It's not that 4 piece tier 9 pushes 3/13/55 above 0/17/54. It's that the lack of 2 piece tier 8 - which you won't actually lose until 4 piece tier 9 - does the trick. Just a slight clarification.
But, yes, there is a point with ArP where 3/13/55 will do more dps, even with 2 piece tier 8. What that exact point is, however, has yet to be determined. You're welcome to use Kahorie's (linked in the OP) to calculate it yourself.
I have about ~280 ArP, and 3/13/55 does about 25 dps less in the last time I ran the sim, for what that shows.
Excellent job Consider, it really needed the update.
One quick question about the rotation though, is there a reason you have OB before BS? I just figured it was just assuming you had desolation up. But on the initial cast, wouldn't you want to make sure it's up before you OB?
It's 200 strength (via Sigil of Virulence) to your first two Blood Strikes and an auto-attack or two or 3-5% damage (depending on how many points you have in Desolation) to your first Obliterate and an auto-attack.
I can't honestly say which is superior. The difference is probably quite minor, if there is a noticeable one at all (Over the course of an entire fight, it's maybe a single digit dps difference. Maybe. Remember, whichever order you choose, it will only make a difference that very first rotation. Thus even if doing one order over the other is a 2,000 damage gap - extremely doubtful - over the course of a 6 minute fight? That's 6 dps).
What I can say is that the simulator does Obliterate first (due to how the priority works) and that SoV will have a slightly higher uptime if you do Obliterate first (due to how the ICD interacts with when your other Obliterates happen after that first one). Thus the reason I list the rotation as it is. Doing your Blood Strikes first, however, isn't wrong, and I would be surprised if you noticed any difference.
Fast questions I figured I would pop this out there for the math dudes.
In terms of DPS pots to pop, wouldn't the Insane strength potions outweigh the speed potions during a heroism? If I am not mistaken, haste from gear [which is dk by dk prolly 5% to 10% on avg] and the heroism already push the Gargs cast time down to 1 second and you cannot push it below that due to global cool down restrictions. Just wanted another set of eyes on this in case I am missing something.
Second question [and a great deal less important given the circumstance] would be, if lets say I just proc a comets trail and do not have access to heroism. Does that get me close enough to the 1 second cast to merit using an insane str potion over a speed potion?
I was curious which spec would win between the two if each of them had BiS gear, DW Frost or Unholy Obliterate.
I took the gear posted in the Optimal BiS Gear thread, added up all the stats, enchants, gems, and my base character stats. Then I created a character in Kahorie's DK Simulator with that information.
Some things to note:
Talent templates used are the default 17-54 Unholy Oblit and 3-51-17 Frost in Kahorie's DK Simulator
Priorities used are the default Unholy Oblit and Frost priorities in Kahorie's DK Simulator
Both tests are in Blood Presence
DW Test is Fallen Crusader / Razor Ice with Sigil of Awareness
Unholy Test is Fallen Crusader with Sigil of Virulence
I am an orc.
I am a Blacksmith and Jewelcrafter, so I took the Bracer, Belt, and Gauntlet extra gem slots and plugged in +34 Str gems into it.
I used a Nightmare's Tear on items with 1blue/1red with a +6str socket bonus to activate my meta.
I used a 10str / 10crit gem on items with 1yellow/1red with a socket bonus of at least +6str.
Characters have 2piece and 4piece T9 Activated
Characters have Death's Choice and DMC: Greatness Activated
These are 100 Hour tests.
Edit: There are now screenshots of tallied up stat sets under the Optimal BiS Gear thread. These character sheets and results now reflect those screenshots added with the base stats of my orc.
As you can see based on the BiS gear posted in the Optimal BiS gear for each spec Unholy Obliterate wins 8945DPS to 8678DPS. I was under the impression the DW Frost was supposed to win in the end, but the sim shows other wise.
If anyone else has tried this, gotten different results, or find that my math on character stats are wrong please feel free to let me know.
I just wanted to post my findings here, and I am sorry if this was already posted.
Last edited by NaeblisHyjal : 08/31/09 at 8:46 PM.
Oddly enough, Unholy outperforms Blood in Kahorie's sim, both pre and post 4p t9. It probably undervalues Hysteria, DRW, and cooldown stacking in general - one of Blood's strengths. It's certainly undervaluing (or overvaluing) something, as there's yet to be any parses reflecting the type of lead the simulator hints at.
It must be the cooldowns. My dps usually skyrockets when I get a good combo like BL+DRW+Hysteria. In almost every fight there are small interruptions and they often fit the DRW timer very well. The sim doesn't stack them.
I've had a read through this thread and I have to say, I'm switching back to unholy now to give this a try so thanks for the solid info on it.
One question though, why 4/5 Necrosis when it has been documented as a fairly low increase in dps compared to other talents. My suggestion would be that since auto-attack is such a big proportion of the dps in this spec then would it not make sense to drop the 2 points that you can spare out of necrosis into Icy Talons?
Feel free to tell me if i'm wrong as I don't have the inclination to do sim tests etc but I would have thought that the extra melee hits would be an improvement?
Icy Talons does not stack with Imp Icy Talons from a frost DK or Windfury from a shaman. Since it's assumed you have one of those two buffs as is, putting a couple points into IT is rather useless.
Originally Posted by haley
Well, this is confusing. Would a rule of thumb look like this (according to this table) ? :
What's confusing about it? T7 4p has always been an amazing set bonus. Remember, those APE values are only the set bonuses, not the set bonus plus the stats (hence why T8 2P + T8 4P does actually beat T7 4P or T8 2P + T7 2P).
In terms of DPS pots to pop, wouldn't the Insane strength potions outweigh the speed potions during a heroism? If I am not mistaken, haste from gear [which is dk by dk prolly 5% to 10% on avg] and the heroism already push the Gargs cast time down to 1 second and you cannot push it below that due to global cool down restrictions. Just wanted another set of eyes on this in case I am missing something.
I don't believe haste from gear and heroism already push the Gargoyle down to the gcd (50% haste on a 2 second cast does not make for a 1 second cast time, remember) or, if it does, that the Gargoyle actually has a 1 second gcd limit. I could be mistaken, though. Haven't really seen much testing on the subject, but it would be interesting to know.
Regardless, It's not all about the Gargoyle. 500 haste is far, far, far more dps for your personally - and for your Ghoul - than 120 Strength. It's the equivalent of ~750 AP compared to ~350 AP, which is a huge gap. Even if your Garg benefits from the latter over the former, you benefit from the former much more. And since you will always be your highest source of damage, it's a no brainer.
The 1 second rune refresh reduction in IUP doesn't provide as much dps as one would expect it to in PvE due to the grace period of runes.
Hard to explain, but hope this helps: When a rune refreshes itself, you essentially have a two second window where as long as you use it in those two seconds, it will come back up ten seconds from when it originally refreshed - not ten seconds after when you actually used it. Think of it as below:
For purposes of this example, let's say all of your runes except one Blood rune are ignored.
0 seconds - Blood Strike, Blood rune goes on CD
10 seconds - Blood rune comes off CD
11 seconds - Blood Strike, Blood rune goes on CD
20 seconds - Blood rune comes off CD
You see, in that example (which is exactly how it works), the rune actually refreshes in nine seconds after the first use. Thus a rune will refresh in 8 to 10 seconds, depending on how long after it comes off CD you use it. What IUP does is make a rune refresh in 8 to 9 seconds, depending on how long after it comes off CD you use it. So in the scenario above, it would refresh one second quicker initially, but that's all. So it does give some benefit, but it's not actually as much as you think - it's certainly not 11% more rune based damage/11% runic power generated. If it was, that on top of the haste would make it actually be worth using as Unholy.
Or such is my understanding of it.
For IUP (and haste in Cataclysm) to be truly effective, that grace period needs to scale as well. The grace period is essential (otherwise you're punished if you are mid GCD when a rune comes up, or if have even the slightest bit of lag, and so on), and the fact that it actually gets smaller with IUP is kinda nonsensical.
Your description of the IUP mechanics is accurate, but your interpretation (bolded above) is not. IUP will reduce your rune cooldown by 1 sec on every rune cooldown, not just the initial. In the above example, your Blood rune will come back 18 sec after its initial usage with IUP. The catch is that the reduced 1-sec grace period means that if your rotation uses runes asymmetrically between rune sets (most commonly BS+BS/OB and PS+IT/OB for Unholy), you have no leeway for latency/reaction delay (so if you have a 0.5-sec latency/reaction delay, your rune will effectively cool down in 9.5 sec instead of 9).
Rotations with symmetrical rune usage between rune sets, on the other hand, will have a 1-sec grace period for latency/reaction delay to get the full benefit (9-sec runes) out of IUP. Currently there is only one known competitive spec/rotation (Unholy DW DC spam) that has symmetrical rune usage to take full advantage of IUP; unfortunately, it is the one spec/rotation Blizzard intends to kill off (Ghostcrawler quote from his recent one-line post: "Die Shadowfrost die!").
I don't believe haste from gear and heroism already push the Gargoyle down to the gcd (50% haste on a 2 second cast does not make for a 1 second cast time, remember) or, if it does, that the Gargoyle actually has a 1 second gcd limit. I could be mistaken, though. Haven't really seen much testing on the subject, but it would be interesting to know.
Regardless, It's not all about the Gargoyle. 500 haste is far, far, far more dps for your personally - and for your Ghoul - than 120 Strength. It's the equivalent of ~750 AP compared to ~350 AP, which is a huge gap. Even if your Garg benefits from the latter over the former, you benefit from the former much more. And since you will always be your highest source of damage, it's a no brainer.
Thank you for the clarification. I think part of the thing that created a question on my part was the cast time which I assumed to be 1.5 seconds from wowhead, so there's where the oversight occurred.
As you can see, the stat weights for the DW Unholy are different from the 2-Hand Unholy versions. Haste is very strong. A version that has more ArP will have a higher Armor Penetration weight.
Using the 2H stats listed for BiS on this thread, using 200ms latency, this was my sim result:
DPS 9144
Total Damage 3291.92m in 100h
Threat Per Second 5991
Generated in 913s
Template: OB001853
Priority: Unholy Oblit
Presence: Blood
Sigil: Virulence
RuneEnchant: FallenCrusader / FallenCrusader
Pet Calculation: True
Edit: I realize that the DW weapons would have different stats than the 2H that was used. I just copied the stats and used Frost's weapons instead. A better comparison would alter the stats based on the change of weapons. My apologies for that bit of inaccuracy. I also forgot to remove the Draenei bonus from the buffs section.
Edit2: The corrected version is further down the page in a later post.
Last edited by Orlgin : 09/01/09 at 11:09 AM.
Reason: yet more info
Like an unchecked cancer, hate corrodes the personality and eats away its vital unity. Hate destroys a man's sense of values and his objectivity. It causes him to describe the beautiful as ugly and the ugly as beautiful, and to confuse the true with the false and the false with the true.
Martin Luther King Jr.
DW Stat Weights added. Also fixed an error in the haste stat weights for the previously listed numbers which was undervaluing haste.
@Sellout | Gargoyle Strike - Spell - World of Warcraft is the Gargoyle Strike you want to look at on Wowhead. The others, some of which are indeed 1.5s cast times, belong to various npcs - Gargoyles in Naxx, Auchenai Crypts, etc.
@Grigori | Yeah, I knew there was something wrong in that post about ten seconds after I posted it. I just wasn't quite sure what. Thanks for pointing out my obvious (in hindsight ><) error. Regardless, that is the reason why I did not include it here. The primary problem with Shadowfrost and such variants isn't so much IUP as it is Glyph of Icy Touch (in combination with Chill of the Grave). It's that mass RP which continues to cause issues. 25 RP from 1 rune is flawed in many ways.
@Sellout | Gargoyle Strike - Spell - World of Warcraft is the Gargoyle Strike you want to look at on Wowhead. The others, some of which are indeed 1.5s cast times, belong to various npcs - Gargoyles in Naxx, Auchenai Crypts, etc.
Yup that's exactly where the confusion came from, the multiple listings etc. Thanks for helping sort that out, its always good to have a second set of eyes on things.
@Orlgin: Exactly what build are you using? Seing the post above me I'm curious if you didn't use WP. Octopi stated earlier in the DW thread that he finds WP to be a low dps per point value (~2% 3/3) which surprises me. Using IUP is probably still as always a preference and I would never go without it if I could possibly get it - but is WP really the place to take the points from?
Also impressed by the output from a non-ToT DW build; would love to see a percentage of total damage from each skill (do Kahorie's list this?) to show how much is Obli and how much is DC (again Octopi stated he has around 14% Obli-damage and 24% DC).
There were two sections to that post. The first (stat weights) used the stats listed. The second has sim DPS value that was created using the BiS stats which are listed earlier in the thread.
I used the same setup as the one earlier in the thread for BiS. That includes both 2pc and 4pc T9. I used the setup NaeblisHyjal suggested in his post. I just forgot to do two things: 1) Change the stats a little since the 1H weapons would have different stats than the 2H and 2) Remove the Draenei bonus as he was an Orc. The purpose was to compare Unholy, Frost, and DW Unholy using the same statset which was best in slot. When I have more time, I'll edit the stats to reflect the new weapons, remove the Draenei bonus and post the result.
Edit: What a difference bad 1H itemization makes:
DPS 8829
Total Damage 3178.5m in 100h
Threat Per Second 5781
Generated in 882s
Template: OB001853
Priority: Unholy Oblit
Presence: Blood
Sigil: Virulence
RuneEnchant: FallenCrusader / FallenCrusader
Pet Calculation: True
This suggests that DW and 2H are roughly the same DPS in Best In Slot Gear. Good to know.
Last edited by Orlgin : 09/01/09 at 11:06 AM.
Reason: Adding fixed sim data
Like an unchecked cancer, hate corrodes the personality and eats away its vital unity. Hate destroys a man's sense of values and his objectivity. It causes him to describe the beautiful as ugly and the ugly as beautiful, and to confuse the true with the false and the false with the true.
Martin Luther King Jr.
I've recently started using the 3/13/55 2H spec after reading this post. I've already had several DK's I've raided with ask me about it... so very good job with the content in the OP.
After having great success with a GoD blood and GoD DW frost spec and rotation, I'm going to try it with the aforementioned unholy spec tomorrow in ToC (our 10-man night is Wednesday). If someone has already simulated 3/13/55 with GoD or has tried it and can link parses, I'd love to see them before I start experimenting.
I'll try to get some stuff posted on Thursday so the theory/math guys can look over it and check for any surprises in addition to getting a critique of the damage breakdowns to see if GoD warrants more research.
I was kinda guessing that already Orlgin, but I did notice the previous poster didn't so good that you point it out. I'm still curious about the damage-spread between skills though! Is the itemization really that bad when it puts them approx. at the same dps though.
I'm actually curious how RoCinderglacier would do for UH DW, I'm not assuming it'd be at the same dps but still.
Consider, what source or post are you quoting your stat weights from?
You suggest they are the currently accepted stat weights, but Zerack's Optimizer contains some values which are quite different. Has he and the hive-mind behind EP values agreed upon these weights recently, and simply not updated his app? If this is the case, Norg's BiS list will also currently be incorrect, and linking to it in this thread would be counter-productive, as they use a different weight scale than that which you provided.
This isn't meant to be a flame, I'm just trying to help steer the min/maxing in the right direction. I have great respect for you and the others for their time and contributions toward the topic.
I simply used the default 2H set in Kahorie's, set it to 1000 hours, and then ran the EP calculations.
I noticed after the fact that they differed slightly from Zerack's, however not significantly so. The ArP values he uses are slightly lower (and that's simply due to the nature of ArP - it depends heavily on how much you already have, and I think Kahorie's default 2H set has 300), and his haste values for 0/17/54 are kinda absurdly high (not quite sure why that's the case), but otherwise, they're all within 10% of one another, and it doesn't change the order of the stats. Make sure you're looking at Zerack's latest version, as his older versions have outdated/incorrect stat weights.
In the end, however, your best bet is to always calculate stats based on your own gear. Although ArP is the most prominent example, all stats fluctuate in value depending on how much of that stat and of others you already have.
I am very confident in saying that whatever weights you personally get, they will resemble these very closely - and these themselves resemble Zerack's (haste aside. Still boggles me why his is so high for 0/17/54. Main reason I didn't use his values, not that I don't trust them - the rest seem fine - but that seems kinda off) almost exactly.
Edit: Also, although I appreciate it, there's really no reason to post "Good guide" and nothing else or those sorts of things. You'll probably just end up getting an infraction. Critcism/discussion/numbers/etc are the sorts of things you should be posting ^^
I've recently started using the 3/13/55 2H spec after reading this post. I've already had several DK's I've raided with ask me about it... so very good job with the content in the OP.
After having great success with a GoD blood and GoD DW frost spec and rotation, I'm going to try it with the aforementioned unholy spec tomorrow in ToC (our 10-man night is Wednesday). If someone has already simulated 3/13/55 with GoD or has tried it and can link parses, I'd love to see them before I start experimenting.
I'll try to get some stuff posted on Thursday so the theory/math guys can look over it and check for any surprises in addition to getting a critique of the damage breakdowns to see if GoD warrants more research.
I ran the sim for this scenario a while back in the old unholy thread. That was with an older sim that didn't roll diseases with the highest AP diseases, only the initial ones. The newest version 1.0.7.0 does however. I reran the sim with the following: 100h, 2h ulduar gear template, blood presence, sigil of virulence (the str helps in favor of a GoD spec), FC rune and unholy oblit priority. 0/17/54 was ran with 5/5 necrosis 3/5 desolation 0/2 IUP 0/1 Frenzy. 3/13/55 was exactly the same with the 1 additional pt in unholy spent in desolation. Lastly, the reason I subbed out GotG was because in my previous post in the old thread, subbing GoD for GotG was always more dps than subbing GoDD out.
0/17/54 GoDD,GotG,GoOB
7104
0/17/54 GoD,GoDD, GoOB
7169
3/13/55 GoDD, GotG, GoOB
7101
3/13/55 GoD, GoDD, GoOB
7191
With the new sim, rolling diseases with the highest AP yields higher DPS when using GoD. One final note, the standard 2h ulduar gear template does not have any trinkets selected. Using trinkets with Str/AP procs would help in favor of a GoD build.
Thank you for the information. I'll swap out GotG for GoD and try to pick up SoV before the raid to provide slightly better test results. Once I get the parse posted, I'll try and be diligent about the various circumstances of the fights so everyone will have a better baseline and look more at timing of events, etc.
I was more concerned with the Hit (To Spell Cap) value. Under 3-13-55, your value is 0.57 whereas Zerack gives 1.67 for the same spec. This trend is actually common for all the builds you list. Depending on whether I use your or his EP values essentially decides if a build should be capped at hit level or capped at spell level, a difference of three or four pieces of gear with quick simulations. I can't wrap my head around how important it is to make sure you never miss an IT or DC, so I don't know which value to put more faith into. I've tried to use Kahorie's Sim to generate EP values for myself, but I have a fairly limited understanding of how it is calculated, and don't particularly trust my own results.
In the mean time, it still stands that Norg is using Zerack's values, and therefore the BiS gear that he lists do not follow your stat weights. One of them will eventually need to be corrected.
Edit: I just took a look at Kahorie's Google Code page and read * EP Mode: Spell hit rating was evaluated too low. With default 50 hit rating, it would over cap. Problem corrected. For 2H spec, the Spell hit value is doubled. under his very latest release (within 28 hours). I don't know if this was the source of some miscalculations or the coming of some new ones, but I'm guessing EP values may need to be reworked again.
08/30/09, 9:51 PM
