I have incorporated some of the suggestions above and added a lot more support for 3.2 specs and items, including presets for ruptureless mutilate and combat specs. There are some UI improvements too, including running simualtion from the front page and saving glyphs into spec presets for convenience.
Also finished inputting all heroic and regular item versions, including both horde and alliance equivalents. All item sockets bonuses now have socket colors next to them to help decide with socket bonuses to activate. The latest link can be found here. Soon I will prepare full lists of BiS gear for each spec along with my DPS estimates.
For those of you who are using EP values on the gear page. Currently the ones listed there are from 3.1 BiS. Obtaining 3.2 values is going to take some time since I run each stats for each spec for 10,000 iteration over about 10 hours per spec. However, for choosing new TOC items in the beginning, the listed 3.1 values should be a good indicator. As before, to make a final decision on gear or spec, you have to actually run the simulation and compare the average DPS.
On the Talents and Glyphs tab, 15/51/5 - Eviscerate build button incorrectly switches the talents to 18/51/2 rupture build. It correctly switches the glyphs though.
I will fix that in the next version, for now you can change the talents manually. I would put 5/5 malice and 2/3 ruthlessness for that spec since it appears to do most damage out of the rest of available 15/51/5 options.
I am going to start a discussion of 3.2.2 changes and it's effects on dps as seen based on the simualtion analysis (new 3.2.2 sheet is already out). There are three major changes that completely change the ranking of specs:
1. HaT, which used to be highest dps spec in 3.2 is getting a nerf which renders it unplayable in PVE. The limit of 1 combo point per second from all group members including the rogue himself, put it 10-15% behind combat and mutilate. Whether it Blizzard developers think that the utility talents in that tree are worth that dps difference or whether it was their intention to completely devoid the spec of PVE use is basically not very interesting. The outcome is what matters most.
2. Armor penetration is getting a nerf via reduction of its conversion coefficient. While it seems like a change to armor penetration only, the interrelation between stats makes it so other EP values are affected as well. In particular, the EP value of haste becomes slightly lower than that of crit rating for mutilate. The effect of armor penetration change is manifold. It effects the BiS spec and gear for mutilate as well as BiS spec and gear for combat. It also bridges the gap between mutilate and combat.
3. Two changes to mutilate that favor envenom: a) AP coefficient is increased from 7% to 9% per combo point, and MP is changed to provide a chance of envenom not dropping DP stacks as opposed to increasing the chance to apply DP poison. The effect of this change is to put envenom and agility based mutilate firmly ahead of other mutilate subspecs.
Now, I am going to discuss BiS gear and spec of mutilate, since this spec once again seems promising in 3.2.2. All the discussion below assumes the target is nonmurderable.
Gear:
Twin Spike Heroic
Stygian Bladebreaker Heroic
BRK-1000
VanCleef's Helmet of Triumph Heroic
Charge of the Demon Lord Heroic
VanCleef's Pauldrons of Triumph Heroic
Vereesa's Dexterity
VanCleef's Breastplate of Triumph Heroic
Armbands of Dark Determination Heroic
VanCleef's Gauntlets of Triumph Heroic
Belt of the Merciless Killer Heroic
Legwraps of the Broken Beast Heroic
Treads of the Icewalker Heroic
Ring of Callous Aggression Heroic
Planestalker Signet Heroic
Comet's Trail
Death's Choice Heroic
JC and BS.
Gemming: 1 prismatic gem in the helmet, 7 deadly ametrines in yellow sockets in items that either have only yellow sockets or yellow-red sockets, 13 delicate rubies, and 3 delicate dragon eyes.
Food: agility.
Total Sum from Gear: STR- 20, STAM - 1544, AP - 2607, AG - 1928, Hit - 541, Haste - 346, ARP - 278, Crit - 745, EXP - 154.
Rotation:
- rupture in rotation, with 1 second delay of other finishers if rupture is about to run out
- no envenom timing (new MP makes it meaningless to time envenoms to DP ticks, pooling energy does not make any dps difference either)
Other notes:
Possibility of offhand dagger switching is still not fully evaluated, but it's likely to add up to 400 dps to this spec at the cost of swapping daggers twice every 12 seconds. The biggest challenge is to evaluate human factor in such involved micromanagement.
Last edited by Mavanas : 09/06/09 at 7:33 PM.
Reason: Cloak enchant and new gun.
It does not take murder into account, even though the spec I listed has murder in it. I put that talent there, from a practical standpoint, thinking that if you are going to have a mutilate offspec, it will most likely have murder talent for those beast and giant encounters in ulduar and toc.
As for dps comparison, I spent a lot of time with mutilate and wanted to spend at least as much with combat, looking for what's best in 3.2.2. However, I can see even now that combat is going to be at least 300 less dps than mutilate on nonmurderable targets. On top of that, MP talent with offhand swapping will increase the gap even further.
That said, I will spend tomorrow analyzing new combat specs in 3.2.2, and I will have a dps estimate by the end of the night.
While it's probably true that Mutilate is better against Murderable targets, I have to wonder if that's really a useful observation given the current content. Considering just ToC for the moment:
Jaraxxus is a Demon
Twin Valkyrs are Undead
Anub'arak is Undead
Faction Champions are Humanoid and thus probably affected by Murder (though given the arena-ish nature of the fight it might be worth verifying that it actually does apply). But I think independent of whether they are or not, what's "good" for that fight has a lot less to do with single-target DPS optimization and a lot more to do with how much it helps control the encounter. And I'd argue having DPS cooldowns to apply during the hardest portions of the fight is fairly advantageous independent of which does more sustained DPS.
Beasts may be Murderable, but it's a Combat fight if there ever was one; the transitions, submerges, freezes, and other interruptions make cooldowns relatively more powerful, the fact that you can blow cooldowns when Icehowl is stunned favors Combat, the fact that you can KSp to avoid getting punted by both the worms and Icehowl prevents a lot of running, etc.
In short: I'd argue that every fight in ToC is either nonMurderable or better for Combat anyway. And even if you consider Ulduar to still be of interest, all the hardest fights aren't murderable either, with the exception of Freya (Yogg, Mim, and Alg are all not). So yes, in an abstract sense Murder might be higher sustained DPS than Combat on Murderable fights, but I'm having a hard time finding fights where that fact actually matters.
Sorry, I must have been not very clear. All the discussion here was for nonmurderable targets. With 3.2.2 changes, mutilate appears to do that much more damage than combat on nonmurderable targets. For murderable targets it will be even further ahead.
Possibly, though surprising if true. And I don't think it really changes the previous analysis very much - like, Jaraxxus and Algalon probably turn into Mutilate fights, and you can debate Twin Valkyrs, but I think Combat's structural advantages will keep it ahead in practice even if it's a couple percent behind in theory.
Gear: (soft arp cap)
Stormpike Cleaver Heroic
Lionhead Slasher Heroic
BRK-1000
VanCleef's Helmet of Triumph Heroic
Charge of the Demon Lord Heroic
VanCleef's Pauldrons of Triumph Heroic
Vereesa's Dexterity
Cuirass of Calamitous Fate Heroic
Armbands of Dark Determination Heroic
VanCleef's Gauntlets of Triumph Heroic
Belt of the Merciless Killer Heroic
VanCleef's Legplates of Triumph Heroic
Treads of the Icewalker Heroic
Ring of Callous Aggression Heroic
Planestalker Signet Heroic
Mjolnir Runestone
Death's Choice Heroic
JC and BS.
Gemming: 1 prismatic gem in the helmet, 6 deadly ametrines in yellow sockets in items that either have only yellow sockets or yellow-red sockets, 14 fractured rubies, 2 delicate dragon eyes and 1 fractured dragon eye.
Food: armor penetration.
Total Sum from Gear: STR- 20, STAM - 1544, AP - 2343, AG - 1632, Hit - 472, Haste - 346, ARP - 697, Crit - 832, EXP - 154.
Rotation:
- rupture in rotation, with 1 second delay of other finishers if rupture is about to run out
Poisons and enchants:
Double wound poison, accuracy enchant on the offhand, rest of the enchants are standard.
To address burst capabilities of combat and abiltity to do more damage to murderable targets as mutilate, I prepared a table for three specs: 1. combat, 2. mutilate and 3. mutilate with murder for murderable targets. I added a fight duration dimension to the comparison to compare burst capabilities of the two specs.
There are other aspects of combat that can make it advantageous for certain fights, such as higher AOE damage and ability to use blade flurry. Also KS is a good way to escape certain knockbacks and fears. However based on this comparison, I'd argue that mutilate will be a spec of choice for a vast majority of fights, definitely for those that are murderable. Even over the span of 30 seconds mutilate with murder beats combat despite 3 cooldowns that can be used in that time.
1) It is my understanding that Swordguard Embroidery has been nerfed; see, for instance, the 5 posts starting here. I haven't tested it myself, but what's provided there would seem to indicate that it has changed. As such, tailoring is no longer remotely interesting or competitive, so I substituted Blacksmithing.
2) With double wound setups, I find that Haste passes Crit in value, resulting in the use of Deft gems instead of Deadly.
3) With these adjustments, I also show BRK-1000 passing Talonstrike, which mandates some resocketing.
With these adjustments, I wind up with 6 Deft gems, 1 Enchanted Tear, 1 Delicate Dragon's Eye, and everything else Fractured, resulting in final stats of 20 str, 1598 agi, 1544 sta, 2343 AP, 772 crit, 472 hit, 154 exp, 406 haste, 731 ArPen.
I'd also note that I show Eviscerate and Rupture cycle dps to be within a hundredth of a percent of each other for almost all these setups; with the setup I provide it's actually a bit ahead, 9597.5 to 9597.1. However, I think in reality, which is ahead doesn't matter much when the margin is so small - the eviscerate cycle is easier to use, and the damage difference is indistinguishably small; as such, I'd advocate using Eviscerate cycles either way.
Note that best results for Eviscerate cycles are typically gotten with 2/3 Improved Eviscerate and 3/3 Ruthlessness, rather than vice versa.
Although the changes coming in 3.2.2 aren't final yet, it would be interesting to know a few details about these new numbers and how they were generated.
In the testruns performed, you use BiS gear for both mutilate and combat to generate the numbers. While I assume it's still in early stages of testing, have you discovered where the breakingpoint is for mutilate vs combat, i.e. with what gear it becomes more beneficial to use one spec or the other? Is mutilate's theoretical advantage over combat active with all "close-to-BiS" gearsetups, or is there a mutilate sweetspot (like, t9 only) that shoots it ahead of combat?
A second question I'd like to ask is similar to the previous one: Have you discovered a point when it becomes more valuable for combat to go with double wound over wound/deadly poison?
And lastly, while I do agree with Aldriana regarding combat's situational benefits over mutilate, do you believe - if these changes remain intact - that mutilate will be the favorable raiding spec when looking to what types of encounters we face in ToC? Is the dps increase worth sacrificing the use of combat cooldowns in an actual raiding environment?
I find it very interesting to see that nonmurder mutilate is pulling ahead of combat on a theoretical level, but to put it into practice may very well prove to be a difficult task.
Last edited by Devlin : 09/06/09 at 7:20 PM.
Reason: rephrased question
Those are good suggestions, and I did not even have that gun in the spreadsheet. It's definitely an upgrade over Talonstrike and Diplomat. I do note that crit rating might be slightly ahead of haste in this setup since changing the 6 gems from deadly to deft resulted in a 6 dps decrease. However, obviously this difference is not statistically significant and could be a result of rng only. Either way I left the deadly gems in for now.
I have changed the embroidery cooldown and proc rate before as I said in the mutilate post, but I made a mistake of leaving the 22 ag from the regular cloak enchant in, so I mistakenly saw a dps increase from tailoring. In reality, JC/BS is the winning combo. Another note about gemming, if you use two delicate dragons eyes and armor penetration food, it should be a slight dps increase even though you are going 2 points over the soft cap.
Eviscerate only cycle is a clear dps decrease in the simsheet, by about 35 dps, which is a statistically significant difference.
Putting the rest of the changes in results in a 45 dps increase (9499 dps in a 5 min fight). Making the changes to the gun and cloak enchant to the mutilate setup brings it up to 9787 dps.
Understanding the 9453 Combat BiS Dps and the 9714 Mutilate BiS dps, I only see a difference of about 2.76 percent. I don't want to be quoted as the authority on this, but it is my opinion that such a small percentage dps advantage hardly negates the situational advantage that combat provides. Dps prediction assumes perfect situation, mutilate simply suffers a greater dps loss the more reality strays from this perfect situation (movements, target switching, dps downtime). Please correct me on my last statement if it turns out to be untrue.
Best
Last edited by Tofuu : 09/06/09 at 7:56 PM.
Reason: silly mistake
7 deadly ametrines in yellow sockets in items that either have only yellow sockets or yellow-red sockets
You seem to have 1 additional ametrine socketed that I can't see where you're putting it. My count shows 7 yellow sockets in total, but one is in the legs which also has a blue socket, so it gets red,red,red.
My bad, there are in fact 6 deadly ametrines in there. I know how I made that error. When I added BRK-1000, I thought I had talonstrike in the setup so I automatically added another orange gem, whereas I actually had the diplomat in my bis setup before, which was already socketed with an orange gem. So yeah 6 deadly ametrines, total agility = 1938, total crit = 735, the rest should stay the same.
And Tofuu, while your argument made sense if the dps difference were in fact 0.22%, seeing how it is close to 3%, it is not a trivial difference at all. We are talking about close to 290 dps here. The are other benefits of combat that are far more important than the one you proposed, such as burst and aoe capabilities. So depending on the fight, one will still want to use combat on occasion. But if the 3% difference is correctly estimated, I'd personally side with mutilate for a vast majority of fights in TOC and Ulduar.
I am assuming that the disparity between mut (non murderable) and combat does not take into account the potential of Mut weapon swapping, as has been discussed a page or so ago? If one considers weapon swapping, will the gap between the two be - too ridiculous to even consider combat?
The reason I do not bring weapon swapping into this yet is because I am not sure how feasible it is to worry about swapping weapons twice every 12 seconds. It might be too much to worry about. I actually wanted to give it a try on the dummies to see how hard and attention intensive it is. But yeah weapon swapping would push mutilate even further ahead. Main reason is that talented instant poison dps is higher than wound poison dps. So while you can also do weapon swapping with combat, the benefit should be lower (did not check this yet). If I can work out a sensible rotation with weapon swapping, I will model it in the simsheet, it should not be very hard to incorporate swing resets and global cooldowns.
If weapon swapping is feasible, it might also turn out that 51/18/2 will do more damage than 51/13/7 due to lightning reflexes.
I'd also note that I show Eviscerate and Rupture cycle dps to be within a hundredth of a percent of each other for almost all these setups; with the setup I provide it's actually a bit ahead, 9597.5 to 9597.1. However, I think in reality, which is ahead doesn't matter much when the margin is so small - the eviscerate cycle is easier to use, and the damage difference is indistinguishably small; as such, I'd advocate using Eviscerate cycles either way.
Note that best results for Eviscerate cycles are typically gotten with 2/3 Improved Eviscerate and 3/3 Ruthlessness, rather than vice versa.
On another note. While I understand that only by calculating BiS gear it is possible to calculate maximum possible damage for a class or spec, I would really like to see you use more "realistic" gear setups or at least using also a 245 ilevel dps-Benchmark.
I think working only with pure 258-BiS gear, setting up 258-BiS-Lists and posting these dps-numbers generates a false sense of "this is the dps we really can achieve", while it is extremely unlikely that even rogues in the best guilds can get all this equipment, due to the droprate of Heroic set tokens only on Anub'arak and depending on attempt-numbers left. Even for Top-100-Guilds it might take 6 to 8 months of farming ToGC with 40+ attempts left to get enough Token drops.
Again, I understand the "scientific approach", I only would love to see some "real world numbers" too
Again, I understand the "scientific approach", I only would love to see some "real world numbers" too
That's part of what the spreadsheets are for. Getting people like Mavanas or Aldriana to work out "BiS pre ToC Hard Mode" lists, and then throwing things off by whether or not you can kill Algalon or not, or whether you run 10 mans... it's just too many variables. Check for yourself!
There *is* an interest in knowing what the best possible combination of current gear is; regardless of what the capabilities of our guild are, we all strive to reach a complete BiS set of gear.
I'd also note that I show Eviscerate and Rupture cycle dps to be within a hundredth of a percent of each other for almost all these setups; with the setup I provide it's actually a bit ahead, 9597.5 to 9597.1. However, I think in reality, which is ahead doesn't matter much when the margin is so small - the eviscerate cycle is easier to use, and the damage difference is indistinguishably small; as such, I'd advocate using Eviscerate cycles either way..
Out of curiosity, since human error does play a role in rupture cycles (not wasting a tick while at the same time, not letting it fall completely), what kind of rupture uptime numbers did you model this observation on? Granted Eviscerate rotations will be easier for most people, but I'm sure there are people who will also find Evis only cycles a bit boring. This isn't an issue for those who can maintain the rupture cycle, but are the numbers crazy high like 95%, or something more manageable?
08/30/09, 6:24 PM
Similar Threads
