force* you to take those, but your choices for those 3 points does gain you something. Deathchill is a wonderful exemplar of what can be a good value or no value at all depending on how much and when you use it.

It could also make life easier on the folks who pick up Imp Icy Talons for their team.

Certainly none of this is a penalty, I like having more options that feel like nice options.

Emi
07-14-2009, 08:19 AM
Well, what i meant to say is that i see no real "value" in seeing a talent having its point cost reduced but that "forces" me to pick others below it that i normally wouldn't(like Deathchill) just to access the higher layer talents.

So in effect, from a personal PoV, im "gaining" "nothing" :D

Dosvi
07-14-2009, 08:34 AM
Dual wielding somehow losing out on Rune Strike is just clamor from people who don't understand the mechanics that well. It may or may not do better, but it won't be some grievous loss, that depends on avoidance and weapon choice. Maybe another bout of napkin math is in order.

I find this to be very interesting. You have no math to support anything you just said. You make the assumption that it is just "clamor from people who don't understand the mechanics." So, here is the supporting math. This is using two desired tanking weapons of the same iLevel.

-Rune Strike:
-Strike the target for 150% weapon damage plus [150 * AP * 10 / 10000]
-Based on 3050 AP

Broken Promise:
274 - 509 Damage
RS: 731.5 - 966.5

Inevitable defeat:
554 - 831 Damage
RS: 1011.5-1288.5

BP / ID (rune strike damage ratio)
.723 - .75
You are losing 25% off a single RS hit.
-------------------------------------------
To compensate, you would need (.75 * x = 1) 4 rune strikes from the one hander to 3 rune strikes from the two hander. That would give you the same RS damage. The assumptions that a one hander weapon will get more SoB procs, as well as more RS hits is correct. The math becomes very complex and it becomes much easier to use simulations and make the computer perform the operations.

With Inevitable Defeat (weapon 1) versus Broken Promise (weapon 2), using 3050Ap, 44% avoidance, 6% miss, 50000s simulation, rank 3 SoB for the following-
---------------------------------------------------------------
Runic power gain/expenditure from your rotation (the following 3 simulations):
0,20,20,10,10:0,20,20,10,15
===================================
Boss Swing Time: 1s

Procs: 22544
Weapon 1 RS hits: 12690
Weapon 2 RS hits: 14968
Weapon 1 RS dps: 420.6735
Weapon 2 RS dps: 361.4772
Weapon2 / weapon1 ratio: 0.8592820798077369
Weapon 1 rp starved misses: 0
Weapon 2 rp starved misses: 87
--------------------------------------
Boss Swing Timer: 2s

Procs: 11340
Weapon 1 RS hits: 9212
Weapon 2 RS hits: 10175
Weapon 1 RS dps: 305.3778
Weapon 2 RS dps: 245.72625
Weapon2 / weapon1 ratio: 0.8046631091061629
Weapon 1 rp starved misses: 0
Weapon 2 rp starved misses: 0
----------------------------------
Boss Swing Timer: 3s

Procs: 7423
Weapon 1 RS hits: 7036
Weapon 2 RS hits: 7423
Weapon 1 RS dps: 233.2434
Weapon 2 RS dps: 179.26545
Weapon2 / weapon1 ratio: 0.768576731431629
Weapon 1 rp starved misses: 0
Weapon 2 rp starved misses: 0
================================
RP gains / expenditures (this one includes a single frost strike at the end of the rotation).
0,20,20,10,10,-33:0,20,20,10,15,-32
================================
Boss Swing (1s)

Procs: 22736
Weapon 1 RS hits: 10068
Weapon 2 RS hits: 11773
Weapon 1 RS dps: 333.7542
Weapon 2 RS dps: 284.31795
Weapon2 / weapon1 ratio: 0.8518782684981941
Weapon 1 rp starved misses: 3082
Weapon 2 rp starved misses: 4193
---------------------------------------
Boss Swing (2s)
Procs: 11297
Weapon 1 RS hits: 8651
Weapon 2 RS hits: 9548
Weapon 1 RS dps: 286.78065
Weapon 2 RS dps: 230.5842
Weapon2 / weapon1 ratio: 0.8040437874731088
Weapon 1 rp starved misses: 782
Weapon 2 rp starved misses: 1128
------------------------------------
Boss Swing (3s)

Procs: 7491
Weapon 1 RS hits: 6975
Weapon 2 RS hits: 7375
Weapon 1 RS dps: 231.22125
Weapon 2 RS dps: 178.10625
Weapon2 / weapon1 ratio: 0.7702849543456755
Weapon 1 rp starved misses: 189
Weapon 2 rp starved misses: 330
=================================

On a ~2s boss swing timer, you lose 20% of your rune strike threat. If RS is 50% of your threat, you lose 10% of your threat. The amount of RS damage per fight varies from about 30% to 50%. Best case (duel-wield) scenario at 30%, 39% of your threat is from RS. This means you lose 7.8% of your threat.

Even at 60% dodge/parry, you lose 17% of RS damage with a 2s boss swing timer. It is significant. Believe it or not, some of us actually do math and don't just clamor.

Simulation source code (Java):
DWTank (http://docs.google.com/View?id=dchgzb93_0fn33btd2)

Satorri
07-14-2009, 10:15 AM
Spare me the insulting tone, we can have a productive discussion without it. I have done math on it some time ago and didn't recall there being sweeping losses. Much of the clamor of which there is plenty, is from people who do not do the math as you do, they simply hear other people say, "we'll lose some threat" and broadcast it to the world, "ZOMG WE'RE ALL GONNA DIE HORRIBLY!!"

Thank you for the simulation, that is helpful. I'm unsure of a couple of elements of your simulation: Did you handle main-hand only using RS procs for dual wielding? (don't think we ever established if that was the case or if off-hands could apply it as well) As for % of your threat coming from RS that seems terribly high, my personal tanking with Frost has put it in the area of 15-30% of my total threat at most.

In the end, the real issue at hand is how the loss in RS balances out against the gains in many other things and the lack of change in a couple items. This is always the consideration in dual wielding, though now the balance is shifting.

RS will slide a bit, white swings and -strikes will jump up a bit, and non-weapon based attacks will remain the same. So here's where simulations don't help. Players will use their own methods and sequences with spells, as mitigated by latency, attention, and other human factors, and practical %'s will vary.

I'll steal some values from a previous set of calculations I did on the buffs to strike values (about 20% gain in FS, BS, OB, PS dmg rounding down to be conservative), along with a recent log for 2h frost tanking I did, and borrow your values for RS (25% loss in damage, 10% increase in RS's used, awful butchery of your values in reduction, forgive me). Here's how the breakdowns compare:

*disclaimor, this is probably pretty horrible for an actual breakdown, but the proportions shouldn't be horribly skewed for the sake of illustrating the trade-offs*

2h Live % breakdown on threat overall (Ulduar25, generalized for ease):
22% Howling Blast
19% Frost Strike
17% Rune Strike
15% Obliterate
12% swings
9% Frost Fever
4% Blood Strike
3% Blood Boil

So, if we lose 25% damage per RS, but get 10% more of them, that's a 17.5% loss in RS threat. HB, FF, and BB will remain the same, all being spell damage and not weapon-based. FS, OB, and BB will get a 20% damage/threat buff from ToT, and white swings should increase by roughly the same proportion.

The new breakdown would be:
21% Frost Strike
21% Howling Blast
17% Obliterate
13% Rune Strike
13% swings
8% Frost Fever
4% Blood Strike
3% Blood Boil

With a net gain of roughly 6.8% threat overall.

The loss is there, yes, but the severity isn't so grand. If your RS is in fact 50% of your total threat, then you will be losing almost 9% of your total threat from loss of RS threat. If that other 50% is a breakdown of say 15% HB, 30% OB+FS+swings+BS, and 5% FF, BB, etc, then you'd also be gaining 6% from your other strikes and you'd be netting a loss.

Out of curiosity, let's find the amount you'd have to lose on RS to cancel out the gains from ToT on strikes, taking 18% loss of RS threat and 20% gain of strike threat:



RS % of total threat FS+OB+melee+BS % of total threat Net Change
50% 30% -3%
45% 33% -1.5%
40% 36% 0%
35% 39% +1.5%
30% 42% +3%
25% 45% +4.5%
20% 48% +6%
15% 51% +7.5%


So drawing from all of our many assumptions, your simulations, and my relative breakdown on potential damage proportions, the break-even point would be with Rune Strike around 40% of your total damage.


Personally, I've only managed to get that high packing hardcore avoidance gear on Patchwerk Hateful tanking, but I can't anticipate how other DK's play. Purely in my thinking, instinctual, not observation-based, if you are getting that much threat from Rune Strike you are not doing what you could be with all of your active move rotations/priorities/sequences.

And in general, losing RS threat will not lose you threat overall, it will simply bring you even with a 2h Frostie, or slightly ahead mitigating your gains from ToT (as if designed that way >.> hmm).

Dosvi
07-14-2009, 11:24 AM
I am sorry if that seemed derogatory. I am just tired of seeing posts like that with no math. I should refrain from letting the negativity slip in. I really do want to know if it will be viable.



Did you handle main-hand only using RS procs for dual wielding? (don't think we ever established if that was the case or if off-hands could apply it as well)

As far as the tests I have seen, RS doesn't do anything to the offhand. So, I didn't factor it in. IF it does affect offhand damage, let me know. If it does cause an offhand proc it surpasses two handed threat by leaps and bounds.

I did remember that I needed to factor in the SoB use by the offhand. If I use duel 2.6 weapons (duel borken promise), it does increase the RS damage by about 2%. So, that should be factored in. I'll add a variable swing offhand later and see about making an actual rotation per damage simulator. That way I don't have so many hypotheticals floating out there.



As for % of your threat coming from RS that seems terribly high, my personal tanking with Frost has put it in the area of 15-30% of my total threat at most.

Personally, it varies on my spec and the encounter. My numbers may be high since I generally tank as blood. Which allows less rp dump time and involves less rp dumping then frost. On high threat encounters (like hodir and general, I have hit ~45% of my damage as RS). I also need to factor in healers that restore rp. Again it depends on your rotation, if you weave frost strikes in your rotation (this stops overflow), versus an RP dump, you will see an increase in rune strike damage.



Players will use their own methods and sequences with spells, as mitigated by latency, attention, and other human factors, and practical %'s will vary.

This is very true. Unfortunately, slower weapons are easier to use. There are fewer procs to worry about and you have a larger time frame to get off RS.

The rest of your post is thorough and I should push that way with my simulation. When you factored the damage increase from everything, did you include KM procs (or are they not affected by DW; wowwiki says KM is affected, but they could be wrong)? I'll see what I can do to make this a full threat simulation so I can figure out how close to that margin we really are. (Because we are very close to it, but certain things do affect it by a large margin. If you want an example, go tank multiple mobs and watch how often you RS. It skyrockets and would push DW up).

Satorri
07-14-2009, 11:43 AM
My stuff didn't account for KM procs. That's another messy problem. Logically I'd expect the effect to remain the same, since it will buff FS and HB (dual wielding shouldn't change the rate/sequence that you use the, though it will make FS bigger). I'm not 100% sure if dual wielding will improve on the turnover rate or not. In theory 2h's may not be getting as many procs as are possible per minute, I'm not sure how that works out.

And off-hand RS procs are something a few of us have been kicking around for months, but none of us can seem to figure out if RS procs with off-hand hits.

Ooooooo lightning bolt to the noggin! Later this evening, if I get a chance, I'll go out with a super crappy white weapon in my main hand and a nice high level off-hand and attack some stuff that I'll dodge/parry a lot. We'll see if I get any bigger RS's, that should give some indication.

NexsusTheFrostDeathKnight
07-14-2009, 05:46 PM
Unfortunately I have no highly detailed math, and in 'somewhat' new to the tanking genre, but here's a proposal (at least for heroics only, I rarely raid) does RS even matter? I've never had rune strike macro'd to any of my attack and alll enemies have stuck to me like glue with HB going off in their faces, although I must say one test realm patch I did bother to tie RS to all the abilitees and it was insanely fun throwing large crits in peoples faces along with HB, FS and OB.

That aside basically what I'm proposing is, does RS even matter 100%? after all you could spend the runic power on frost strike and hit with both weapons, also what must be some big obliterate numbers.

Feanorr
07-14-2009, 09:17 PM
Personally, it varies on my spec and the encounter. My numbers may be high since I generally tank as blood. Which allows less rp dump time and involves less rp dumping then frost. On high threat encounters (like hodir and general, I have hit ~45% of my damage as RS). I also need to factor in healers that restore rp. Again it depends on your rotation, if you weave frost strikes in your rotation (this stops overflow), versus an RP dump, you will see an increase in rune strike damage.

RS's % is definitely not that high as a frost tank; I didnt tank as frost since some time so I cant find some stats, but as far as I remember, RS was behind both FS and OB in term of threat.

Dosvi
07-14-2009, 11:28 PM
Unfortunately I have no highly detailed math, and in 'somewhat' new to the tanking genre, but here's a proposal (at least for heroics only, I rarely raid) does RS even matter?

In general, no it doesn't matter in heroics or even naxx. Threat is not an issue for most tanks unless your dps are extremely geared (full uld25 gear) or it is a gimmick fight. (Fights like Maly, Hodir, General. These fights push up dps and cause them to ride your threat). You could by all means tank lower level content while duel wielding. However, until it is proven to be superior, pugs will be less likely to take you.



RS's % is definitely not that high as a frost tank; I didnt tank as frost since some time so I cant find some stats, but as far as I remember, RS was behind both FS and OB in term of threat.

Yes, I know that RS will be lower for frost then it will be for blood. However, if you have the correct rotation, you can easily place it equal to FS and OB.

The above is more for end game tanking viability. That much of a threat loss is a non issue until you get into hardmodes in end game. In those scenarios, every bit of threat you get is significant. I'm trying to find out exactly what the change to threat will be.

Emi
07-15-2009, 01:41 AM
Frost Tank here (no HB Glyph). If its any help this is a log from last time i was at Vezax : World of Logs - Real Time Raid Analysis (http://www.worldoflogs.com/reports/G3td4GlJC43N9xj4/details/6/?s=3436&e=3883) .

Due to 3/3 on SoB i find myself having enough RP to use whenever KM procs or just to dump thus you see FS figuring prominently. In any case the % use of FS, Ob and RS is within 3% as you can see.

Satorri
07-15-2009, 06:24 AM
As your gear improves, specifically your avoidance, RS becomes a major threat tool, and I do mean major.

In Naxx25, that's where my threat started sky rocketing, and it does more threat now (and less damage, /sigh miss my 10k+ crits). You can get away without it, but why bother? It is the most efficient threat you can get from RP regardless of spec or gear.

lyd
07-16-2009, 02:06 PM
It's pretty easy to macro all your attacks to use RS if it's available. I haven't done it personally because i'm quick at the keyboard and and my knowledge of macros is lacking, but it is wise to do so. I will get around to it eventually.

Also, it's a pretty important ability to use should you decide to DW. the math that tears down the parry-haste fallacy starts to falls apart if you're not actively replacing MH swings with RS at every opportunity.

Satorri
07-16-2009, 02:40 PM
The macro method is quite simple, though I've seen some funny unnecessary additions.

#showtooltip spell name
/cast spell name
/cast Rune Strike

That's all that's required and the button will show whatever move you tacked it onto on mouseover, and when you hit it it will activate RS for your next swing if it's available.

Dosvi
07-16-2009, 09:41 PM
The macro method is quite simple, though I've seen some funny unnecessary additions.

#showtooltip spell name
/cast spell name
/cast !Rune Strike

That's all that's required and the button will show whatever move you tacked it onto on mouseover, and when you hit it it will activate RS for your next swing if it's available.

Fixed that for you.

/cast !Rune Strike

Using the exclamation point won't toggle rune strike on and off; /cast Rune Strike, will cause it to toggle. The exclamation point stops this.

Splug
07-17-2009, 11:29 AM
Back before the 3.1 removal of BoS RP generaton, I was trying to crunch out the autoattack implications of DW vs 2h using rune strikes. I ended up throwing together a mini-spreadsheet based off this and concluded that dual wielding with infinite RP does outdamage 2h damage, particularly for fast-attacking bosses such as Patchwerk. I can't find the spreadsheet I put together, but I did find the post I put together with the probability formulas it was based off. Again, this was done during the infinite RP era of 3.0.8... so there's certainly going to be less RP efficiency, and SoB is not accounted for either. But I did find higher throughput when I threw it at the spreadsheet. EDIT: A few talents have changed evidently as well... Tundra Stalker is still being referenced as a 10% increase there.

---------------------

If a player's mainhand attack occurs at mh_speed and a boss attacks at boss_speed, the number of attacks taken per attack made is:

boss_speed / mh_speed

If combined parry/dodge-based avoidance is defined as avoid% (note - NOT including miss chance), then the probability of an arbitrary autoattack lighting up as a rune strike is:

Rune%_per_swing = 1 - (1 - avoid%)^(boss_speed/mh_speed)

The damage-per-swing* generated by autoattack and rune strike are (boss_glance% = .25, glance_penalty = .3):

mh_swing_damage = (mh_base + AP/14*mh_speed)*(1 + swing_crit% - boss_miss% - boss_dodge% - boss_parry% - boss_glance%*glance_penalty - boss_block%) + (mh_base + AP/14*mh_speed - boss_block_value)*boss_block%

oh_swing_damage = (oh_base + AP/14*oh_speed)*oh_penalty*(1 + swing_crit% - boss_miss% - boss_dodge% - boss_parry% - boss_glance%*glance_penalty - boss_block%) + [(oh_base + AP/14*oh_speed)*oh_penalty -boss_block_value]*boss_block%

rune_damage = [(mh_base + AP/14*mh_speed)*1.5 + .15*AP]*(1 + rs_crit% - boss_miss%)*rs_mod
*- All attacks are passing through the same value of target armor. Thus, for these calculations armor value has been ignored. %-based modifiers are also being ignored, as they should also be globally effective. The only exception would be talents such as tundra stalker and rage of rivendare, which can be accounted for in rs_mod. For threat values, multiply rs_mod by 1.5.

Combining that all into one equation and assuming runic power cost is a non-factor (only accurate in some cases), the total autoattack-plus-runestrike damage yielded by a tank who is firing rune strike as fast as it comes off cooldown should be as follows:

Rune_plus_swing_dps = [1 - (1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*1.5 + .15*AP]*(1 + rs_crit% - boss_miss%)*rs_mod/mh_speed + [(1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*(1 + swing_crit% - boss_miss% - boss_dodge% - boss_parry% - boss_glance%*glance_penalty - boss_block%) + (mh_base + AP/14*mh_speed - boss_block_value)*boss_block%]/mh_speed + {(oh_base + AP/14*oh_speed)*oh_penalty*(1 + swing_crit% - boss_miss% - boss_dodge% - boss_parry% - boss_glance%*glance_penalty - boss_block%) + [(oh_base + AP/14*oh_speed)*oh_penalty - boss_block_value]*boss_block%}/oh_speed

Breaking that down into a few special cases, to use hit/expertise stats instead of miss rates:

-- Known constants:
rs_miss% = .08 - melee_hit% // Don't forget possible +.03 from talents
(2h) swing_miss% = .08 - melee_hit%
(dw) swing_miss% = .27 - melee_hit% // Don't forget possible +.03 from talents
boss_dodge% = .065 - exp% // Until exp% > .065%, then it's zero...
boss_parry% = .14 - exp%
boss_block% = .065
boss_glance%*glance_penalty = .075

For dual wielding at expertise <= 26:
Rune_plus_swing_dps = [1 - (1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*1.5 + .15*AP]*(.92 + rs_crit% + melee_hit%)*rs_mod/mh_speed + [(1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*(.385 + swing_crit% + melee_hit% + 2*exp%) + .065*(mh_base + AP/14*mh_speed - boss_block_value)]/mh_speed + {(oh_base + AP/14*oh_speed)*oh_penalty*(.385 + swing_crit% + melee_hit% + 2*exp%) + .065*[(oh_base + AP/14*oh_speed)*oh_penalty - boss_block_value]}/oh_speed

For dual wielding at expertise >= 26:
Rune_plus_swing_dps = [1 - (1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*1.5 + .15*AP]*(.92 + rs_crit% + melee_hit%)*rs_mod/mh_speed + [(1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*(.45 + swing_crit% + melee_hit% + exp%) + .065*(mh_base + AP/14*mh_speed - boss_block_value)]/mh_speed + {(oh_base + AP/14*oh_speed)*oh_penalty*(.45 + swing_crit% + melee_hit% + exp%) + .065*[(oh_base + AP/14*oh_speed)*oh_penalty - boss_block_value]}/oh_speed

For two-handing at expertise <= 26:
Rune_plus_swing_dps = [1 - (1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*1.5 + .15*AP]*(.92 + rs_crit% + melee_hit%)*rs_mod/mh_speed + [(1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*(.575 + swing_crit% + melee_hit% + 2*exp%) + .065*(mh_base + AP/14*mh_speed - boss_block_value)]/mh_speed

For two-handing at expertise >= 26:
Rune_plus_swing_dps = [1 - (1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*1.5 + .15*AP]*(.92 + rs_crit% + melee_hit%)*rs_mod/mh_speed + [(1 - avoid%)^(boss_speed/mh_speed)]*[(mh_base + AP/14*mh_speed)*(.64 + swing_crit% + melee_hit% + exp%) + .065*(mh_base + AP/14*mh_speed - boss_block_value)]/mh_speed

Now, some of those values are determined by your talents, and what you're trying to calculate:
rs_mod = 1 (for damage dealt, without rage of rivendare or tundra stalker)
rs_mod = 1.1 (for damage dealt, with rage of rivendare or tundra stalker)
rs_mod = 1.5 (for pre-frost presence threat, without rage of rivendare or tundra stalker)
rs_mod = 1.65 (for pre-frost presence threat, with rage of rivendare or tundra stalker)
oh_penalty = .5 (without nerves of cold steel)
oh_penalty = .575 (with nerves of cold steel)

-Splug

Satorri
07-18-2009, 04:14 AM
Actually Dosvi, that's totally unnecessary, one of the confusions pieces I'm talking about. When the macro as I typed it is used (which I spam), it doesn't toggle it off at all.

That became a popular notion some time back, but it is a fallacy.

Dosvi
07-19-2009, 10:40 AM
Actually Dosvi, that's totally unnecessary, one of the confusions pieces I'm talking about. When the macro as I typed it is used (which I spam), it doesn't toggle it off at all.

That became a popular notion some time back, but it is a fallacy.

I stand corrected. I should have tested it myself. The confusion stems from the 2.3.2 patch notes. "/cast will toggle spells again unless the name is prefixed with an exclamation mark, e.g. /castsequence Steady Shot, !Auto Shot." But running my own tests, rune strike doesn't toggle.

Satorri
07-20-2009, 06:06 AM
=) very small thing that people won't notice at all in or out, and a single character is a small thing in a macro. But misinformation is nice to squelch when you can.

NexsusTheFrostDeathKnight
07-20-2009, 06:52 AM
lets not derail this with macros >.>.
so is the optimal slow/slow? or will I be able to start off with red swords of courage?

Satorri
07-20-2009, 07:32 AM
Optimal and available are going to be a tricky balance.

By the current mechanics, slow/slow is optimal, fast/fast still surpasses a 2h for total strike damage, and probably the choice you will run into is slow/slow both as dps weapons, or fast/fast tank weapons, if you have the luxury to pick (many will not for a long time), at the highest level of gear offering.

It seems a threat vs survival value really, and the dps of the weapon will usually outweigh the speed. Dual wielding Sorthalis and/or Shiver would be a great balance of threat and survival methinkst.

junkilo
07-20-2009, 11:34 AM
DW tanking is already very possible, it just pulls lower TPS at a higher cost with a more strict gear dependency.

No, it isnt.

Look at Satorri's post (#24). Double the parrys than when using a single weapon. Tanks get one shot as it is with just one parry landing at the same time as a melee swing and magic damange. I'm not willing to play chicken with the RNG like that.

If DW was more dps then 2h tanking and you could get near the expertise hard cap then sure.

lyd
07-20-2009, 11:38 AM
slow/slow for tanking. as far as i can tell, dual BP's is still the optimal tanking configuration for DW. off-hand fast weapons won't kill you, but a mh fast weapon considerably reduces the effectiveness of your weapon damage-based abilities. Right now I off-hand a red sword of courage for no other reason than i have yet to get an opportunity to replace it >.<



No, it isnt.

Look at Satorri's post (#24). Double the parrys than when using a single weapon. Tanks get one shot as it is with just one parry landing at the same time as a melee swing and magic damange. I'm not willing to play chicken with the RNG like that.

If DW was more dps then 2h tanking and you could get near the expertise hard cap then sure. When the math behind DW tanking was first being fleshed out on these forums, I believe it was shown that warriors with a fast tanking weapon generate more parry-haste opportunities off a lv 83 boss than a frost dk DW'ing slow/slow. This is partly because 3 of the attacks in their rotation are either spells or not parryable (IT, HB, FS). I believe Satorri was actually the one who did the theorycraft. It follows, therefore, that if DW'ing is not a valid tanking style based on the number of parry-haste opportunities they generate, then the entire warrior class, ipso facto, is not valid for tanking either, given that they generate even more. you also can't use block as an excuse, since using DK mitigations on a cycle generally serves the same purpose. Better, actually, given the current state of block, and the proposed nerfs to DK's.

since we all know warriors who are successfully tanking in 25-man raids right now, I think we can, once more, put this to rest. this is yet another example of how badly the parry-haste mechanic is exagerrated in terms of understanding its effect on DK tank survivability.

Satorri
07-20-2009, 11:39 AM
Junkilo, you need to read the other 200 or so posts on this. Dual wielding is quite doable, it is not that big an added vulnerability, and in fact at the expertise soft cap (which most tanks aim for anyway) you won't notice much difference in incoming damage or pacing.

You can get to the expertise hard cap, but it's neither necessary, nor really worth it.

It's not that bad Lyd, if you have reasonably high level items, we run out of slow tanking weapons with Naxx, so it may be worth using fast weapons even for threat if you want to get nice tanking values out of the weapons as well.

junkilo
07-20-2009, 03:00 PM
Junkilo, you need to read the other 200 or so posts on this. Dual wielding is quite doable, it is not that big an added vulnerability, and in fact at the expertise soft cap (which most tanks aim for anyway) you won't notice much difference in incoming damage or pacing..

Point taken. However you assume that the DK tank is DW'ing broken promise and the war tank has a fast 1.6 speed one-hander. You need to go apples for apples for these comparisons. I don't know of a single parse that item level for item level can match DK 2h dps either. If there is please point me to it. Perhaps DK dps is different for tanking and it makes DW dps better?

Feanorr
07-20-2009, 07:57 PM
Point taken. However you assume that the DK tank is DW'ing broken promise and the war tank has a fast 1.6 speed one-hander. You need to go apples for apples for these comparisons. I don't know of a single parse that item level for item level can match DK 2h dps either. If there is please point me to it. Perhaps DK dps is different for tanking and it makes DW dps better?

Read the post before answering maybe?

First, you may not have noticed but we are talking about DWing with the next patch (and the change to frost DW). So parse will be hard to find (not many people raid on the PTR, even less post there parses).
Second, Satorri showed some maths proving that even with 2 fast weapon (worst scenario since 2 slow should be the best, but the more possible scenario since slow tanking weapon doesnt exist) the DPS/threat of DW was superior than with a 2H (with the 3.2 change, of course).

VerticalEvent
07-21-2009, 04:49 AM
Point taken. However you assume that the DK tank is DW'ing broken promise and the war tank has a fast 1.6 speed one-hander. You need to go apples for apples for these comparisons. I don't know of a single parse that item level for item level can match DK 2h dps either. If there is please point me to it. Perhaps DK dps is different for tanking and it makes DW dps better?

I think I can find a few weapons that can out DPS a 2-H weapon

Item Comparison Tool - World of Warcraft (http://www.wowhead.com/?compare=40406:39292;40402:40345:40531;45165:39292 ;45892:45110:40531)

A comparison of Weapons from Naxx25 and Ulduar25.

For Naxx25:
2-H Weapon:
NAME: Inevitable Defeat
DPS: 203.7

1-H Weapon:
NAME: Last Laugh
DPS: 171.6

NAME: Broken Promise
DPS: 156.6

TOTAL DPS: 171.6 + 156.6 *.65 (talented) = 273.39 dps

In Naxx, the white damage generated is 33% greater with 2 1-H Weapons, then a 2-H weapon for tanking weapons. We can even increase this by DWing 2 Last Laughs, instead of mixing up the weapons.

And, if I remember correctly, strike damage is all standardized to a specific weapon speed (3.6s, I do believe), so Frost Strike and Obliterate's damage is not affected by the swing speed of the weapon, but the overall DPS.

Satorri
07-21-2009, 06:06 AM
It all gets a bit sticky in the live patch, a hard balance.

This thread is dedicated to the rise of dual wielding in 3.2. There is no good reason then that a dual wield frost tank can't tear it up as good or better than any other setup.

I am not a dual wield tank myself (I've played with it, it works, it's not my style). I come out as an advocate in the community because of the raging nay-sayers who come to the forums to blast the very idea and anyone who entertains it.

It is viable.
It can be done with the same gear set as any other DK tank.
It can be improved on by adjusting your talents and moves to match its strengths/weaknesses.
In 3.2 Frost Dual Wielding gets a MASSIVE prop up.

And as a side note: theorycrafting, simulations, and napkin math do not prove anything on the practical application of dual wield tanking. You can use any of them to prove how awful or how functional it can be, but that has little bearing since the only true test is doing it. If you want, we have an assortment of past and present dual wield tanks who can tell you they're doing just fine tanking current content. What's worst is the people who don't back their arguments with even made-up out of game math, they just have heard it can't be done without zomgparrygibs and they feel the need to come here and spout their superior understanding.

Dosvi
07-21-2009, 08:41 AM
I think I can find a few weapons that can out DPS a 2-H weapon.... You have some flaws in the data here.

1. You are comparing a 226 and a 213 weapon (for DW) to a 213 weapon. Which really isn't fair. At the end, you suggest comparing DW 226x2 to the 213 weapon.
2. You don't factor in miss. You are going to miss 19% more with DW.
3. The attacks are normalized. But, that normalization depends on the weapon being a two-hander or a one hander.
Normalization - WoWWiki - Your guide to the World of Warcraft (http://www.wowwiki.com/Normalization)

Yes, math shows the damage to be higher with DW then it is with a two-hander. It's not as drastic as your post is showing.

Satorri
07-21-2009, 08:45 AM
Nitpick:

*Miss 19% more on auto-attacks

Also, of course, the picture is a lot more complicated because you have to factor all the damage components, AND more importantly method of delivery. A dual wielder spec'd and playing like a 2h DK will not get the best results the style is capable of.

VerticalEvent
07-21-2009, 10:30 AM
You have some flaws in the data here.

1. You are comparing a 226 and a 213 weapon (for DW) to a 213 weapon. Which really isn't fair. At the end, you suggest comparing DW 226x2 to the 213 weapon.
2. You don't factor in miss. You are going to miss 19% more with DW.
3. The attacks are normalized. But, that normalization depends on the weapon being a two-hander or a one hander.
Normalization - WoWWiki - Your guide to the World of Warcraft (http://www.wowwiki.com/Normalization)

Yes, math shows the damage to be higher with DW then it is with a two-hander. It's not as drastic as your post is showing.

1. Well, I could take Betrayer of Humanity, the best 2-H weapon in Naxx25, which has a DPS of 222.9 and the two 1-H weapons still beat it by 51 dps, and one of them is iLevel 213 (Broken Promise). I was comparing them based on where they dropped to and less on iLevel. Two Last Laughs would give me a dps of 283.14. Also, remember that Betrayer of Humanity lacks the tanking stats that the Inevitable Defeat has.

2. I think you should re-phrase that to being you have an increase of 19% chance to miss. As well, talented for dual wielding, that increase of 19% goes to 16% (+3% chance to hit with single handed weapons, for both weapons, when talented). As well, that only counts for pure white damage, while the yellow strike damage is based purely on if the main handed weapon lands or not, from what I can gather from the PTR and the new talent. For those of us using a +Defense trinket, Dual Wielding 1-H weapons will free up a trinket slot (if we go with tanking weapons) and let use a +Hit or +Expertise trinket, to compensate.

3. Thank you for the link. So, in order for the standardized damage of a 2-H weapon to do more then a 1-H weapon, the 1-h Weapon needs to do about 27% more dps then the 2-H weapon.

There's also the idea of it being easier in getting 1-H tanking weapons over the 2-H weapons, where almost every melee dps is rolling on it with the tank, where as a 1-H tanking weapons quickly go to offspec after two or three runs.

junkilo
07-21-2009, 11:19 AM
Read the post before answering maybe?

First, you may not have noticed but we are talking about DWing with the next patch (and the change to frost DW). So parse will be hard to find (not many people raid on the PTR, even less post there parses).

gotcha.

VerticalEvent
07-21-2009, 04:34 PM
There's another thing that my post was missing - attack power.

Since attack power gets applied to both weapons, and I can easily have about 3000 ap (without buffs). Dual wielding, that's like me have 3000 ap x 1.65 = 4950 ap, or an increase of 353 dps when dual wielding, compared to 214 dps I'd have with a single 2-H weapon, or an increase of 139 dps.

Torn
07-21-2009, 07:25 PM
A message to all naysayers regarding DW tanking:
I've been DW tanking since I first stepped into Naxxramas in February. Meanwhile, I've successfully tanked everything up to and including General Vezax in Ulduar. All I can say is: Yes, the generated threat is lower while DW tanking. No, this doesn't mean, that you can't tank the bosses. Threat has been a non-issue throughout WotLK (anyone remember the Shattered Halls pulls in BC?) and DW tanking today simply means that you shouldn't go afk while tanking.

Still, I'm looking forward to patch 3.2.

NexsusTheFrostDeathKnight
07-22-2009, 01:19 AM
A message to all naysayers regarding DW tanking:
I've been DW tanking since I first stepped into Naxxramas in February. Meanwhile, I've successfully tanked everything up to and including General Vezax in Ulduar. All I can say is: Yes, the generated threat is lower while DW tanking. No, this doesn't mean, that you can't tank the bosses. Threat has been a non-issue throughout WotLK (anyone remember the Shattered Halls pulls in BC?) and DW tanking today simply means that you shouldn't go afk while tanking.

Still, I'm looking forward to patch 3.2.


My respects to you sir! *salute* it's rare to find an honest man, let alone a DW tank :P grats on batting off any of the nay sayers, I have to ask though; how?
how in the sense that, how does your guild let you dw tank? Most from what I see are wanna be elitists, and is your survival as DW ever thrown into question?

Inspiri
07-22-2009, 07:35 AM
I too also DW tank; been doing it since I got Broken promise and Red sword of courage, then was able to upgrade to last laugh and now I DW titanguards. I push 4.5k to 5k tps and I can spike well over 10k if "the stars align" (and a tricks). My spec is very odd in that it, for now, is a hybrid along all three tree, but it works. Come patch I did not see a jump in my threat yet; but this could have been to the fact I dont have the frost rotation yet.

But all in all to get to a point, I think the expertise hard cap will be MORE important come patch because now we cant rely on frost strike not to miss and now all our strikes have an extra chance to be parried. We no have to worry about our off hands; which is why I dont believe I saw a tps jump yet. I am going to try gemming for expertise and stam and see if that makes up the difference; I will try to post some numbers if I can get around to it!

Great discussions here always an enjoy to read and please comment about the OH and how we now have to worry about it being parried on all our strikes.

lyd
07-22-2009, 09:01 AM
My respects to you sir! *salute* it's rare to find an honest man, let alone a DW tank :P grats on batting off any of the nay sayers, I have to ask though; how?
how in the sense that, how does your guild let you dw tank? Most from what I see are wanna be elitists, and is your survival as DW ever thrown into question?you get your guild to let you DW tank by knowing why the math works and running with people who aren't idiots. anyone who heals a properly-geared dw tank sees that there is little difference in effective incoming damage. I raid in pugs regularly and no one has ever questioned my playstyle. Probably because they see that i'm loaded up with epics and they assume that i must know what i'm doing if i've made it this far.

That being said, I have been questioned before. We have a raiding alliance for 25-mans, and the conversation went something like this:

Q:"how do you account for the massive amount of parry-hasted attacks from DW'ing?"

A:"with the removal of crushing blows, parry-gibs were basically removed from the game. parry-hasted attacks only account for a small amount of incoming damage. and by gearing for expertise, you essentially eliminate the gap between 2h and DW that parry-haste creates."

Q: "aren't you gearing for expertise at the expense of other stats?"

A: "no, because you're question assumes that expertise is a bad stat to gear for in the first place. Expertise increases both threat and mitigation from decreased parry-haste chance up to the soft cap, which benefits any tank regardless of class. once you hit the soft cap, parry-haste accounts for so little incoming damage that it becomes irrelevant. not to mention warriors generate more parry-haste opportunities than a DW tank, on average. so saying a DW tank is too vulnerable to parry haste is also saying a prot warrior is too vulnerable to parry haste."

Q: "ah, i see."




I think the expertise hard cap will be MORE important come patch because now we cant rely on frost strike not to miss and now all our strikes have an extra chance to be parried.i was under the impression that ToT procs worked like enhancement shaman Stormstrikes? if the mainhand hits, the damage is dealt automatically as a function of off-hand damage, but is not literally a "strike." is this not the case? did i completely misunderstand the mechanics of stormstrike?

Splug
07-22-2009, 09:13 AM
i was under the impression that ToT procs worked like enhancement shaman Stormstrikes? if the mainhand hits, the damage is dealt automatically as a function of off-hand damage, but is not literally a "strike." is this not the case? did i completely misunderstand the mechanics of stormstrike?From my brief testing on the PTR, you are correct: there's one roll to determine whether the attack connects. Only a single dodge/parry/miss will be recorded on failure, and two successful swings are logged on hit. I never got a hit/miss combination.

As for the dual wield survivability question, I gain just over 2% avoidance from using a pair of XT hammers instead of Voldrethar. It actually reduces the throughput damage taken; the real loss I see is from the lower strength values on a pair of tanking weapons relative to two-handed weapons, which cuts into non-weapon based strikes as well. If the threat is workable, the damage intake should actually drop.

-Splug

lyd
07-22-2009, 10:09 AM
great job splug. It's interesting to hear that you experienced it that way. the added avoidance from DW at the expense of threat actually *decreases* total damage taken.

just shows further how small the amount of damage that parry-haste from DW causes actually is, so much that when combined with the higher avoidance that tends to accompany DW'ing, it actually makes you more survivable than a 2h, not less.

lol, i'd love to see some WWS reports comparing your results. the purists aren't going to like this one...

Splug
07-22-2009, 11:40 AM
Hrm, I suppose I could throw some hammers on for Freya tonight, but the fight has a lot of movement and I do occasionally get nipped in the back (or caught by a silence so I can't keep blade barrier/cooldowns up). Vezax would be a better target but I'm leery about threat demands; it gets close even going full bore with the 2h.

Algalon would actually be the best test case since threat's totally irrelevant after the initial 30 seconds or so, but we don't normally record WMO's from 10-man and I'm dps for that fight on 25.

-Splug

Splug
07-22-2009, 09:18 PM
Last week's raid (2h tanking) (http://www.wowmeteronline.com/combat/log/824435) - Damage taken for Vezax (http://www.wowmeteronline.com/combat/6088054#damagetaken)

This week's raid (dual wielding) (http://www.wowmeteronline.com/combat/log/835553) - Damage taken for Vezax (http://www.wowmeteronline.com/combat/6221908#damagetaken)

Results are exactly as expected, though perhaps more extreme: drastic drop in threat output, drastic drop in damage intake. Damage income per second dropped from 3,593 to 2,559. Spec and equipment were exactly the same, with the exception of swapping Worldcarver for a pair of Sorthalis, Hammer of the Watchers. However, I believe my bone shield uptime was drastically better tonight, both due to longer sustainance after casting and a poor refresh rate the week before (5 casts tonight vs 2 last week). There was certainly some operator error involved, and that will need to be addressed - looking directly at the hit intake count should accomidate that, as well as absorb shields.

Looking directly at the incoming swings, Vezax connected with 92 hits over 9:05 while dual wielding (avg 5.924 sec between hits, or 0.168807 hits per second), and 91 over 8:26 while 2h tanking (avg 5.5604 sec between hits, or 0.179842 hits per second). Ironically, a major contributing factor may have been the faster mainhand swing - with more rune strikes fired, Shadow of Death (sigil proc) had a higher uptime (88 rune strikes vs 65 - only two more total activations, but significantly more refreshes result in 64% uptime vs 55%). There was also one additional activation of my avoidance trinkets tonight relative to last night (resulting in 17% uptime vs 16% uptime - possibly not a major factor since the fight lasted longer as well).

Overall, the net change in damage increase seems to largely be due to the bone shield uptime (41% tonight... which seems extremely high), and the way that interacts with power word: shield consuming a disproportionately high amount of damage. However, there IS a small reduction in hits intaken per second as well, and a net gain to throughput avoidance due to sigil proc uptime. I apologize for the small sample size, but we one-shot everything tonight and only had three bosses left before we ran out of things to tank. I hope these are still useful.

-Splug

lyd
07-23-2009, 06:51 AM
this is outstanding Splug. I'm sure everyone appreciates you taking the time to run the reports. I certainly do :)

am i reading that right? Melee miss rate on Freya went up by 7% while DW'ing?

WoW Meter Online - Combatlog Replay (http://www.wowmeteronline.com/combat/detail/163676893#damagetaken)

WoW Meter Online - Combatlog Replay (http://www.wowmeteronline.com/combat/detail/166813752#damagetaken)

54.3% versus 61.9%. amazing.

Splug
07-23-2009, 07:56 AM
There's a lot of movement on that encounter, and I tend to get poked in the back occasionally. Data for avoidance on Freya is going to be sketchy... though misses should be omnidirectional, so that shouldn't be effected. I net lose defense and the SSG rune (and thus miss chance), it's just raw dodge/parry that go up. I'm guessing the miss difference there is pure RNG. EDIT: Oh, you meant total avoidance, not pure miss. Yeah - a big factor there is going to be hits in the back while repositioning. Usually, "run like hell and get away from the raid with Nature's Fury" is less total damage done than backing/strafing slowly away to ensure she's in front of me. So it's not random, I may have just had better facing last night. 62% avoidance is about right for my gear, 54.3% is low. That said, it should statistically have increased - just that it's hard to measure how much of that increase was better play and how much was the weapon change. I wouldn't throw it out completely, but I would caveat any observations on avoidance in that encounter.

Anyway, glad I could help. The Vezax results should be pure (it's just a really long tank and spank from my end, which is perfect for this) and back the math I'd done back before WotLK even hit. The only consideration would be using an avoidance 2h such as ID or Rune Edge to bump up 2h avoidance by another 1-2%, but even then it should be a very small gain.

-Splug

Dosvi
07-23-2009, 10:23 AM
Splug, something you may want to try (and something I should go try), is to go beat on Anachronos. (He's the dragon outside caverns of time). He doesn't hit hard, so you will stay up easy. He has 1.6m health and despawns at 20%. You could just soft reset him after a certain point and would give you fairly stable data.

Torn
08-03-2009, 11:02 PM
Slow tanking weapons - will there be any in 3.2?

Satorri
08-04-2009, 07:18 AM
None yet discovered that I've heard of, but I haven't been paying terribly close attention.


Also, bear in mind, the bosses in the Coliseum raids will be released one at a time.

Marrz
08-04-2009, 08:35 AM
So, this thread has some good information and theroycrafting about DW tanking in 3.2. I personally would love to try it.

That said, what would you guys recommend for our combat stats before charging in to Ulduar 10 with a weapon in each hand?

I'm assuming the following, but would like to hear your opinions.

Expertise: 6.5% (Can't be dodged, can still be parried)
Hit: Somewhere above 8%. Far above? Obviously can't cap white hits.

I would also assume that I would be taking Virulence and Nerves of Cold Steel.

Thanks in advance.


-Marrz

Satorri
08-04-2009, 08:59 AM
Nerves of Cold Steel, yes, Virulence, maybe, it's worth making your HB's hit more often for sure. Especially if you won't have a boomkin or spriest which for many will be far less common in 10-mans. That said, I'd prioritize survival over Virulence, as generally as is possible.

Splug
08-04-2009, 02:46 PM
There aren't enough points available for me to pursue Virulence; I'm fairly confident Scent of Blood and Bladed Armor both have better returns, and I'll need 52 in frost to reach everything I want from the bottom of the tree there.

I'm not too worried about the hard hit/expertise values. Staying over the dodge cap will be worthwhile, pursuing beyond that or stacking more than about 4% hit is pushing excess threat. If gear makes more conveniently available, cool. If not, I'll run lower threat gear and continue to carry around my Pyrite Infuser as a quick swap-in fix for encounters with asymetric damage levels.

-Splug

Krehinth
08-04-2009, 07:24 PM
So slow/slow, fast/slow, slow/fast, or fast/fast? And y?

Feanorr
08-04-2009, 08:39 PM
As already mentionned in this thread, idealy slow/slow; but for tanking there isnt any slow weapon as far as I know but even with a fast/fast combo, the dmg/threat on specials should be better than with a 2H so...

Zivh
08-04-2009, 10:18 PM
So I am curious....I am mostly all geared in Ulduar 10/25 gear but my 1h weapons for DW tanking (which I tried out briefly as Unholy) are only ilvl 200s...how do you think the threat generation would be with those? Probably too low? Should I hold off on trying out the DW until I can get a couple new ones from Ulduar/CC?

I had stopped rolling on the 1handers after having some threat issues with DW and switching back to 2H Blood. I am thinking maybe I should have been rolling on those instead of letting them go to the dogs for offspec.

Also, has anyone figured out a viable DW Frost build since the patch dropped?

Edgewalker
08-04-2009, 10:21 PM
After the first 25 boss in Ulduar, and another HM Ulduar 25 clear, I find DW to be very lackluster. Unbreakable Armor isn't as good as Vamp Blood, you lose Hysteria, you lose VoTW, and you lose Death Strike as a constant heal. For the return I was hoping for a spec that allowed consistently higher threat generation, but was quite let down.
Meh. Was using the very top end DW gear as well, double Invars, 13% hit and capped 56 expertise.


Edit - 13/53/5

Splug
08-04-2009, 11:19 PM
Dual wielding threat and throughput mitigation are fine. I was playing around with a death-strike based DW spec initially, then tweaked to obliterate-based. There was a non-trivial threat/survivability tradeoff in there, but ultimately the throughput on either side was fine. The threat sacrifice to use death strike as a primary attack was suprisingly small; by focussing on heavy RP generation via fast-swinging weapons and SoB,

The problem is that frost is extremely hurt for cooldown power. It was behind before, but the 1 minute IBF cooldown somewhat bandaged the problem with the longer uptime. The relative value of Guile of Gorefiend has been halved from a throughput perspective, and outright crippled from a "secondary cooldown standin" perspective. From this angle, frost plays similar to a 3.1 paladin, with a 2-minute recast on holy shield (unbreakable armor). Substituting death strike for obliterate assuaged the lack of a shield slightly, but the breaking point far and away was the lack of mitigation on demand.

-Splug

Edgewalker
08-04-2009, 11:51 PM
Dual wielding threat and throughput mitigation are fine. I was playing around with a death-strike based DW spec initially, then tweaked to obliterate-based. There was a non-trivial threat/survivability tradeoff in there, but ultimately the throughput on either side was fine. The threat sacrifice to use death strike as a primary attack was suprisingly small; by focussing on heavy RP generation via fast-swinging weapons and SoB,

The problem is that frost is extremely hurt for cooldown power. It was behind before, but the 1 minute IBF cooldown somewhat bandaged the problem with the longer uptime. The relative value of Guile of Gorefiend has been halved from a throughput perspective, and outright crippled from a "secondary cooldown standin" perspective. From this angle, frost plays similar to a 3.1 paladin, with a 2-minute recast on holy shield (unbreakable armor). Substituting death strike for obliterate assuaged the lack of a shield slightly, but the breaking point far and away was the lack of mitigation on demand.

-Splug

That's exactly by point. We knew the trade-offs already existed, which it could have made up for by having exceptional value elsewhere. It didn't. And the threat is extremely sub-par. While it certainly CAN hold agro, it is still quite a bit behind a traditional blood or frost spec. For most raids, 3000 TPS on Vezax simply wouldn't cut it. I do roughly 6500+ with some parses over 7500 and still have DPS nipping at me off and on.

As for tanking weapons, new Ulduar gear lets you easily use 2 DPS weapons and maintain defense cap. I was sitting at almost 570 defense after getting 2 new pieces before I shuffled things around.

NexsusTheFrostDeathKnight
08-04-2009, 11:52 PM
I find DW to be very lackluster. Unbreakable Armor isn't as good as Vamp Blood,


I disagree, unbreakable armor is the only one that will give a good dps buff as well, to me, vamp blood was only attractive back when it had a 1 min cooldown.
But that's just my opinion, I'd like to tank as blood, but I've it so 'paniced' when compared to frost.

Satorri
08-05-2009, 06:16 AM
Mmm, Vamp Blood is a wonderfully delicious defensive CD. Unbreakable Armor is actually the only tank CD DK's have that could actually be used for a threat boost as well.

I'm looking forward to a night of stability where I can actually start testing the new lay of the land against real situations.

NexsusTheFrostDeathKnight
08-05-2009, 03:27 PM
I'm looking forward to a night of stability where I can actually start testing the new lay of the land against real situations.

Im looking forward to hearing about it :D
Thereś nothing Id want more than to DW tank

Feanorr
08-05-2009, 09:41 PM
Mmm, Vamp Blood is a wonderfully delicious defensive CD. Unbreakable Armor is actually the only tank CD DK's have that could actually be used for a threat boost as well.

Imo, it's not even a tank CD, it's a threat/dps cd; the mitigation part is worst than pal/war block and it's only up 20sec/2min. I cant see any situation where it could be used like any other tanks cd (i.e. to prevent/reduce a damage spike period).

That's my main concern about frsot tank, we only have one real tank CD (IBF) who have been reduced to 2min; I will test it a bit more but I am already thinking about going back to blood, to at least being at same HP than war (cause all the comparison showing that we was at same HP was in blood, as far as I know; frost and unholy are behind).

Fini
08-05-2009, 09:56 PM
Im looking forward to hearing about it :D
Thereś nothing Id want more than to DW tank

That's what dual-spec is for :) I kept my original blood tanking spec and picked up a 'standard' deep frost DW spec. First night we tried out 10 man colliseum. Not the best place to try out a new spec, especially if your guild hasn't piled up enough T8. But I saw alot of potential there, threat was through the roof, but I want a good Naxx or Ulduar run so I can look at the parses.

Satorri
08-06-2009, 06:29 AM
Fea, I find your comments disturbing, you are usually far more insightful.

Similar to the complaints about shield block for warriors. It's not a game winning massive survival effect, but it is a solid and steady damage reduction tool. It doesn't shine like IBF may, but it certainly has its value. Maybe this will make more sense in perspective when I finish my next installment on the Survival Contribution Breakdowns.

Pruke
08-06-2009, 07:44 AM
I haven't had the chance to raid with 0/53/18 yet but I can see the potential for threat problems since I kill 81 elites in IC very very quickly. That much burst has to be a threat hog.

Only problem I have is using the standard rotation of IT-PS-OB-BSx2 (dump) 3xOB (dump) that if rime and KM arent up I am sitting at full RP and if I dump with FS I will eat the KM proc (since KM usually proc's during the 1st part of the rotation). Typically during the 1st part of the rotation KM will be up but as I finish out the 2nd BS I am at 120 RP and have to chose to dump with FS burning KM or just sitting there with RP and starting the OBx3 and praying Rime procs during that time. If not the KM is eaten with the IT. I havent tested it enough to determine if I should just sit there at full RP and save the KM or use it on the FS dump and hope it procs again. Anyone else tested this and have any input?

Edgewalker
08-06-2009, 07:56 AM
Fea, I find your comments disturbing, you are usually far more insightful.

Similar to the complaints about shield block for warriors. It's not a game winning massive survival effect, but it is a solid and steady damage reduction tool. It doesn't shine like IBF may, but it certainly has its value. Maybe this will make more sense in perspective when I finish my next installment on the Survival Contribution Breakdowns.

It's not that it doesn't have clear value as a damage mitigator, it just isn't even in the same ballpark as Vamp. Blood or Whirling Bones. In addition it barely scales and uses a frost rune... Sometimes you really don't need survival breakdowns to figure out somethings practical use in game. It's fine for Kologarn's rubble and Naxxramas' spider packs, but when it's an actually difficult fight it just doesn't cut it.

Satorri
08-06-2009, 08:56 AM
Yay, math time.

First, nitpick: Vamp Blood costs a Blood rune, Bone Shield costs an Unholy rune.

Second, boss swing, gross simplification!!

Here's our DK tank: 48k health, 65% rdx from armor (~31k armor), 9.75% rdx from Frost Pres/Blade Barrier, and 58% avoidance (and 1200 Str)
Our Opponent: 60k unmitigated hits every 2 sec.

Vamp Blood will give you 20 sec with 55.2k health and +35% on all incoming heals. Assuming you have good heals incoming, this is very powerful though it doesn't reduce your damage taken whatsoever. Functionally, we'll just assume heals are sufficient and you aren't taking hits big enough to one-shot you. If all incoming heals, adjusted for proportion of total healing done to you, average out to 50% overhealing (this may be generous or conservative, depending on the situation), then you could say this is effectively 17.5% damage reduction on everything that hits you during the duration. This is not fixed, Vamp Blood's relative value is dependent on the incoming heals, proportional overhealing, and requires that you don't need additional mitigation to survive the hits. Strong, yay. If we use 17.5% apparent rdx, that means the DK will appear to take about 3365 dps (3859 actual) in 18.4k swings as much as every 2 sec.

Now, let's look at Bone Shield (we'll compare them all unglyphed). With 58% avoidance we can approximate the average uptime at about 20 sec (convenient, eh?). During that time, all (4) swings taken will be reduced by 20%. That means the tank will take 3087 dps in 14.7k swings as much as every 2 sec.

And finally, Unbreakable Armor. Note the example is an average swinging boss with big fat hits. UA will be up for 20 sec and remove 1550 dmg from each hit, and from the Str buff will get ~1.2% more parry. That means that the tank will take 3432 dps in 16.8k hits as much as every 2 sec.

So, for matching uptimes, during the effects, the Blood tank will take the most damage but be the most healable (assuming no big hit spikes kill him), the Unholy tank will take the least damage (assuming the avoidance holds out for optimal duration), and the Frost tank will take slightly more damage than the Unholy tank, less than the Blood tank, and will have unvarying protection.

I don't think there is a clear winner in the comparison, though certain styles will appeal to some over others.

If you want to add the complication of "but there are other forms of damage incoming!!" The Blood tank still won't mitigate any of them, and the Unholy tank will run a higher risk of consuming the bones faster and shortening the duration. I love Vamp Blood, it is surely my favorite, but I don't think Unbreakable Armor is somehow dwarfed by the other two.

Splug
08-06-2009, 09:04 AM
Similar to the complaints about shield block for warriors. It's not a game winning massive survival effect, but it is a solid and steady damage reduction tool. It doesn't shine like IBF may, but it certainly has its value. Maybe this will make more sense in perspective when I finish my next installment on the Survival Contribution Breakdowns.
This is more or less the problem: it's an extra tertiary cooldown (comparable to shield block) instead of a secondary (comparable to last stand); its effectiveness is on the wrong tier. From an overall throughput analysis, this is offset by frost's longer duration on icebound fortitude. From actual field testing, the extra six seconds of IBF often are excess - to paraphrase something a friend often said regarding shield wall, if the first twelve seconds at half damage aren't going to save you, are the next six?

The main issue is just an inability to rotate a percentage based cooldown every minute. This is an option for every other tank in 3.2, as well as both other deathknight trees. The fact that paladins were reliant on Glyph of Salvation as a secondary/tertiary cooldown was deemed an unacceptable problem and brought about the 3.2 version of the Ardent Defender. Unbreakable armor is not worthless in all circumstances, but it falls drastically short in a fairly common case.

-Splug

Satorri
08-06-2009, 09:29 AM
Did you read the above?

Is any amount of attempted comparison going to open your minds to a new appreciation for the different sorts of damage prevention in their relative value?

Krenian
08-06-2009, 10:14 AM
I haven't had the chance to raid with 0/53/18 yet but I can see the potential for threat problems since I kill 81 elites in IC very very quickly. That much burst has to be a threat hog.

Only problem I have is using the standard rotation of IT-PS-OB-BSx2 (dump) 3xOB (dump) that if rime and KM arent up I am sitting at full RP and if I dump with FS I will eat the KM proc (since KM usually proc's during the 1st part of the rotation). Typically during the 1st part of the rotation KM will be up but as I finish out the 2nd BS I am at 120 RP and have to chose to dump with FS burning KM or just sitting there with RP and starting the OBx3 and praying Rime procs during that time. If not the KM is eaten with the IT. I havent tested it enough to determine if I should just sit there at full RP and save the KM or use it on the FS dump and hope it procs again. Anyone else tested this and have any input?

/cough...That's a DPS question in a whole lotta tanking posts. I'll cover this.

Eat the KM proc. It's not a big deal however if you have Rime come up, you'll have it up for I think 20-30 seconds. This allows you to wait for a KM proc to happen. You should prioritize KM for Howling Blast, but if you cannot, just eat it with Frost Strike, and it won't be the end of the world. Don't worry about it and eaaaaaaaaaaat the KM buff :D

As another note, a whole lotta stuff having to be learnt for tanking. You people hurt my head. -.-

Pruke
08-06-2009, 10:51 AM
Duh really didnt even consider the fact it was a tank thread :(

Thx though sorry for the derail.

Splug
08-06-2009, 11:46 AM
Did you read the above?

Is any amount of attempted comparison going to open your minds to a new appreciation for the different sorts of damage prevention in their relative value?The problem is that you're using a strict throughput damage analysis. That's fine for some cases - in fact, it would likely be both valid and favorable for Algalon. But the problem isn't throughput - it's TTL, particularly when using the secondary cooldown to rotate with IBF for high-surge damage scenarios. It doesn't shave enough damage off to survive a ~55k fusion punch, 30k/sec plasma blast (~5.3% reduction), or 40k unbalancing strike with surrounding autoattacks. I don't have the numbers for Vezax's autoattack during surge of darkness, but I suspect it's not particularly promising there either (then again, bone shield was too close for comfort as well). There are at least three encounters where the spec simply cannot fulfill roles the other trees are able to without fairly drastic strategy adjustments involving remapping additional external cooldowns.

It works acceptably on paper, against a target doing more or less fixed damage at a fairly normal damage level. The problem is using the cooldown where it's needed - abnormal damage levels, which occur fairly frequently in Ulduar - comes up short.

-Splug

Edgewalker
08-06-2009, 02:41 PM
Yay, math time

You use way too much math and not enough common sense relating to how Blizzard actually designs bosses and scenarios where tanks ACTUALLY die. Math is fun. You can make great gear comparisons for EH, you can parse exact TPS increases... but you can't ever measure intangibles. Would Armsman to gloves be a better threat enchant than 20 hit? In most scenarios, probably. Will Armsman to gloves matter if you miss 3 attacks in a row off an Assault Bot spawn and a DPS kills themself? No it won't. (Please don't attempt to break down a hyperbolic example). In a mathematical scenario will you take less damage with 2 dodge trinkets then the 2 HP trinkets over the course of a fight? Sure. Are you still better suited overall with a higher stamina pool and clickable trinkets though when you are actually in danger of death? Yes.
UB is the EXACT same way. It does reduce damage by a decent amount, and does a good job of constant overall mitigation, but where tanks actually die is spike damage. Plasma Blast, Rune Punch, Unbalancing Strike, Every Other Vezax surge. The damage mitigation isn't broken down into a nice little math package then, but a messy spike of parry hasted melees, Sif frostbolts, and Doomfire. It comes up short, and in a BIG, very noticeable way.

Feanorr
08-06-2009, 09:33 PM
Fea, I find your comments disturbing, you are usually far more insightful.

Well I am sorry to hear that but even after reading your maths (that are always welcome) I keep it; it may be as good as BS for total damage reduced, but that's not what kill a tank in wotk (or pre-BS nerf DK would have been god, we was taking half the damage of others tank at this time; and we werent even considered OP, like in 3.1).
Like splug said, the important thing about a CD is to be able to help survive a high-surge scenario (see splug's examples), and UA isnt doing the job.
It may change if futur bosses are more about total damage but for now, every wotk boss has been about high-spike damage. Even when mana is a problem (vezax by example) it still most of the time the high spike period that will kill the tank, not the overall damage.

I cant think of any boss in Ulduar (or even before) where I would consider UA to be enough to use it as I use other CD. At best I could use it like we were using BS before, everytime it's up to reduce total damage; but it would be up only 1/6th of the time.

Krenian
08-06-2009, 09:35 PM
UA's actually pretty damn terrible. I actually hate the ability to be honest with you. I don't think it's a good enough ability.

Zivh
08-06-2009, 11:42 PM
Nice breakdown above, Satorri.

So has anyone actually tried Ulduar 25 in the new Frost spec with decent gear rotating in UBA? As much as I QQ'd about the patch I am liking it in 2H Blood so far, but I am still debating on whether or not I want to try a DW Frost in raids this week. I personally don't think UBA is that bad of a CD, dam reduction on every incoming hit, plus 25% strength for a bunch of extra parry and threat. Its no Vampiric Blood, but hey, it got me through Naxx 25 just fine back in the day.

Feanorr
08-07-2009, 12:17 AM
I just though about a much similar cd in the result: UA is pretty close in effect to Mark of blood, exept for the parry part. With my gear, Mark of blood is aprox 1K5 heal and UA is 1200 dmg reduction every hit. does that seem accurate?
And Mark of blood can have it's use, but it's no way near good as Vampiric blood or even BS

Satorri
08-07-2009, 07:04 AM
Every ability has its values, and it's easy to highlight where something fails and another ability succeeds, my desire is only to highlight where the rather unpopular abilities do just fine, if only for the sake of mitigating the lemming herd of blind followers who will go parrot to trade chat, "Unbreakable Armor is worthless! [I read it on TankSpot]"

Unbreakable Armor is less helpful in some of the places where we want to rely on CDs most, but it is anything but a broken or unhelpful ability. I just want to keep that in perspective.

And Edge, I use math, yes, to support practical understandings, and also to rein in some "practical" understandings that get out of line or inflated with filtered perception. Your sin in this is that you flout opinions without every backing them up with supporting data or calculations. It's a hard distinction but there is such a thing as Opinion-Truth and Factual-Truth. Just because you have done it, doesn't mean you understand the nature of how you did it. Just because it works doesn't mean it is "best." And just because something is the "best" doesn't make it the "only viable choice." Rhetoric which is grossly abused in our community.

You can say, "Unbreakable Armor is awful" without demonstrating, and you are speaking your truth, you really believe that. That doesn't make it factual, nor is that at all supported, even by circumstantial math.

Seeing as this is a forum for discussing and elevating knowledge on abilities, unsupported opinions are rather out of place other than to illustrate the mental state of the community, be it ignorant or actually informed.

Splug
08-07-2009, 01:07 PM
The problem is that there are cases where damage comes in rapid, small bursts - and there unbreakable armor is superior, but bone shield / vampiric blood can get the job done. In the reverse cases, where damage comes in abnormally large bombs, unbreakable armor is not only inferior, it is incapable of providing survival. It is difficult to justify a talent for across-the-board balance when the difference in some cases becomes a binary death/non-death scenario, where the alternative does not have any equivalent binary shortcomings in current content.

That said, on a note of irony I was in a position last night where I would have greatly preferred unbreakable armor, but was required to apply ebon plaguebringer and thus unable to dual spec back to frost. Bone shield was about as far from ideal as possible, even though it was still sufficient for the application (barely).

-Splug

EDIT: I think we're beating a dead horse at this point; everyone has expressed their viewpoints fairly well. To get back to the original question in the thread - dual wielding's threat as frost was viable in my experience, and dual wielding's survivability with respect to the same tree as 2h is comparable. Inter-tree comparisons are an entirely different beast.

Edgewalker
08-07-2009, 09:45 PM
And Edge, I use math, yes, to support practical understandings, and also to rein in some "practical" understandings that get out of line or inflated with filtered perception. Your sin in this is that you flout opinions without every backing them up with supporting data or calculations. It's a hard distinction but there is such a thing as Opinion-Truth and Factual-Truth. Just because you have done it, doesn't mean you understand the nature of how you did it. Just because it works doesn't mean it is "best." And just because something is the "best" doesn't make it the "only viable choice." Rhetoric which is grossly abused in our community.


You misjudge what I said. Math has it's purposes, but making up a scenario that mathematically turns Unbreakable Armor into a tanking cooldown that is even remotely optimal for a progression oriented tank is not one of them. Encounters of difficulty where Unbreakable Armor is even in the same ballpark as Vamp. Blood or Bones don't exist. In this way it IS a broken cooldown. The threat increase doesn't account for the loss in practical use and something in the ability needs to, and will change. I've been contributing here since the start of this site, I don't always bother writing a 15 page essay on why abilities are and aren't good, but that never changes the final verdict.

Xequecal
08-08-2009, 04:17 AM
I've tried DW tanking for a bit and found it to be far superior to Blood in 3.2. I use a 13/51/7 spec, I wouldn't tank Vezax with it but everything else is great. Hodir is no problem due to you having Frost Strike and Howling Blast to combo with Singed.

Unbreakable Armor kinda sucks but the avoidance advantage of Frost with 2 tank weapons is immense. You end up like 7-11% ahead of the other tanks in avoidance, depending on your gear level and race choice. (Warriors don't have a 200 dodge sigil/libram, Paladins can't be Night Elf.) There's also some really damn good 3.2 itemization for DW in the game, you can dual-wield these (http://db.mmo-champion.com/i/48044/ardent-guard/) which is utterly amazing.

Satorri
08-10-2009, 07:22 AM
Edge, you have yet to contribute anything to this discussion other than your statement: Unbreakable Armor doesn't work.


Saying it doesn't make it so.

Edgewalker
08-10-2009, 09:27 AM
Edge, you have yet to contribute anything to this discussion other than your statement: Unbreakable Armor doesn't work.


Saying it doesn't make it so.

Sure I have, you just can't seem to grasp it or understand it.
Let me say it again, slowly.
Unbreakable. Armor. Is. Static. Damage. Reduction.
This is a fact.
This is why it does not work acceptably as a survival cooldown in comparison to VE or Bones (or Barkskin/SI or Last Stand/SW or AD/DP if it is a total comparison). When bosses hit hard, it is less effective.
Tanks die when bosses hit harder.

I could make you a damned flowchart or make up a page of math like you LOVE to do, but frankly it's, as described earlier, a waste of time.

Satorri
08-10-2009, 09:42 AM
Oh well. Now I see the light, this time when you say the same thing, it totally has opened my eyes.

Edgewalker
08-10-2009, 09:54 AM
Oh well. Now I see the light, this time when you say the same thing, it totally has opened my eyes.

You being stubborn doesn't take away from me being right.
Unless you feel like refuting it?
Explain how UA is effective as a second surge Vezax CD over bones or VB, or explain how it would work better for a 3 charged Steelbreaker doing a 52K rune punch over VB or bones, or explain how it matches any other tanking CD on a 9 stacked Thorim?
You can't, because it doesn't, and that's the point.
I don't have to say anything else because I don't NEED to.
Tanking cooldowns aren't about taking the lowest amounts of incoming damage in given periods of time. They are about surviving. Period.

Ciderhelm
08-10-2009, 09:57 AM
To stir the pot a little: How is Edgewalker incorrect? His argument is similar to one we've used for a long time as Warriors/Paladins in regards to Block. Burst kills.

Satorri
08-10-2009, 09:57 AM
Not interested in butting heads, my only interest was to illustrate the value of mechanics instead of editorializing things as "good" or "bad" without supporting information.



My point is simply that. We've established that Block and UA (essentially the same sort of mechanic) are weak in specific scenarios, i.e. against a boss that will hit you with a great big swing. That said, I did the math above not to refute that, but to flavor it. Even against a big hitting boss, it still holds its own in the net effect of damage reduction.

I'm not trying to sell it as a better cooldown than anything, I happen to agree that it is the weakest of the three, I just also think that unsupported "this ability is worthless" statements is misleading. If you want people to blindly follow what you say, that's all well and good, if you want them to trust your words, show them why.

GravityDK
08-10-2009, 05:08 PM
Last week's raid (2h tanking) (http://www.wowmeteronline.com/combat/log/824435) - Damage taken for Vezax (http://www.wowmeteronline.com/combat/6088054#damagetaken)

This week's raid (dual wielding) (http://www.wowmeteronline.com/combat/log/835553) - Damage taken for Vezax (http://www.wowmeteronline.com/combat/6221908#damagetaken)

Looking directly at the incoming swings, Vezax connected with 92 hits over 9:05 while dual wielding (avg 5.924 sec between hits, or 0.168807 hits per second), and 91 over 8:26 while 2h tanking (avg 5.5604 sec between hits, or 0.179842 hits per second).


Good data Splug.
I wonder if we could find any incoming parry-haste attacks in that, bit late for me to check right now.

Big loss of threat, but you felt it was still viable overall. Do you recall how close the dps got to catching you, or were you miles ahead?

Lastly, has anyone recalculated the risk of parry-gibs now that Frost Strike can be parried? The previous maths we used on this analysis relied on FS being an unavoidable attack.

jaydee
08-11-2009, 02:26 AM
How long is Vezax's Surge of Darkness? Will a glyphed 4-piece'd AMS + UA + dodge trinket provide you enough cover when IBF is on CD?

Feanorr
08-11-2009, 02:39 AM
My point is simply that. We've established that Block and UA (essentially the same sort of mechanic)

Personnaly that's all what I was saying: UA is similar to block (or holy shield, exept it's only up 20 sec every 2min), and not usable like VB, BW, AD, LS, etc. So as frost we got only one "hard hiting" CD

GravityDK
08-11-2009, 02:45 AM
OK I've used the 'browse events' feature and log browser on WMO to find how many times Vezax parried you Splug. Count them using 'find' in your browser, for 'parried'. Results are worrying.

Two-handed fight (events (http://www.wowmeteronline.com/combat/eventfilter/6088054/1247756362000/1247756872000/0x0400000002A44DCF/0xF1300081F702CDBD/HIT/nil))


14 times over 8:26 = 1.6600 parry/minute

Dual-wield fight (events (http://www.wowmeteronline.com/combat/eventfilter/6221908/1248268353000/1248268903000/0x0400000002A44DCF/0xF1300081F7002864/HIT/nil))


45 times over 9:05 = 4.9541 parry/minute

Sure it's a small sample, but 3x as many incoming parries shows an order of magnitude that is so large I don't think the small sample size would be a factor.

If I've not made a mistake, that is a significant parry-haste increase for the boss, which with 18K melee strikes could be a problem for some healers (perhaps not yours, Splug, but in general). The spike damage is concerning.

Could someone verify my count and conclusion?

lyd
08-11-2009, 05:55 AM
I always liked unbreakable armor :P although I can understand the arguments being made against it.

I think the "intangibles" argument works both for and against VB and BA, only because by definition you can't "math out" intangibles in such a fashion to show how grave the disparity between the abilities is, so the evidence is anecdotal. I also think we should be viewing these abilities in a "total package" format. How does VB fit in to the rest of blood's talents/abilities? Does frost have better passive bonuses that account for VB being such a powerful cooldown and UA being less so (same with BA)? how do you measure the relative impact of both in a "burst damage" situation?

I think the only real way to show these discrepancies indiscriminately would be to have a BiS tank stand in front of a boss capable of 1-shotting a tank sans cooldown, and apply each cooldown. If with UA the tank still gets one-shot, but with the other cooldowns he does not, then you have a quantifiable disparity. The question then becomes: how often does that situation occur, and it is enough to justify a redesign of UA? or do other aspects of Frost tanking prevent UA's smaller relative burst mitigation potential from being a game breaker?

Perhaps the answer is just simply better scaling. I can understand how a 5,000,000,000,000k fusion punch might make UA less desirable than the other 2 cooldowns. But i also get the feeling that blizzard wants 3 different, unique tanking cooldowns, and I really can't think of how they would do that with Frost without a static DR. Perhaps they could make it function similar to BA with charges, but have it start out more powerful than it is now, and reduces in efficacy with each successful hit landed on the tank. So overall mitigation remains the same, but it becomes more effective for burst on the first press for big game-stoppers like fusion punch.

Zivh
08-11-2009, 06:52 AM
Perhaps they could make it function similar to BA with charges, but have it start out more powerful than it is now, and reduces in efficacy with each successful hit landed on the tank. So overall mitigation remains the same, but it becomes more effective for burst on the first press for big game-stoppers like fusion punch.

Yes.

EDIT: Let me rephrase:

HELL YES.

Bashal
08-11-2009, 08:38 AM
If I've not made a mistake, that is a significant parry-haste increase for the boss, which with 18K melee strikes could be a problem for some healers (perhaps not yours, Splug, but in general). The spike damage is concerning.

Could someone verify my count and conclusion?

I took those same logs, and instead looked at the timing between the hits Splug took. I put both logs in a spreadsheet, and had it compute the time between hits.

With the 2-hander:
* Splug took damage an average of once every 5.566 seconds.
* 30 of the hits were in intervals of time less than 2 seconds apart
* shortest interval between hits was 1.521 seconds
* number of runs where hits were <2 seconds apart: 5
* Average length of runs where the time between hits was < 2 seconds apart: 2.8
* longest run of hits in a row that were < 2 seconds apart: 4

While dual-wielding:
* Splug took damage an average of once every 5.593 seconds.
* 27 of the hits were in intervals of time less than 2 seconds apart.
* shortest interval between hits was 1.580 seconds.
* number of runs where hits were < 2 seconds apart: 5
* Average length of runs where the time between hits was < 2 seconds apart: 2.2
* longest run of hits in a row that were < 2 seconds apart: 3

Overall I'd say dual-wielding comes out about even, or possibly marginally better (depending on how you look at the data) as far as the pattern of damage goes.

As far as this one set of encounters goes, I don't think taking heavy spike damage as a result of additional parries is an issue at all.

EDIT: if anyone is wondering why I picked 2 seconds as a threshold, I figured that 2 seconds was the minimum amount of time where healers could comfortably (relatively speaking) heal a tank enough so he could survive another hit.

Splug
08-11-2009, 10:14 AM
Thanks for the analysis Bashal - I scraped together a few numbers very quickly, but didn't go into that level of detail at all. Essentially, the demonstration is exactly what I was hoping to see - the increased avoidance offsets the parry haste by enough that even in 2-second burst time, there is not an appreciable risk gain.


Big loss of threat, but you felt it was still viable overall. Do you recall how close the dps got to catching you, or were you miles ahead?The data there was for dual wielding as unholy, which really isn't supported. For that encounter we use a lot of salvations and threat resets, so there's always someone riding very close - it's just a matter of them getting thrown back to zero (or near it) when it becomes a problem. I do not consider dual wielding threat as unholy reasonable; the goal was to compare incoming damage for the two weapon styles, then extrapolate that to frost builds where the threat would be reasonable in 3.2. Mostly, I just wanted to assuage some of the parry-haste concerns by putting their effect into perspective.


Perhaps they could make it function similar to BA with charges, but have it start out more powerful than it is now, and reduces in efficacy with each successful hit landed on the tank. So overall mitigation remains the same, but it becomes more effective for burst on the first press for big game-stoppers like fusion punch.While I could see that working much better and being interesting, something seems internally inconsistent with calling the ability "Unbreakable Armor" and then having it break down as damage is applied... ;)

-Splug

Bashal
08-11-2009, 10:36 AM
Thanks for the analysis Bashal

Anytime. My DK alt is a frost-spec tank, and I'm playing around with dual-wield tanking to see if I like it better, so this thread is of great interest to me.

One observation I made was that while the rune strikes are individually weaker, they are more frequent, making it harder for me to accidentally cancel a rune strike during a FS RP-dump, so I get more rune strikes off successfully. My threat actually went up some (I haven't 2h-tanked yet since 3.2 so I have no data, this is strictly anecdotal at this point).

But others more skilled than I am at tanking (of which there may be many) may not encounter that issue nearly as much.

GravityDK
08-11-2009, 12:34 PM
There is something logically inconsistent between my quick count and your solid analyais, Bashal.

Argument: If I counted nearly 3x as many parry-haste events, and Splug only had about 2% higher avoidance, you would expect to see some noticeable parry-haste. But your data shows no increases in overall haste (in fact, you had the opposite). Where did all the parry-haste go? It couldn't have been dodged, since he only increased avoidance by 2% (which is about 5% relative increase); not enough. Let's argue that 50% of the parry events came in the portion of the boss swing-timer where it doesn't actually speed him up, but if so, that'd still leave a lot of parry-hasted-attacks, which I can't see in your data either.

Problem: I think either (a) one of us is has the data wrong, or (b) my that argument is flawed.

Splug: thanks for clarifying you were Unholy DW tanking. I missed that. I'll ignore any data on its threat output.

Bashal
08-11-2009, 12:55 PM
All I can say is: that's what the logs said. You could argue that it was an anomalous (i.e. "lucky") run while dual-wielding.

However:

His avoidance should have actually been higher than an additional 2%. I don't have time to double-check facts, so please feel free to refute any of my conclusions:

* Dual swordbreaker adds 4% undiminishable parry
* Depending on the tanking weapons used he would have more (diminishable) dodge and/or parry.
* The amount of haste depends on when the parry actually occurred during the boss' swing timer. An early parry adds significant haste; a late parry does essentially nothing.

All these factors could account for the discrepancy. I said that it came out about even, rather than saying it was definitely better, because you could account for the slight variations (e.g., a longer "shortest interval" or 3 fewer swings less than 2 seconds) as statistically insignificant. In other words, it could have easily been the other way around -- 30 hits < 2 seconds while dual wielding and 27 hits < 2 seconds while using a 2-hander. It's not enough of a variance to say it was because he was dual wielding as opposed to using a 2-hander. But the stats are close enough to say -- for that specific data set -- that additional parry haste didn't seem to cause a problem.

Also, it's just one set of fights. It's not enough, on it's own, to definitively conclude anything.

We need more datasets for comparison to know for certain what is going on.

EDIT: one thing I can do is check the logs to see what his apparent avoidance was, i.e. how many dodges and parries, etc. I'll post a follow-up with that info.

Splug
08-11-2009, 01:22 PM
The 2% is about right. +4% from the swordbreakings, -3% for not having stoneskin gargoyle, +2*~1% pre-DR from the tanking weapons, -~.5% from not having the 2h (forceful deflection). The absolute avoidance increase should fall somewhere between 2% and 3%.

It's also worth noting that the dw fight was longer by about 40 seconds, and that both weapons are the same speed (and very near to the GCD) - so a few of those parries may be synchronized. Any time more than one parry occurs in the same boss swing cycle, it's almost assured that the second or third parry will fall into the dead zone where no further parry haste can occur. Both of those effects are relatively minor though; I can't see either one accounting for a factor of 3 shift. Then again, we're quickly accumulating minor factors (2% avoidance, synchronized parries, fight duration). There may not be any single feature causing the observed discrepancy, but a combination of smaller ones - particularly given the effects of dumb luck with a coarse granularity of data.

-Splug

Bashal
08-11-2009, 01:51 PM
Ok, looked at the logs again:

2-handed:
----------
* attacked by Vezax 217 times total
* parried 41 times (18.89%)
* dodged 58 attacks (26.73%)
* missed 27 times (12.44%)
* hit 91 times (41.94%)

Apparent total avoidance: 58.06%


Dual-wielding:
-------------
* attacked by Vezax 231 times total
* parried 44 times (19.05%)
* dodged 77 times (33.33%)
* missed 18 times (7.79%)
* hit 92 times (39.83%)

Apparent total avoidance: 60.17%


As predicted, you had about 2% more avoidance; although, the source of that additional avoidance seems to have come from dodge more than anything else. Dumb luck? You tell me.

Bashal
08-11-2009, 01:57 PM
The 2% is about right. +4% from the swordbreakings, -3% for not having stoneskin gargoyle, +2*~1% pre-DR from the tanking weapons, -~.5% from not having the 2h (forceful deflection). The absolute avoidance increase should fall somewhere between 2% and 3%.

Good points. I hadn't considered what you might have been losing in conjunction with what you were gaining.

GravityDK
08-11-2009, 04:14 PM
Then again, we're quickly accumulating minor factors (2% avoidance, synchronized parries, fight duration). There may not be any single feature causing the observed discrepancy, but a combination of smaller ones - particularly given the effects of dumb luck with a coarse granularity of data.


Maybe it is like the butterfly analogy used in chaos theory :).
It's puzzling.

lyd
08-12-2009, 05:34 AM
While I could see that working much better and being interesting, something seems internally inconsistent with calling the ability "Unbreakable Armor" and then having it break down as damage is applied...

well obviously they'd have to change the name ;)

Splug
08-12-2009, 09:11 AM
"Somewhat Breakable But Certainly Not Fragile Armor?"

Bashal
08-12-2009, 11:02 AM
"Somewhat Breakable But Certainly Not Fragile Armor?"

"Ablative Armor" :)

lyd
08-12-2009, 12:07 PM
If i may borrow one from the wow forums...



Let's ask Ollie Williams. What should we call the new UA Ollie?

...............…………..„--~~""”””~~-„……………………………………...........
…………………..„-~”:::::::::::::::::::::::::”~,………………………………………… .
……………….„~”\,::::::::::::::::::::::::::::::::: ”-,………………………………….......
…………….../;;;;;;|::::::::::::::::::::::::::::::::::::\……………… ……………............
……………..|;;;;;;;/:::::::””~~--::::::::::::::-~~”::\……………………………...........
……………./”~,;;/:::::::/’\::::::::::::::::/’’\………………………….........
…………….\ : ::::::::\ . . • . ,/::::::::::::::\„ . • . ,/……………………………….......
……………/’::::::::::::::::”~~~”:::::/””~::::’\”~~”.::’\,………………………………….
………….../:::::::::::::::::::::::::::::::’\,__„„/’:::::::::::’\………………………………...
…………..|:::::::::::::::::::::::::::::::„----„_„---,::::::::::\………………………...........
…………..|:::::::::::::::::::::::::::„~"----„__„„-“::::::::::|………………………….......
…………..|::::::::::::::::::::::::::/’;;;;;;;;;/::::::::::::::::::|MO' ARMOR!
………..„~”\:::::::::::::::::::::::::\__;;;;;;\--„::::::::::::::/…………………………........
……..,-“;:;\ .\,::::::::::::::::::::::::’\„_„-“:::::::::::::/………………………….........
….„~”;:\;:;:;\, .”~„:::::::::::::::::::::::::::::::::::::,-“;\………………………...........
.,~”;:;:;:;\;:;:;:”-, . ”~--„:::::::::::::~--„„___„„~“::::„-‘ . |;:'\”~-„…………………………
“;:;:;:;:;:;:\;:;:;:;:;”-, . . .”~~---„„„„_______„„„--~” . ./ |;:;:’\;:;”-,……………………...
;:;:;:;:;:;:;:;\____;:;”-, ~-„ . . . . . . . . /;;;;;’\, . . .,/ . |;:,~”;:;:;”-,……………..........
;:;:;:;:;:;:;:;:;:;:;:;/;:;:;:;”~, “~-„_ . . . /„;;;;;;;/”-„~” . . |;:;”~„;;’\……………….....
;:;:;:;:;:;:;:;:;:;:;:/;:;:;:;:;:;:”~„ . .”~-’ . “~,;;;’\ . . . . .|;:;:;:;|;:;’\…………….....
;:;:;:;:;:;:;:;:;:;:;:;”~--„;:;:;:;:;:”~„ . . . . . .|;;;;;;’\ . . . |;:;:;:;/;:”~:\………………
;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:”~--„;:;:;:;:”~„ . . . |;;;;;;;;’\, . .|;:;:;:/;:;:;;’\;:;:
;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:”~-„;:;:;:”~„ .|;;;;;;;;;;;”~,|;:;:;/;:;:;::\;:;:
;:;:;:;:;:;:;:;:;:|;:;:;:;:;:;:;:;:;:;:;:;:;:”~-„;:;:”~„;;;;;;;;;;;;/;:;,/’;:;:;:;::;\;:;:
;:;:;:;:;:;:;:;:;:|,/;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:”~-„;”~„;;;;;;/;:,/;:;:;:;:;;:;:;\;:;:'
;:;:;:;:;:;:;:;:;:/;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;”~-„’\,;/ ,/;:;:;:;:;:;:;:;:\;:;:|…
;:;:;:;:;:;:;:;:;|;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;: ;:;:;:;:;:;:;:;:\„/;:;:;:;:;: ;:;:;:;|;:;:|……….
;:;:;:;:;:;:;:;:;|;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;: ;:;:;:;:;:;:;:;:;:;|;:;:;:;:;:;::::;:|;:;:|....... ....
;:;:;:;:;:;:;:;:;|;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;: ;:;:;:;:;:;:;:;:;:;|;:;:;:::;:;:;:;:;:|;:;:|……….
;:;:;:;:;:;:;:;:;|;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;: ;:;:;:;:;:;:;:;:;:;|;:;:;:;:;:;:;:;:;:|;:;:|……….
_;:;:;:;:;:;:;:;|;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:; :;:;:;:;:;:;:;:;:;|;:;:;:;:;:;:;;:;:;:|;:;:|...... .....
;:;””~~--„„_|_;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:;:; :;:|;:;o;:;:;:;:;:;:;::;:;|-~/…….....
;:;:;:;:;:;:;:;:;:;:””~~----„„„__;:;:;:;:;:;:;:;:;:;:;;:;:;:;|;:;:;:;:;:;:;:;: ;:;:;:;|~“…………


Thanks Ollie.

Feanorr
08-13-2009, 04:14 AM
There is something logically inconsistent between my quick count and your solid analyais, Bashal.

Argument: If I counted nearly 3x as many parry-haste events, and Splug only had about 2% higher avoidance, you would expect to see some noticeable parry-haste. But your data shows no increases in overall haste (in fact, you had the opposite). Where did all the parry-haste go? It couldn't have been dodged, since he only increased avoidance by 2% (which is about 5% relative increase); not enough. Let's argue that 50% of the parry events came in the portion of the boss swing-timer where it doesn't actually speed him up, but if so, that'd still leave a lot of parry-hasted-attacks, which I can't see in your data either.

Problem: I think either (a) one of us is has the data wrong, or (b) my that argument is flawed.

Splug: thanks for clarifying you were Unholy DW tanking. I missed that. I'll ignore any data on its threat output.


Or maybe, the boss just doesnt parry-haste ^^

Bashal
08-13-2009, 09:29 AM
Or maybe, the boss just doesnt parry-haste ^^

Possibly, but from what I could find out, he supposedly still has it.

Can anyone confirm definitively if he has parry haste?

Splug
08-13-2009, 09:35 AM
If he didn't, you should observe zero events where two attacks connect in under 2.0, correct? So the data already throws that option out...

-Splug

GravityDK
08-18-2009, 10:51 AM
I wish we could unravel this mystery. I think about it now and then, and can't work it out.

DwayneDibley
08-19-2009, 02:54 AM
Ok, I'll unravel this mystery for you.... The boss doesn't parry haste.

If you look at the logs, and list all hits on the tank whether hit/miss/dodge or parry, you'll see that the bosses hit speed is approx 1.7 secs, and it never gets any faster than that (except perhaps due to a small amount of measurement latency)

Here is an example of the logs i produced, from 2H:
(First column is time, 2nd is the event, 3rd is the difference in time between the current attack and the next one)

59:22.97 General Vezax's hit Spyte for 20597 damage. 00:01.707
59:24.68 General Vezax's hit Spyte for 18805 damage. 00:01.705
59:26.38 General Vezax's was parried by Spyte. 00:01.713
59:28.09 General Vezax's hit Spyte for 14463 damage. 00:01.801
59:29.90 General Vezax's hit Spyte for 16646 damage. 00:01.648
59:31.54 General Vezax's was dodged by Spyte. 00:01.729
59:33.27 General Vezax's hit Spyte for 14943 damage. 00:03.454
59:36.73 General Vezax's hit Spyte for 14023 damage. 00:01.700
59:38.43 General Vezax's missed Spyte. 00:01.726
59:40.15 General Vezax's was dodged by Spyte. 00:01.770
59:41.92 General Vezax's hit Spyte for 14068 damage. 00:01.664
59:43.59 General Vezax's missed Spyte. 00:04.393
59:47.98 General Vezax's hit Spyte for 16069 damage. 00:01.712
59:49.69 General Vezax's was dodged by Spyte. 00:03.139
59:52.83 General Vezax's hit Spyte for 17393 damage. 00:02.383
59:55.21 General Vezax's hit Spyte for 16104 damage. 00:01.706
59:56.92 General Vezax's hit Spyte for 17091 damage. 00:01.725
59:58.64 General Vezax's hit Spyte for 17416 damage. 00:01.732
00:00.38 General Vezax's hit Spyte for 17782 damage. 00:01.687
00:02.06 General Vezax's was dodged by Spyte. 00:01.721

And From DW:

12:33.60 General Vezax's was parried by Spyte. 00:01.609
12:35.21 General Vezax's was parried by Spyte. 00:01.705
12:36.91 General Vezax's hit Spyte for 13633 damage. 00:01.731
12:38.64 General Vezax's hit Spyte for 15184 damage. 00:01.717
12:40.36 General Vezax's was dodged by Spyte. 00:01.686
12:42.04 General Vezax's was dodged by Spyte. 00:01.729
12:43.77 General Vezax's was parried by Spyte. 00:01.691
12:45.47 General Vezax's was parried by Spyte. 00:04.208
12:49.67 General Vezax's was parried by Spyte. 00:01.716
12:51.39 General Vezax's was dodged by Spyte. 00:03.106
12:54.50 General Vezax's was parried by Spyte. 00:02.411
12:56.91 General Vezax's was dodged by Spyte. 00:01.699
12:58.61 General Vezax's missed Spyte. 00:04.353
13:02.96 General Vezax's missed Spyte. 00:01.710
13:04.67 General Vezax's was dodged by Spyte. 00:01.723
13:06.39 General Vezax's was parried by Spyte. 00:04.208
13:10.60 General Vezax's was dodged by Spyte. 00:01.680
13:12.28 General Vezax's hit Spyte for 9478 damage. 00:01.715
13:13.99 General Vezax's hit Spyte for 12155 damage. 00:01.789
13:15.78 General Vezax's was parried by Spyte. 00:04.442

Now from Dw, with the first 3 parries from the tank merged in:

12:33.60 General Vezax's was parried by Spyte. 00:01.61
12:35.21 General Vezax's was parried by Spyte. 00:01.71
12:36.91 General Vezax's hit Spyte for 13633 damage. 00:01.73
12:38.13 **Spyte's was parried by General Vezax.
12:38.64 General Vezax's hit Spyte for 15184 damage. 00:01.72
12:40.32 **Spyte's Blood Strike(R6) was parried by General Vezax.
12:40.36 General Vezax's was dodged by Spyte. 00:01.69
12:42.04 General Vezax's was dodged by Spyte. 00:01.73
12:43.77 General Vezax's was parried by Spyte. 00:01.69
12:45.47 General Vezax's was parried by Spyte. 00:03.89
12:49.35 **Spyte's was parried by General Vezax.
12:49.67 General Vezax's was parried by Spyte. 00:04.21
12:51.39 General Vezax's was dodged by Spyte. 00:03.11
12:54.50 General Vezax's was parried by Spyte. 00:02.41
12:56.91 General Vezax's was dodged by Spyte. 00:01.70

My Hypothesis is as follows:
The boss has a basic 1.7 speed swing.
Any small variation in swing speeds is down to latency.
For a parry to occur, you'd expect up to a 40% increase on swing time, which means you'd expect some swings as fast as 1.02 seconds... There are no swings any faster than 1.5secs in the entire fight (which can be attributed to recording latency).
Any durations between swings much greater than 1.7 secs is due to boss movement or boss cast times.

Agreed? I have the logs I can post on request.

DwayneDibley
08-19-2009, 05:36 AM
Further info:

In the DW log there was a potentially catastrophic event - the tank caused the boss to parry 3 times in 0.7secs:

21:11.1 Spyte's was parried by General Vezax.
21:11.3 Spyte's Death Strike(R5) was parried by General Vezax.
21:11.8 Spyte's was parried by General Vezax.

If the boss did parry haste, the tank likely wouldn't haved live to tell the tale.

Another bad section is here (it shows no appreciable parry hasting, despite 6 parries in 21 seconds):
17:53.40General Vezax's was parried by Spyte. 00:01.812
17:54.03Spyte's was parried by General Vezax.
17:55.21General Vezax's hit Spyte for 18217 damage. 00:03.102
17:57.66Spyte's Blood Strike(R6) was parried by General Vezax.
17:58.31General Vezax's hit Spyte for 22455 damage. 00:01.648
17:59.96General Vezax's was dodged by Spyte. 00:01.581
18:00.90Spyte's Death Strike(R5) was parried by General Vezax.
18:01.54General Vezax's was parried by Spyte. 00:01.560
18:01.54Spyte's was parried by General Vezax.
18:03.10General Vezax's was dodged by Spyte. 00:01.671
18:07.94General Vezax's hit Spyte for 16375 damage. 00:01.680
18:08.12Spyte's was parried by General Vezax.
18:09.62General Vezax's was dodged by Spyte. 00:01.708
18:11.32General Vezax's was dodged by Spyte. 00:01.759
18:13.08General Vezax's hit Spyte for 13732 damage. 00:01.717
18:14.39Spyte's was parried by General Vezax.
18:14.80General Vezax's was dodged by Spyte. 00:05.602

Satorri
08-19-2009, 06:16 AM
Small problem with the above though, in the first post comparing incoming swings to parries, all of the parries listed happened in the last 20% of the swing so no parry haste could occur if it is possible. That data won't settle it.

In the second to last portion, 3 parry-hastes in 0.7 seconds could still only bump that swing up 0.68 sec sooner on a 1.7 sec swing timer. It would only be one faster swing.

In the final colorful piece, notice he parried/dodged 7 of 11 incoming strikes and the typical strike hit for (we'll average up to be conservative) ~18k. Of the (statistically completely implausible) 6 parries in 20 seconds, one of the 6 was not in a swing timer range that would generate haste anyway. So, the other 5, in total, would've sped up his attacks by a total of ~1.88 sec total, so he would've squeezed in one additional hit in that 20 sec. With this size hit, for a tank on Vezax (40k+ health is a safe assumption), the only way that would've turned deadly is parry #2 or #3 was actually followed by a hit instead of a parry/dodge, and there was a convenient 5 second hole where the tank wasn't healed.

Hyperbole doesn't help your case when it's checked for accuracy, and I think you misunderstand exactly how parry haste works.

Fetzie
08-19-2009, 06:27 AM
Vezax does parry haste, although I no longer have the log to prove it. I have taken two melee attacks from him within 0.9 seconds of each other (I checked the death report after every fight to see why I died. This is coherent with a 40% parry haste on the second swing. This was on 25 man difficulty (although before 3.2, I havent tanked him since the patch, what with extended raid locks in place to at last kill yogg)

DwayneDibley
08-19-2009, 06:35 AM
You are right about the first post satorri, I had looked at the timing of the data I posted and decided to post the second.

This swing for example should *DEFINATELY* exhibit haste, yet it doesnt.

18:07.94General Vezax's hit Spyte for 16375 damage. 00:01.680
18:08.12Spyte's was parried by General Vezax.
18:09.62General Vezax's was dodged by Spyte. 00:01.708

If the fact that not a single swing faster than 1.52 seconds is observed in 9 minutes isn't proof enough. (and the fact its 1.52 seconds at all is likely down to measurement inaccuracies - the average parry hasted attack should see 24% haste - hence be 1.292 secs, and there are no strikes of this order, even though there were 45 parries.)

lyd
08-19-2009, 08:29 AM
good info Dwayne. what's interesting about the last log is that Spyte avoided 4/6 of the hits that followed each parry. If those sort of avoidance returns could be shown consistently, then it follows that, in most cases of a "potentially catastrophic" event, i.e. a triple parry chain, the DK in this parse would could reasonably be expected to avoid 60+% of the hits.

However, we need not even consider avoidance in this particular case, because I think that the "catastrophic event" in question may actually a false demon, depending on the swing speed of the boss. something in the nature of parry-haste mitigates the danger of this event: the time span in which it occurs. that chain takes place over the course of .7 seconds. let's assume a boss swing speed of 1.7 seconds, as your data suggests. And while it is highly unlikely that they would perfectly align like this, let's also assume the first parry occurs at the beginning of the boss's swing timer:

i'll copy the part of the event log that i'm interested in here for quick reference:



21:11.1 Spyte's was parried by General Vezax.
21:11.3 Spyte's Death Strike(R5) was parried by General Vezax.
21:11.8 Spyte's was parried by General Vezax.k, now let's paste the parry rule for quick reference:

When you parry an attack, the remaining time on your current swing is reduced by 40% of your weapon speed, unless this would result in a reduction to less than 20% of your swing time remaining.

parries occur at 0.0, 0.2 seconds , and 0.7 seconds respectively.

40% of 1.7 = .68 (we'll call this the haste coefficient)

anytime a parry haste brings the swing speed into the dead zone, i will write *stop* to denote a haste failure:

1st parry occurs: 1.7 > 1.02
2nd parry occurs .2 seconds later: .82 > .14 *STOP*

.14/.82 = 17%. this is less than 20% of the remaining swing-time, so the second parry has no effect on the boss's haste. so we stay at .82.

3rd parry occurs .5 seconds later: .32 > (negative number, N/A)

this would result in a negative number, so it obviously is less than 20% of the swing time remaining and will have no haste effect on the remaining swing time. so in this particular parry chain, if the boss was parry-hasting, this chain would have only produced 1 parry-haste effect.

Now is 1.7 swing speed typical? let's do the same analysis at 2.0 and 2.4:

40% of 2.0 = .8

1st parry occurs: 2.0 > 1.2
2nd parry occurs: 1.0 > .2 (this is EXACTLY 20%. so according to our model, this attack does get hasted again. the next parry occurs .5 seconds later, which is .3 seconds into the next swing)
3rd parry occurs: 1.7 > .9

so at 2.0, these same three parries will produce 2 separate hasted attacks instead of 1.

now for 2.4:

40% of 2.4 = .96

1st parry occurs: 2.4 > 1.44
2nd parry occurs: 1.24 > .28 (.28/1.24 = 22.5%, so this attack does get hasted again. next parry occurs .5 seconds later, which is .22 seconds into the next swing)
3rd parry occurs: 2.18 > 1.22

so in this example as well, two separate attacks are hasted.

end result of this chain (counting the hit just prior to the swing timer resetting):

@ 1.7 swing speed: 3 parries caused 1 hasted attack, removing a total of .68 seconds off of the boss's swing time, resulting in 3 hits in 2.72 seconds.

@ 2.0 swing speed: 3 parries caused 2 haste effects on 1 attack and 1 haste effect on another, resulting in 2nd attack .4 seconds after first reset, and 3rd attack 1.4 seconds after second reset, for a total of 3 attacks in 1.8 seconds (gah!)

@2.4 swing speed: 3 parries caused 2 haste effects on 1 attack and 1 haste effect on another, resulting in 2nd attack .48 seconds after first reset, and 3rd attack 1.72 seconds after 2nd reset, for a total of 3 attacks in 2.2 seconds (slightly less gah!)


Things to consider:

-the likelihood of this parry chain, which is in of itself unlikely, occuring in sync with a boss's swing right at the beginning of his reset is highly improbable.

-a more realistic example would be using a middle ground, such as running the 2.0 calculation with 2.4 haste coefficent (.96). which would change things considerably:

1st parry occurs: 2.0 > 1.02
2nd parry occurs: 1.02 > .06 *STOP*

if the parry chain started at 2.0 on a boss with a 2.4 swing speed, the 2nd and 3rd parries in this chain would occur inside the "dead zone" (1.02 and .52) where the haste coefficient would reduce it to less than 20% of the remaining swing time, thereby having no additional effect. in fact, with a 2.4 swing speed, any parry chain of this nature occuring any time between roughly 2.3 and 1.6~7 seconds during the swing timer of a boss with 2.4 swing speed would still only result in 1 parry-hasted attack:

1st parry occurs: 2.3 > 1.34
2nd parry occurs: 1.14 > .18 *STOP*
3rd parry occurs: .64 > (negative number) *STOP*

1st parry occurs: 1.67 > .7~1
2nd parry occurs: .5~1 > (negative number) *STOP*
3rd parry occurs: .0~1 (negative number) *STOP*

-in the initial 2.0 example, a miniscule adjustment to the beginning of the swing timer would cause the 2nd parry to occur inside of the 20% marker, causing this example to lose it's volatility, with only 1 hasted attack effectively occuring. given the statistical unlikelihood of ever being EXACTLY at the start of the swing timer at the time of a parry (i don't know how precise Blizzard's combat generator is so this may be more or less so), I think the 2nd example is, in fact, unlikely, because in the event that a triple-parry chain does occur, it is infinitely more likely that this will not be the case than that it will be.

-in order to gain a true appreciation for the actual danger this puts the tank in, you must factor in avoidance, the chance of a physical mitigation cooldown being active at the time of the chain, and the chance that an absorption effect of some sort exists on the tank to mitigate the size of at least one of the hits. since we're talking about the danger of a wipe-causing chain in which heals don't have time to land, all of these things must be taken into consideration.

more importantly, i want to point out just how unlikely this event is, and how well Splug actually ran the odds by actually procc'ing a triple parry chain:

probability of a 3x parry chain @ 26 expertise (14% boss parry according to Satrina's numbers: .075*.075*.075 = 0.000421875, or .042%. In other words, you will have to attack a boss with any sort of melee attack roughly 2468 times in order for this even to be statistically guaranteed to occur once. at a swing speed of 1.2 (fast/fast, hasted), without RS replacement, you would attack 100/minute w 2 weapons. a melee attack every 1.5 second GCD would give you an additional 90 attacks/minute, for a total of 190 melee attacks/minute. so not replacing for spell use (howling blast, death coil, etc), and not accounting for Rune Strike replacement, this event will happen roughly once for every 13 minutes of boss fight time. i'm going to ballpark and say that with RS replacement and spell replacement (IT, HB, DC, etc), this would increase significantly, probably upward in the 16-20 minute range. I'd run some math based on common percentages, but i don't have the time as of right now, unfortunately : \

now if the event occurs on a boss with a 2.4 swing time, who is between 2.3 and 1.67 on their swing timer, it effectively only grants one hasted attack, as shown above. no worse than if only 1 or 2 had been parried. i imagine a similar deadzone can be shown for the 1.7 speed.

after this, we will now factor in the possibility of avoidance. call it 50%+. so even if this hasted effect doesn't fall into any of the deadzones described above, there's a better chance than not that one of the attacks will be avoided. so there are 2 factors right there working against this scenario.

a third factor was mentioned before, which is the possibility of cooldown mitigation being present during this parry-chain. if IBF or Bone Armor was mitigating one or all of these hits, then it's not even an issue, because the relative size of the hits, even if they didn't fall into a deadzone, and even if at least one of them wasn't avoided entirely, would still be mitigated heavily by cooldowns, making it extremely unlikely that a non-dead-zone, unavoided chain like this would actually ever down a tank. we're basically multiplying .042%*50%*x, where x is equal to the chance of a mitigation cooldown not being present during the time the chain occurs. this number is very, very small.

another thing remains to be observed: What is blizzard's standard for removing parry-haste on a boss? If these logs are accurate, the data implies that Vazax does not parry-haste. so we now have two hard-hitting bosses off the top of my head, Vazax and Patchwerk, that don't seem to exhibit any parry-haste. the question is this: does blizzard remove the parry-haste mechanic on bosses who hit hard and fast enough to make this mechanic a liability?

more testing is most definitely needed :P but the catastrophic event observed requires what i believe can be shown to be a statistically improbable chain of events to occur which would actually put the tank in danger of being 2-3 shot without a chance for heals to go off. in other words, if you wiped to this event, it's because God wanted you to.

Splug
08-19-2009, 10:11 AM
It's probably worth mentioning that I used weapons with synchronized swing speed. The only way I'd have two autoattacks .7 seconds apart would be if I had parried an incoming attack. While that does increase the risk of multiple swings in a small window, the premise it was caused under indicates that at least one of the swings was avoided, and thus it's still a surge of less-than-normal swing time delay.

It's possible to have both weapons land directly on top of each other, which would result in essentially ensuring the next incoming swing goes into the parry dead zone. At absolute worst case, it's the equivalent of 1.5x the normal speed increase for a single parry.

I'd read that Vezax was capable of a parry-hasting effect, but that was second hand information. This seems to point toward that being incorrect. It's more likely that the occasional ~20% faster attack is a result of downtime in icy touch.

I can dredge up some records for Algalon if we want to look at another fight with heavy tank damage. I don't have any data on 2h-tanking him though, just dw frost. Then again, maybe he doesn't parry haste either.

-Splug

Satorri
08-19-2009, 10:20 AM
I'm fairly certain Algalon does not. He's a dual wield style boss, who are traditionally parry-haste-disabled (Patchwerk, Mother, etc).

GravityDK
08-19-2009, 10:47 AM
Ok, I'll unravel this mystery for you.... The boss doesn't parry haste.
...

My Hypothesis is as follows:
The boss has a basic 1.7 speed swing.
Any small variation in swing speeds is down to latency.
For a parry to occur, you'd expect up to a 40% increase on swing time, which means you'd expect some swings as fast as 1.02 seconds... There are no swings any faster than 1.5secs in the entire fight (which can be attributed to recording latency).
Any durations between swings much greater than 1.7 secs is due to boss movement or boss cast times.



I think that's it. Gets my vote.
Perhaps parry-haste was removed from this boss after Pyrea observed it killing him, in a silent hotfix.

Fascinating. I think we went from trying to work out DW incoming pain to discovering something about Vezax :).

We need more logs on slow bosses now to explore Lyd's thinking :). XT002 is probably the easiest, since its tank/spank.

Satorri
08-19-2009, 11:12 AM
*coughhardmodecough*

lyd
08-19-2009, 11:40 AM
It's possible to have both weapons land directly on top of each other, which would result in essentially ensuring the next incoming swing goes into the parry dead zone. At absolute worst case, it's the equivalent of 1.5x the normal speed increase for a single parry.

I'd read that Vezax was capable of a parry-hasting effect, but that was second hand information. This seems to point toward that being incorrect. It's more likely that the occasional ~20% faster attack is a result of downtime in icy touch.

love these observations. great thinking splug. the first observation implies that weapons with sync'd weapon speeds diminish parry haste by tending to force a dead zone push. The second one helps explain why the data samples have a speed variance. However, matched weapon speeds also seem to like to desync over the course of a fight. anyone have any idea on how likely it is for weapon strikes to fall out of sync on weapons with matched speeds? I'm thinking it's probably very likely, honestly. When I raided with my fury warrior in TBC it seemed like it was practically guaranteed that my white hits would desync over the course of the fight...

DwayneDibley
08-19-2009, 12:53 PM
Yeah, fair enough, 3 parries in 0.7 seconds doesn't happen to be quiet as catastrophic as you may think it would be.
It does show how frequent parries may end up being, and how they could potentially change the outcome of a fight, if they happen to synch in a particularly arkward way.

In one 9min fight, the improbable has happened twice. 3 parries in 0.7 secs and 6 parries in 21. This clearly shows that sh1t does indeed happen ;-) And the more chance you give it, the more it takes!

As a tank I need to be considered a 'safe pair of hands' but the potential for DW to cause gibbing worries me, and is why I personally will steer well clear of it.

GravityDK
08-19-2009, 01:04 PM
I do like how it looks though. :)
I DW for dps.

lyd
08-20-2009, 06:18 AM
In one 9min fight, the improbable has happened twice. 3 parries in 0.7 secs and 6 parries in 21. This clearly shows that sh1t does indeed happen ;-) And the more chance you give it, the more it takes!that's a fair statement. but remember: it's all about combining probabilities to predict the potential of a wipe-causing string of attacks. what are the chances that the tank has 6 parries in 21 seconds AND fails to avoid at least half of the hasted attacks with better than 50% avoidance? glancing at the data, my point is made manifest: out of 11 total boss hits in the sample 21 second period, Spyte avoided 7 of them. the only eye-brow raising event occurs after the first parry, in which he takes 40.6+ k damage in 4.28 seconds:

17:54.03Spyte's was parried by General Vezax.
17:55.21General Vezax's hit Spyte for 18217 damage. 00:03.102
17:57.66Spyte's Blood Strike(R6) was parried by General Vezax.
17:58.31General Vezax's hit Spyte for 22455 damage. 00:01.648

now there's no appreciable parry haste here. so the boss appears to have been .5 seconds into his swing timer at the time of the first parry. assuming the boss did parry haste, it would have made the next swing fall at 17:54.55, thus resetting the swing timer. another attack would've occurred at 17:56.25. at the time of the second parry, 17:57.66, the boss would have been 1.41 seconds into his swing timer, so no parry haste would occur for the second attack, which would then fall at 17:57.95. so what we end up with is 40.6k damage in 3.40 seconds instead of 4.28, still a more than reasonable amount of time to expect heals to reach the tank before the last strike falls.

I think the best way to look at it is that there are a series of failsafes built into the mechanics of the game to prevent a real parry-gib from ever occurring in all but the "act of god" genre of circumstances, assuming human error isn't skewing the math. triple or better parry strings are already an unlikely occurrence, but definitely not so unlikely that they can't be guranteed to occur. let's assume it'll happen once every other boss fight. once the event has run its own existential probability tables, it has to result in more than 1 parry haste to be dangerous. to do this, they have to tend with dead-zone math. If the event rolls past the dead-zone math, it now has to make it past 50+% avoidance tables. if it makes it past the avoidance tables, it now has to make it past the mitigation cooldown tables. If it makes it past the mitigation cooldown tables, it now has to make it past the hps RNG, assuming a heal may or may not be landing at the right time. Once it makes it past the hps RNG, the tank is now in danger.

the data shows that the "unlikely" can and does occur. But it also shows that the additional probability matrices for events that are contingent to the "unlikely" can be reasonably depended on to keep it from actually putting the tank in danger of a wipe in a consistent fashion.

Splug
08-20-2009, 09:02 AM
I think a better thing to walk away from it with is that Blizzard turns off parry haste on bosses where tank damage is a major focus. It could be a concern on fights where boss melee is not the encounter's focus, but at that point... is it really a threat?

-Splug

lyd
08-20-2009, 09:08 AM
that is a legitimate induction as well Splug, and the fact that they do it at all implies they recognize a possible RNG threat that is possibly unfair to the players, which would also seem to imply that most, if not all hard-hitting, fast bosses would have these mechanics removed to prevent RNG from trumping execution and player skill. it wouldn't make any sense otherwise from a game design perspective to remove parry haste from some hard-hitting bosses but not others. Blizzard's design process is many things, but it is anything but arbitrary.

DwayneDibley
08-21-2009, 12:53 AM
I think a better thing to walk away from it with is that Blizzard turns off parry haste on bosses where tank damage is a major focus. It could be a concern on fights where boss melee is not the encounter's focus, but at that point... is it really a threat?
-Splug

That does seem to be a reasonable idea.
I have however been parry gibbed a few times on auriaya for example - in this case there is often quite a bit of raid damage going on, which may cause a healer to take his eye off a tank for just a GCD, which when combined with a non avoided parry hasted attack, can, and has been fatal.

If you look at it this way, yes it can be considered a threat. When keeping the tank up is not the entire focus of the healing team, a sudden burst of tank damage when the healers are up to something else can very easily be fatal.
Perhaps if your healers know that there is an increased chance for these burst events to happen, they can mitigate it though with increased focus or rigid healing assignment.

Satorri
08-21-2009, 06:51 AM
Truth be told, I don't think healers would notice the difference. Consider the way things are:

1.) No crushing blows
2.) High avoidance
3.) High boss damage

Any tank, regardless of weapon style, will take big hits and often it is only a matter of how your avoidance lines up that will determine how many hits get clustered together and how often. Dual wielding will present a degree of increase in incoming swings (see: my post second on the page (http://www.tankspot.com/forums/f14/54264-deathknight-dw-tanking-discussion-2.html)), but relative to other tanks, and heavily depending on the amount of expertise you have, the amount will be no more remarkable than the variance caused by avoidance.

Something to keep in mind in *thinking* about how it would actually played out, the log snippets above are really worth looking at. With a tank avoiding 60% of incoming strikes, you can essentially wipe half the extra swings taken off the map for consideration since they will do no damage.

If you are dual wielding fast/fast for survival weapons, and have the normal 26 expertise any other tank would aim for (i.e. no stretching or extra gemming/chanting), you could expect 4 more swings at you per minute on average, than wielding your 2-h, and if we factor the avoidance, that's maybe 2 more hits on you per minute. Timing is always an issue, but I don't expect that the increase in damage would actually be noticed, or require special preparation.


It is always in our interest to reduce the damage we take, but it doesn't have to be min/max'd to the absolute least damage taken. Every tank trades some survival for threat in several of their choices in spec and gear, this is just one more choice in that vein, and not even as obvious. It could, as Lyd pointed out, not be a loss in survival at all, as a pair of tanking weapons could conceivably boost your survival to reduce incoming damage beyond the increase caused by parry haste (when you can even be parry-hasted). This is a big part of why I've been trying to regulate these discussions with data.

Parry-haste as a liability for a DK tank is pretty over-exaggerated, frequently by people who don't actually appreciate what it means. They have a different idea of it in their head that has the tank taking MASSIVE "spikes" of damage.

Zephyre
08-21-2009, 10:33 AM
I just thought I'd weigh in on this one. I ditched my warrior about 3 months back after rolling my DK. I've been Frost tank specced since lvl 55.

With the changes introduced with 3.2 I realized it'd be foolish of me not to at least give DWFrost a shot. So I did. Granted my weapons are not the greatest (their too fast) I'm still producing around 4500 tps with an oblit spam rotation.

Thank you Satorri for illuminating the benifits of Glyph of Disease in your Blood discussion. I've modified it to fit my particular play style. After IT and PS have been used I settle into a Pest, BS, OB, OB rotation dropping FS and HBs when RP and procs occur.

I'm actually showing better TPS than I did with my old 2H frost build.
This is my new Frost build for DW.
The World of Warcraft Armory (http://www.wowarmory.com/character-talents.xml?r=Uldum&n=Austru&group=2)

lyd
08-26-2009, 08:29 AM
If you drop acclimation and take 2 points out of subversion, then max out Bladed Armor, i think you'll see a poignant dps/tps increase.

i used to be a huge fan of acclimation until someone pointed out to me that acclimation is only effective if you're taking fast, regular bursts of magic damage, and there are many fights where that's not the case, meaning those points are more or less ineffective. on the same note, the additional threat and damage is 100% effective on all boss fights, regardless of magic damage or frequency.

Satorri
08-26-2009, 11:47 AM
That's half true, Lyd. It does apply 50 resistance to most hits at least, which is nothing to scoff at (~5% mitigation and bumped partial resist chances, roughly). On fights where it is consistently stacking Acclimation the value jumps up pretty noticeably.

That said, I don't take it the way I haven't taken Spell Deflection. It's great but I haven't been focusing on covering my butt from magic so much as I've been greedily stacking threat talents. Do not make the mistake of thinking that it's not worth doing though. People generally under-appreciate the amount and prevalence of magic damage out there.

(Oh, but also listen to Lyd, Bladed Armor is your single best generic threat talent as a DK tank, loads of AP).

Splug
08-26-2009, 12:06 PM
If the span between magic attacks of the same element is greater than the duration of acclimation, it does exactly nothing. It's not a matter of luck with stacking, it's just that the effect occurs in a timeframe where it is not useful.

It is strong in specific environments, but it is very situational and at a location in the talent tree that is difficult to fill out, especially for dual wield builds.

-Splug

Edgewalker
08-26-2009, 10:55 PM
On a totally unrelated note, Spell Deflection is a very under appreciated talent :(

jaydee
08-26-2009, 11:27 PM
It's the direct damage part of the talent that makes it not so attractive. If only they would remove that part. But then again, that might seem too much, given that tanks usually have ~20% parry.

Satorri
08-27-2009, 08:29 AM
Unless I'm mistaken Splug, the Resistance applies to the hit that procs it as well. Take that and add that the duration is 18 seconds and it gets more love than you might think.

The funny thing is that Spell Deflection and Acclimation are just about on par with Magic Suppression (though the 100% from AMS is kind of epic), but they're not often as well appreciated.

Also, Jaydee, this was something that took me time to appreciate. Direct Damage here is as opposed to periodic damage, not AoE or Cone type attacks. It will work against non-targeted dragon's breath, for example. It's actually nicer than you think it just won't protect you from the odd DoT or Aura.

Edgewalker
08-27-2009, 09:42 AM
I like spell deflection over Acclimation for that reason generally. Acclimation helps mitigate constant low magical pulses in encounters, but is pretty poor generally against the heavier strikes with cooldowns because it simply isn't stacked high enough in most situations. Spell Deflection protects against the spike damage of single target spells, which is generally worse for a healer.

lyd
08-27-2009, 09:57 AM
I agree with you on Spell Deflection Edgewalker. I think it's a wonderful talent and shouldn't be passed up in a blood tanking build. I would definitely prefer it to acclimation if I didn't prefer frost tanking in general to Blood tanking. What's nice about it is that it's efficacy grows as your parry rating increases, so you're not only increasing melee avoidance, but marginally increasing magic damage mitigation.

Satorri
08-27-2009, 10:09 AM
Acclimation is actually sexay when used well against stuff like Hodir and Steelbreaker. The aura from Steelbreaker and the ample opportunity for Frost damage in Hodir means you can keep a 3 stack up all the time. 150 Frost resist on top of a couple pieces of Frost resist gear and a totem is another 15% reduction of Frost damage, or if you use it instead of Frost Resist gear it will be worth more like 20-25%.

Against Steelbreaker it'll be a good 20-25% reduction against the aura and his Fusion Punch, which is delectable when it matters. Plus for those fights it's always on.

It's a funny nuance in the balance.

Magic Suppression = 6% magic reduction, period. Plus AMS is immunity unless you're soaking more than half your health (not often, not a bad value). This was once laughably overpowered before they added the health limiter, I used to use it to be immune to massive kill-shots of magic. 5 second magic damage immunity, /sigh good times.

Spell Deflection = 45% reduction ~20% of the time (average parry chance). A 20% parry chance averages out to a 9% magic damage reduction over all non-periodic magic damage ticks (good against Fusion Punches, nothing against his Aura). A 22% parry chance, for illustration, is a 9.9% average reduction. The big thing is you'll see this in spikes instead of constant low level reduction. This can be game-winning against massive hits like Fusion Punch/Frozen Blows (never tested it, does Plasma Blast count as periodic?).

Acclimation = ~4-7% reduction against random spell damage of any kind depending on resistance already on from aura's and totems. Scales up to 15-25% reduction on every magic effect when there is a lot to absorb. It scales well.

Everybody gets their tools, it's all good, just different flavors.

Gendrake
08-27-2009, 10:58 AM
The biggest problem with Acclimation that I see is its placement in the tree. We have too many important tanking talents in that region of the tree, particularly in a DW spec. It makes it very hard to fit it in without losing out on a lot of really good threat talents in blood or unholy.

Here's hoping the pruning of the trees in Cataclysm leaves us with an easier route to taking it.

lyd
08-27-2009, 11:10 AM
that was the first thing I noticed when ToT was introduced Gen. I knew I was going to have to choose between the new SoB and acclimation. I decided on SoB, since Acclimation is a victim of being situational, whereas SoB is just as effective on any boss no matter what. DW also gets better returns off SoB than 2h, making it that much more attractive to me.

Satorri
08-28-2009, 06:39 AM
All that I ever championed Magic damage reduction (mostly back when I was all about Unholy), I have made the same choices, Lyd. =(

I think you're right Gen. The talents can only be taken at the expense of other very central stuff you want, though you could say the same looking from a different angle. If that were not the case, conceivably tanks would complain that they have to take the talent (situational some complain) instead of taking something they want more but comes in a different place.

Warriors, Pallies, and Bears have reached a point in their talent trees where they just sort of get the 6% magic damage reduction (boring by comparison, right? =D) with other stuff they just normally get otherwise.

I'm hopeful too. I like this new paradigm with talent trees of losing the passive buffs everywhere and make talents more about interesting flavor and play style changes. It could lead to talents being more fun.

GravityDK
09-22-2009, 03:07 PM
I ran a simulation (http://pwnwear.com/2009/09/22/3-22-death-knight-dual-wield-tanking/). Result: did good threat, slightly more than 2Her!

I'm keen to hear how it goes on live. With 30 expertise skill, DW DK tanking looks very promising.

@Lyd, I also updated the dk.info sticky you'd mentioned.

Malygor
09-23-2009, 09:02 AM
I ran a simulation (http://pwnwear.com/2009/09/22/3-22-death-knight-dual-wield-tanking/). Result: did good threat, slightly more than 2Her!

I'm keen to hear how it goes on live. With 30 expertise skill, DW DK tanking looks very promising.

@Lyd, I also updated the dk.info sticky you'd mentioned.

Looking forward to testing it out tonight. I will test a couple of specs for it. Just a question: what weapons are you using?

lyd
09-23-2009, 09:40 AM
the RS change, as I'm sure you can imagine, is very satisfying. according to Omen3 I was pushing 8.5k threat last night in ONY 10 without any trouble.

Ferfey
09-23-2009, 06:13 PM
I took a look at your build on your blog gravity and am wondering, with the recent change to subversion, if taking 2 points from epidemic and 1 from black ice to 3/3 in subversion would be a net increase in threat. With a glyphed howling blast I'm not sure epidemic is needed.

GravityDK
09-24-2009, 06:45 AM
Interesting question Ferfey.

I'd simulate it but sadly Kahorie's sim is not so good when you exclude Epidemic (details (http://elitistjerks.com/f72/t50274-kahories_dk_simulator_beta_1_0_a/p19/#post1398083)). I'd have to use a rotation file instead, which would mean I don't have a like-for-like comparison anymore.

With shorter disease duration, you would have the problem though, of blood plague dropping off in the 2nd half of your standard rotation, and that would mean obliterate would do less damage then. I wonder too how it'd pan out but my gut-feeling is it'd be a net loss.

lyd
09-24-2009, 07:35 AM
I can't speak from a tank perspective on this question because I haven't done this particular comparison, but back when I was running with a 15/51/5 dps build for DW frost, i decided one day to pull 2 points out of conviction and grab epidemic, and i saw an average dps gain of about 100 points. I am unsure if the same would remain true for subversion from a tanking perspective. I don't currently tank with epidemic because my current rotation adjusts for the shorter disease up-time and my threat is ample. My suggestion, as it usually is with these types of things, is to test it yourself and see if you notice a positive change.

Satorri
09-24-2009, 08:44 AM
Disease duration is a hard value to track for its benefits to your dps/threat, but it is very real, and becomes more valuable in practical application than it does in ideal number crunching.

GravityDK
09-24-2009, 01:09 PM
Looking forward to testing it out tonight. I will test a couple of specs for it. Just a question: what weapons are you using?

I didn't use real weapons in my simulation; just entered their speed and dps.