Elitist Jerks
Register
Blogs
Forums


Go Back   Elitist Jerks » Class Mechanics » Death Knights

Closed Thread
 
LinkBack Thread Tools
Old 06/25/09, 3:18 PM   #16
AtheistGod
Piston Honda
 
AtheistGod's Avatar
 
Orc Death Knight
 
Zul'Jin
I did testing with a blood build of 23/48/0
Because of the high number of death runes available in this build the 15 second disease timer isn't that detrimental as you can refresh diseases almost any time you wish.

I did 4 tests. All with my current gear with the following changes

T7.5 Legs instead of t8.5
Fury of the Five Flights instead of grim toll
Broken Promise with Fallen crusader in Main Hand
Angry Dread with Cinder Glacier in Off Hand

Stats:
1491 STR
4048 AP
174 Hit Rating
138 haste Rating
39.95% ArP
29.07% melee crit
20.36% spell crit
5 expertise

All 4 tests were in Unholy Presence.

I Did not use ghoul or any form of cooldowns in any of the tests

The first test was 3.5 minutes last 3 were 5 minutes each.
Ranges are the lowest number from any test to highest both of the other 2 tests were in between.

All 4 of my tests had very similar damage breakdowns
Blood Strike 36.9%-39.9%
Frost Strike 24.7%-25.2%
Melee 14.7%-16.2%
Obliterate 10.7%-11.7%
Icy Touch 2.6%-3.1%
Both diseases 2.4%-2.8%
Plague Strike 1.8-2.2%

Blood Strike Damage ranges
Hit: 54.5%-61.8%
Min: 451-500
Avg: 1302-1404
Max: 2061-2142
Crit: 32.4%-41.6%
Min: 1133-1689
Avg: 3385-3445
Max: 4926-5242
Dodge: 4.0%-5.9%

Frost Strike Damage ranges
Hit 50.7%-54.1%
Min: 644-710
Avg: 1154-1225
Max: 1734-2007
Crit: 39.8%-44.9%
Min: 1649-1783
Avg: 2825-3034
Max: 4203-4864
Dodge: 2.9%-6.3%

From this Nerves of Cold Steel definitely works on all attacks as my 5.31% hit chance didn't see any missed specials
Blood Strike is a better investment of a Death Rune than IT. IT averaged about 1500 for me (counting both crits and normal hits) With Blood Strike hitting hitting harder than Frost Strike the half a Frost Strike added onto IT doesn't make up for the Blood Strike when main hands are only considered, let alone with offhand hits.

I believe it is very likely that strikes proc Cinder Glacier. I had Cinder Glacier procs occur much more than Killing Machine procs with CG procs occurring at least once and often twice during one 15 second disease interval.

With Fury of the Five Flights off hand attacks only stacked it when they don't hit at the same time as the main hand. Strikes only added one stack.

With Grim Toll and Mirror of Truth I used [Broken Promise] in my main hand and [Club] in my off hand to create a overwhelming difference between off hand and main hand hits. I then witnessed a proc of Mirror of Truth from an offhand crit (mainhand miss, offhand glancing, offhand crit, mainhand hit, was the section of hits that procced the mirror). I also witnessed a proc of grim toll immediately after an off hand hit. The Grim Toll proc could have just been server lag but the fact that the mirror of truth procced when only my offhand had crit signifies that offhand hits/crits can proc trinkets.

Last edited by AtheistGod : 06/25/09 at 3:57 PM.

Offline
Old 06/25/09, 3:52 PM   #17
Kyruski
Piston Honda
 
Troll Death Knight
 
Lightning's Blade
Ok, I took a stab at the Blood Strike vs. Obliterate for Death runes ordeal.

Right now, Rime Procs produce 5RP. For now, let's ignore the 2pT9 set bonus.

For testing, we need to include HB in our OB damage because of Rime Procs. Average calculations for BS, OB, and HB should be as follows:
Originally Posted by Calculations Template
BS: (non-crit avg damage * proportion of non-crits) + (crit avg damage * proportion of crits) = Expected Avg Damage of a BS.

OB Option 1: (((non-crit avg damage * proportion of non-crits) + (crit avg damage * proportion of crits)) * 2) + (Expected Avg Damage of a HB * .15(Rime Proc Chance)) + (Expected Avg Damage of a FS * .15(Rime Proc Chance) * .15625(The number of FS a Rime Proc produces from RP)) = Expected Avg Damage of an OB.

OB Option 2: ((non-crit avg damage * proportion of non-crits) + (crit avg damage * proportion of crits) + (Expected Avg Damage of a HB * .15(Rime Proc Chance)) + (Expected Avg Damage of a FS * .15(Rime Proc Chance) * .15625(The number of FS a Rime Proc produces from RP))) * 2 = Expected Avg Damage of an OB.

HB: (non-crit avg damage * proportion of non-crits) + (crit avg damage * proportion of crits) = Expected Avg Damage of a HB.

FS: ((non-crit avg damage * proportion of non-crits) + (crit avg damage * proportion of crits)) * 2 = Expected Avg Damage of a FS.
There are two options for OB. OB Option 1 is both swings of OB have only 1 chance to Proc Rime. OB Option 2 is each swing of OB has a chance to Proc Rime.

If there is enough time without pushing back other GCDs and 2 BS > 1 OB, then 2 BS.

Testing Circumstances
  • Spec used: 10/53/8 (2 points spare in this for Merciless Combat, Removed for testing).
  • I did 100 strikes (100 MH, 100 OH) of BS and 100 of OB, as well as 100 of HB and 100 of FS (For the contribution that it has VIA OB Rime Procs).
  • Weapons: 2x[Shortsword], No FC on either for Consistent Uptime.
  • No Trinkets used.
  • OB Glyph was included and so was 4pT8.
  • Tested on a Test Dummy.
  • No Horn of Winter.
  • 0/5 Killing Machine used for the Howling Blast test (KM procs would be used on FS in a fight).
  • 5/5 Killing Machine used for the Frost Strike test (KM procs would be used on FS in a fight).
  • SotVH used.
  • Just did a Dump system at the end for FS, Didn't pay attention to KM Procs

Recount Totals
OB Crit: Avg = 2309, % = 43.5%
OB Non-Crit: Avg = 1010, % = 56.5%
BS Crit: Avg = 2081, % = 29%
BS Non-Crit: Avg = 860, % = 71%
HB Crit: Avg = 4764, % = 12%
HB Non-Crit: Avg = 2150, % = 88%
FS Crit: Avg = 2290, % = 57.5%
FS Non-Crit: Avg = 939, % = 42.5%
Remember, These are averages of OH and MH, So OH hits lower on average than this and MH hits higher.

Now for the Calculations:
Originally Posted by Calculations
BS: ((860*.71)+(2081*.29))*2 = 2428.18

OB Option 1: (((1010*.565)+(2309*.435))* 2)+(2464.16*.15)+(3431.65*.15*.15625) = 3600.183297

OB Option 2: ((1010*.565)+(2309*.435)+(2464.16*.15)+(3431.65*.15*.15625))*2 = 4050.236594

HB: (2150*.88)+(4768*.12) = 2464.16.

FS: ((939*.425)+(2290*.575))*2 = 3431.65.
So now let's see if 2 BS > 1 OB.

2428.18*2 = 4856.36
4856.36 > OB Option 1 and 2

So short answer is Yes, 2 BS is better than 1 OB.

Now let's go into a little more detail.

Things I found on the PTR and things to make note of
  • MH Misses/Dodges would not use runes but still do OH Damage.
  • It seemed impossible to keep a rotation as long as PS>IT>OB>BS>BS>Dump>OB>OB>OB>Dump. Changing to OB>OB>BS>BS Would be even harder because of Length. Including Rime procs is a lot harder. (The difficulty may be due to Lag on the PTR)
  • OB has a higher Weapon % than BS, so think about that while choosing.

For sake of Rotation lengths, It might be best to Keep it OB>OB>OB instead of OB>OB>BS>BS if you want to stay in BP (We might do better DPS in UP). It could also be worthwhile to look into the Disease (I know Frost and Blood have looked into it for 2H and proved it wasn't a DPS gain, though there is a possibility that we are different). Remember to take everything I've shown here with a grain of salt and Interpret the data yourself. We alos need to look into the idea of not including HB for Rotation's Sake. Hope this helps.

Offline
Old 06/25/09, 4:23 PM   #18
Fargom
Piston Honda
 
Human Death Knight
 
Lightbringer
@ syrellia:

You are correct in saying that shorter disease rotations can indeed compete. Correct me if I'm wrong, but 2h frost currently has three builds, all of which are neck and neck, and one of them does infact use short disease rotations? I'm not going to update the OP with anything in regards to Epidemic vs short rotations until someone proves it one way or the other. Thank you for pointing out that we need to consider all options.

@ Concept84:

The Frost strike VS Deathcoil issue needs to be tested, instead of posting what we "think" will be better lets do some actual testing. Frost strike has had a reduction in power, and a build that uses deathcoil would be able to take advantage of the new UB and would place a lower value on expertise as a result of using UB. There is no way to argue at this point for one or the other, but we sure can test things.

@ Atheistgod:

Your numbers for Frost strike vs Blood strike are troublesom. I'm sure the numbers you are reporting are accurate, but a max frist strike crit of 4000ish seems very low. Is it possible that Recount is only looking at the first strike? If your numbers are correct, it very well might mean that BS is the best option for a deathrune, as it will do more damage than the icy touch + extra RP.

Nice observations regarding offhand procs, the Fury of the five flights issue makes me think that there is a very short internal cooldown on procs, which makes sense when applied to your observation about MH and OH hits landing at the same time. If the hits land at the same time, only one charge is applied to FOTFF, however, if a short delay is put inbetween each hand does infact proc a charge. This clearly shows that offhand hits are capable of proccing various things, can you please edit in a clipped SS of the combat log? I will consider this a closed issue and have the OP point to your Screenshot.

Offline
Old 06/25/09, 4:27 PM   #19
dr_AllCOM3
Great Tiger
 
dr_AllCOM3's Avatar
 
Orc Death Knight
 
Blutkessel (EU)
Originally Posted by AtheistGod View Post
[...] I believe it is very likely that strikes proc Cinder Glacier. I had Cinder Glacier procs occur much more than Killing Machine procs with CG procs occurring at least once and often twice during one 15 second disease interval.
With Fury of the Five Flights off hand attacks only stacked it when they don't hit at the same time as the main hand. Strikes only added one stack. [...]
Originally Posted by Kyruski View Post
[...] There are two options for OB. OB Option 1 is both swings of OB have only 1 chance to Proc Rime. OB Option 2 is each swing of OB has a chance to Proc Rime. [...]
Because of that, OB most likely doesn't have an increased chance to proc Rime.

Strikes proc stuff, that has always been the case.


Offline
Old 06/25/09, 4:40 PM   #20
 Darkside
I find your lack of faith disturbing
 
Darkside's Avatar
 
Kroot
Orc Death Knight
 
No WoW Account
@Kyruski:

In your calculations, why are you multiplying the crit damage portion of each attack by a factor of 2?

Ex: ((860*.71)+(2081*.29))*2 = 2428.18

I would assume that the average damage would simply be the weighted average of the crits/non-crits.

[e]: It would appear as if I can't read.

Last edited by Darkside : 06/25/09 at 4:52 PM.

Originally Posted by Silmeria View Post
See this is how engineers argue! Why the fuck we gotta have 17 page threads on how much Diablo 3 sucks I blame liberal arts majors

United States Online
Old 06/25/09, 4:45 PM   #21
Kyruski
Piston Honda
 
Troll Death Knight
 
Lightning's Blade
Originally Posted by Darkside View Post
@Kyruski:

In your calculations, why are you multiplying the crit damage portion of each attack by a factor of 2?

Ex: ((860*.71)+(2081*.29))*2 = 2428.18

I would assume that the average damage would simply be the weighted average of the crits/non-crits.
That is actually multiplying the sum of the Avg Non-Crit/Percentage + Avg Crit/Percentage by 2. If you notice, the Crit+Non-Crit is inside the Parentheses. This is because Recount Counts each time you use BS or OB as 2 because of the Swings, 1 for MH and 1 for OH. So without multiplying, it becomes the Average damage of (MH+OH)/2, so multiplying by 2 bings it back closer to the Actual of the two combined hits.

Offline
Old 06/25/09, 5:03 PM   #22
AtheistGod
Piston Honda
 
AtheistGod's Avatar
 
Orc Death Knight
 
Zul'Jin
Recount counts main hand and off hand separately. Those max numbers are main hand hits only.

I forgot to take a screenshot of the first occurrence but I was able to get another such situation.
Remember I'm using [Broken Promise] in mainhand and [Club] in offhand so mainhand hits for about 600-800 offhands hit for about 200.

Crit of offhand procs Reflections of Torment

Offline
Old 06/25/09, 5:11 PM   #23
AtheistGod
Piston Honda
 
AtheistGod's Avatar
 
Orc Death Knight
 
Zul'Jin
I was asked to figure out if Death Strike heals for 10% or more because of the offhand attack. In testing this I came across the issue that Death Strike still only hits with the mainhand. For some reason ToT does not work on Death Strike on the PTR.

Offline
Old 06/25/09, 5:16 PM   #24
Aezoc
Piston Honda
 
Orc Death Knight
 
Scilla
@Kyruski: The blood strike specific portion of your testing will need to be redone next PTR patch, as the bonus damage from diseases is being cut in half (50%->25%). Also, I believe the better test is whether 2x BS > 2x IT + FS, as that's currently the ideal way to spend those death runes as frost. I guess with the FS nerfs the value of 2x IT + FS might need to be re-evaluated compared to Ob, however. And I can confirm that CG (all proc-based rune enchants, actually) do proc from strikes. The easiest way to test this is to gradually face away from the dummy until you can no longer hit it with auto attacks. If you do it incrementally enough, you should still be able to use special attacks. Spam them for a few seconds until you get a proc, and voila.

Blood strike - Is it now doing enough damage to justify using Death runes for BS instead of other attacks?
Unfortunately I don't think this question has a simple yes/no answer. It's actually a bit of a chicken and egg problem, in that we only care about the answer for the highest DPS DW build, but we're not sure yet which that is. I'll take a stab at this for deep frost, but 23/48 and deep unholy would need a separate set of calculations.

Assumptions
  • BS damage is increased by 25% per disease, using the latest patch notes.
  • Not factoring in any sigils, since SotVH likely won't be BiS next patch
  • Using glyphs of IT and FS
  • For ease of calculation, assume paired weapons of the same speed and DPS. This lets us calculate total BS/FS damage by multiplying the MH damage by 1.575
  • Assume raid buffs and spell hit cap. BS/FS are mostly affected by the same buffs - BS will gain 4% damage relative to FS (Blood Frenzy), and IT will gain 13% damage.
  • I'm ignoring crit chance for now, because it's pretty complicated, especially with KM procs. But see below on why I don't think that affects the outcome.
  • Given the glyph choices, the comparison is really 2x BS vs. 2x IT + 93.75% of a FS

Weapon damage = wpn_base + AP / 14 * 2.4

BS damage = (0.4 * wpn + 305.6) * 1.5 * 1.1 * 1.575 * 1.04
FS damage = (0.55 * wpn + 138) * 1.1 * 1.1 * 1.15 * 1.2 * 1.575 * .9375
IT damage = (236 + AP * .1) * 1.1 * 1.15 * 1.15 * 1.2 * 1.13

Feel free to plug in various AP and weapon damage numbers. There is no one answer, even given the assumptions listed, because the BS option scales worse with AP but better with weapon damage. However, what I consider to be all sane values ITx2 wins handily. Now, I didn't consider crit. Ignoring Dark Conviction, IT has 15% crit from talents, FS has 3%, and BS has 12%. If we generously assume that 1% crit = 1% increased damage (it's less, but bear with me) and we ignore the crit that IT and FS get, then with 5500 AP and two 2.6s ilvl 239 weapons, 2x BS is better than 2x IT + FS. Except we haven't considered armor. BS is affected by it, 2x IT + FS isn't. Again, this is somewhat subjective because everyone will have a different idea of what a reasonable amount of ArP is, but if we assume the boss has even 10% damage reduction from armor, then 2x IT + FS pulls ahead again by a decent margin.

Conclusions
I think that calculating the exact average damage of each option will require a full-fledged spreadsheet, or ideally, an updated simulator. However, I feel that the scenario above is proof that for the average case, ITx2 is a better use of death runes for a 51 frost spec that is glyphed for IT and FS, since it outperforms 2x BS in a situation that is very favorable for 2x BS (only 10% reduction from armor, and ignoring IT/FS damage increases from crit). This raises several new questions though, like whether the current deep frost glyphs will still be the best choices in 3.2, and how deep frost stacks up to 23/48 overall. So while I don't have time to calculate the answer to this question for the various other specs, hopefully this will prove useful in getting the ball rolling.

Offline
Old 06/25/09, 5:17 PM   #25
Melchior
Piston Honda
 
Melchior's Avatar
 
Orc Death Knight
 
Skywall
Originally Posted by dr_AllCOM3 View Post
Everything is explained in the second post. Why didn't you read it? I did and it was well worth it.
The offhand strikes, so slow it better. I'd say it's speed>stats>dps, but mostly stats and dps go hand in hand. The ilvl239 weapons e.g. aren't good. My sheet was correct about the strikes, so feel free to test it out yourself.
Edit - I was screwing my own results because I had SoA equipped, nevermind.

Last edited by Melchior : 06/25/09 at 5:46 PM.


Offline
Old 06/25/09, 5:22 PM   #26
concept84
Piston Honda
 
Blood Elf Death Knight
 
Darrowmere
Originally Posted by Fargom View Post
@ Concept84:

The Frost strike VS Deathcoil issue needs to be tested, instead of posting what we "think" will be better lets do some actual testing. Frost strike has had a reduction in power, and a build that uses deathcoil would be able to take advantage of the new UB and would place a lower value on expertise as a result of using UB. There is no way to argue at this point for one or the other, but we sure can test things.
While I agree this needs to be tested again, and admittedly would rather use Death Coil, we have to remember back in the old DW days Frost Strike beat out Death Coil by a very large margin, and I just don't feel a 30% DoT can make up for that, especially with the increased runic power generation from OB with CoTG and Dirge, plus the FS Glyph lowering its cost.

Also remember that Frost Strike hit harder than Death Coil before when Death Coil was fully buffed by Impurity at 25% and Frost Strike was hitting based solely off the Main Hand Weapon's damage.

So my post was not really a "guess" at all, merely an observation from the recent pass, and unless I'm missing something very large here, I believe this to be the case come 3.2.

Offline
Old 06/25/09, 5:27 PM   #27
Kyruski
Piston Honda
 
Troll Death Knight
 
Lightning's Blade
Originally Posted by Aezoc View Post
@Kyruski: The blood strike specific portion of your testing will need to be redone next PTR patch, as the bonus damage from diseases is being cut in half (50%->25%).
I thought that was currently implemented, the 25% per disease part. Also if I was having trouble keeping a rotation that's OB>OB>BS>BS>Dump, IT will still be somewhat hard, though haste will help. Like I said, it's up to you for interpretation.

Offline
Old 06/25/09, 5:31 PM   #28
AtheistGod
Piston Honda
 
AtheistGod's Avatar
 
Orc Death Knight
 
Zul'Jin
Originally Posted by Melchior View Post
When I equip a 1.1 DPS MH and a Razorscale Talon OH, the OH will not strike harder than the MH, which it should if it were actually looking at the OH weapon shouldn't it? If I switch the two, the OH still strikes for roughly the value you mentioned (57.5% of the MH damage). I don't see where the OH weapon is actually used.
My testing goes against that. My two hits when using a a 1.1 DPS mainhand with Broken Promise offhand are less than 25% different.

Offline
Old 06/25/09, 5:33 PM   #29
Aezoc
Piston Honda
 
Orc Death Knight
 
Scilla
Originally Posted by Kyruski View Post
I thought that was currently implemented, the 25% per disease part.
The tooltip still reads 50% per disease. Messing around with the dummy also seems to indicate that it's doing basically double damage with diseases up.

Offline
Old 06/25/09, 5:35 PM   #30
Melchior
Piston Honda
 
Melchior's Avatar
 
Orc Death Knight
 
Skywall
Originally Posted by AtheistGod View Post
My testing goes against that. My two hits when using a a 1.1 DPS mainhand with Broken Promise offhand are less than 25% different.
Edit - testing more.


Offline
Closed Thread

Go Back   Elitist Jerks » Class Mechanics » Death Knights

Thread Tools

Similar Threads
Thread Thread Starter Forum Replies Last Post
[Elemental] 3.1 PTR changes and testing Ezareth Shamans 448 06/05/09 7:43 AM
Recount 2 Testing Cryect User Interface and AddOns 41 01/15/09 5:08 PM
Shaman DPS PTR testing Durnitol Class Mechanics 9 04/28/07 5:59 PM