Jump to content


Photo

Frost DPS 5.4 - Howling Blargh


  • This topic is locked This topic is locked
86 replies to this topic

#21 Jessamy

Jessamy

    Struck by Diax's Rake

  • Moderators
  • 4803 posts

Posted 08 March 2013 - 10:18 PM

As a follow up to my earlier post about transitional gear decisions, I wondered how good the tier 15 set bonuses are. Are they good enough to warrant choosing 502 tier pieces instead of 522 off set items? So I created a profile with the 522 item list from my earlier post, but replaced all the 522 tier pieces with 502 tier pieces. Then I created a second profile with those items replaced with 522 non-tier items.

The gear lists I used are spoilered below. They're exported from Mr. Robot, and reforging may not be optimal, especially the hit & expertise. The point however isn't so much that 502 tier is dramatically better (it isn't), but that the two gear sets are comparable. Taking the LFR tier pieces for the fun bonuses isn't a loss, and won't hurt your performance. Replacing them with normal tier is a smooth transition.

Spoiler


#22 CausalXXLinkXx

CausalXXLinkXx

    Glass Joe

  • Members
  • 10 posts

Posted 09 March 2013 - 01:23 PM

I searched this thread and didn't find this, but if you have the 2 set, you are absolutely able to death pact off of your troll undead warriors. To some people this doesn't matter, but to me this means I can now keybind my raise dead and feel like I can use death pact any time I need it thanks to the 2 set.

#23 Scyzek

Scyzek

    Glass Joe

  • Members
  • 5 posts

Posted 13 March 2013 - 03:09 PM

I want to comment on and how it's been running for me. I'm 2h Frost. I run between 10k-10.5k haste and I do run Heroic . I have seen uptimes of 55%-70%.
Jin'rok - 16 procs, 57% uptime, 3:23 fight
Horridon - 53 procs, 69% uptime, 9:48 fight
Council - 45 procs, 65% uptime, 8:08 fight
Tortos - n/a
Megera - 34 procs, 58% uptime, 8:03 fight
Jin'kun - 29 procs, 53% uptime, 7:10 fight (I was on nests several times, thus low uptime)

All of these fights contained a heroism. I don't have any racial dps CD's. My melee uptimes were in the 99% range for most fights except megera and jin'kun.

While this is amazing uptime, the stack variation can vary greatly. Sometimes it'll run 1 stack and fall, then proc up to a 5 stack and run godmode. Fortunately, This buff (rampage) can be monitored and stacked easily with pots/pillar. I really really really want to get a spark (maybe even lfr version just to test) and see how well this str stack can do. I still have 4 piece t14 so I expect close to the 400k+ mark on OB crits.

#24 fangless

fangless

    Von Kaiser

  • Members
  • 34 posts

Posted 13 March 2013 - 08:31 PM

I assume this was before the hotfix just implemented that boosted the rates for the 5.2 trinkets?

Real Proc Per Minute (RPPM) Trinket Changes - Forums - World of Warcraft

#25 Mendenbarr

Mendenbarr

    Oh, I never leave home without my party cannon.

  • • Guide Author
  • 181 posts

Posted 13 March 2013 - 11:13 PM

Thanks to the many of you that sent me logs containing the new trinkets, it has really helped me find out as much out about how they work as possible, but with the changes mentioned in the post fangless linked, I now have to all but start over. Preliminary trinket rankings should be done by some time tomorrow, and I only need logs from tuesday or later at this point.

Also, if you are posting logs/uptimes for rage/gaze, please list uptimes for each stack type, eg. 1, 2, 3, 4, 5.

#26 Scyzek

Scyzek

    Glass Joe

  • Members
  • 5 posts

Posted 15 March 2013 - 01:12 PM

Blizzard says they buffed proc rates but my has tanked down to 50% uptimes. Mendenbarr- I'm not sure how to track the stack per proc in my logs. My guild did twin consorts and lei shen last night and I had good uptimes (99.7% on twins, 80% on lei shen). I've noticed the lack of stacks while fighting, because I can usually time a pillar on a 5 stack if I wait a few seconds. I almost came in second on twin consorts because of this, my guild might think I'm going soft if this doesn't go back up! Or it's back to lei shen's final orders for me (normal version).

edit: Looking forward to the quick ranking spreadsheet whenever that's ready. If spark behaves the same way after this 10% buff I'll probably be going for a feather and vp talisman.

#27 fangless

fangless

    Von Kaiser

  • Members
  • 34 posts

Posted 15 March 2013 - 02:21 PM

The Feather just seems so OP, here is a recent kill of ours on a largely patchwerk style fight;

Details for Waylandyr - 14-03 20:34 - Static - World of Logs

Our Unholy DK ended up with 50% uptime. Doesn't stack haste any more than just collateral stats. Other fights it's been 35%+ average. That's a lot more than I got with simcraft (using the latest svn code). I picked up the VP trinket since I didn't really plan to spend VP on any of the other slots, but I think I am going to shoot for the Feather as well.

#28 Mendenbarr

Mendenbarr

    Oh, I never leave home without my party cannon.

  • • Guide Author
  • 181 posts

Posted 15 March 2013 - 11:01 PM

Unholy rppm scales with unholy pres and unholy frenzy, but frost rppm does not scale with icy talons, so unholy uptime will always be higher.

First run at rppm new system modeling is live on simc, and here is the first set of results I've done. Linking only here for now, will update OP when I've finished a full set for all specs, and heroic mode trinkets:
This is in no way final.
522 ilevel trinkets, 522 ilevel DW frost:
Profile had 4804 haste without spark, and 5685 haste with spark.

[TABLE] DPS | % | Trinkets
148617 | 10.3% | Feather/Shadopan
147384 | 10.2% | Spark/Feather
147029 | 10.1% | Rage/Feather
146741 | 10.1% | Twins/Feather
144289 | 10.0% | Spark/Shadopan
144125 | 9.9% | Rage/Shadopan
143747 | 9.9% | Twins/Shadopan
142850 | 9.9% | Spark/Rage
142221 | 9.8% | Spark/Twins
142022 | 9.8% | Rage/Twins[/TABLE]

Suggested order: Feather > Shadopan > Spark > Rage > Twins


Uptimes:
Trinket on side has uptime listed when matched with trinket on top.

[TABLE]X | Feather | Shadopan | Spark | Rage | Twins
Feather | X | 26.53% | 26.87% | 26.40% | 26.54%
Shadopan | 20.44% | X | 20.44% | 20.45% | 20.43%
Spark | 22.43% | 22.39% | X | 22.41% | 22.43%
Rage | 56.12% | 56.02% | 56.77% | X | 56.05%
Twins | 37.87% | 37.96% | 38.30% | 38.25% | X[/TABLE]


Edit: Also worth noting that FB with heroic thunderforged feather and shadopan trinket is looking like the highest single target sustained damage spec of any class in the game.

#29 Mendenbarr

Mendenbarr

    Oh, I never leave home without my party cannon.

  • • Guide Author
  • 181 posts

Posted 16 March 2013 - 08:02 AM

Profile used for trinkets:
Please let me know if you find any mistakes, or can write a program to do this automatically, because it took me way too long to write this.


DW double:
Spoiler




2-hand Double:
Spoiler



DW single:
Spoiler



2-hand single:
Spoiler


#30 fangless

fangless

    Von Kaiser

  • Members
  • 34 posts

Posted 16 March 2013 - 03:07 PM

Notice that you ran the results with the 5.1 stuff, were they just too low to include in that results table?

#31 Mendenbarr

Mendenbarr

    Oh, I never leave home without my party cannon.

  • • Guide Author
  • 181 posts

Posted 16 March 2013 - 09:59 PM

Notice that you ran the results with the 5.1 stuff, were they just too low to include in that results table?


They are in the suggested ranking and full results table, Relic = Relic of Xuen, HLSFO2 = Heroic Lei Shen's Final Orders 2/2, and HDMV2 = Heroic Dark Mist Vortex 2/2. In the OP.

#32 Sanctimony

Sanctimony

    Piston Honda

  • Guild Members
  • 228 posts

Posted 19 March 2013 - 04:49 AM

Currently, razorfrost in the mainhand and fallen crusader in the offhand may be optimal, due to what I suspect is a bug: in this configuration most razorfrost damage procs are doubled. You can see this pretty clearly here, during a fight in which I runeforged as described, and Necrus had the swapped configuration. You can also see the doubling clearly by equipping yourself appropriately and hitting a training dummy with the combat log turned on.

I'm not sure if any other runeforges have any weird doubling effects like this, but I don't believe fallen crusader does as our procs over time have been roughly comparable.

#33 Mendenbarr

Mendenbarr

    Oh, I never leave home without my party cannon.

  • • Guide Author
  • 181 posts

Posted 19 March 2013 - 06:47 AM

Currently, razorfrost in the mainhand and fallen crusader in the offhand may be optimal, due to what I suspect is a bug: in this configuration most razorfrost damage procs are doubled. You can see this pretty clearly here, during a fight in which I runeforged as described, and Necrus had the swapped configuration. You can also see the doubling clearly by setting yourself appropriately and hitting a training dummy with the combat log turned on.

I'm not sure if any other runeforges have any weird doubling effects like this, but I don't believe fallen crusader does as our procs over time have been roughly comparable.



My results:
Simc:
RotFC on mainhand sims 400 dps higher.
RotFC being applied by soul reaper leads to a noticeable increase in uptime, from 74.34% to 76.48%.
Razor ice hits for the same amount on either weapon (835), but is cast more often with razor ice on the mainhand, 497.6 counts versus 466.7. This difference happens to be the exact same numner of soul reapers (31), so we can presume that razor ice change happens because of the lack of SR hits. However, in the sim, this still leads to a net 400 dps loss.

In-game:
Note: Not using log as they seem to not be recording every attack. Recount data will hopefully suffice. Including logs anyway in the hope they show the same trend.
10 minutes on a training dummy, no procs of any kind.

RI in offhand:
Details for Mendenbarr - 19-03 00:41 - Divide - World of Logs
728 melee attacks, 134 miss, 594 melee attacks that actually hit.
291 RI hits

Assuming a perfect ratio of offhand to mainhand attacks, we would expect 297 RI, but whether due to a reporting error or more misses on the offhand, we get 291. Aceptable difference.
RI damage averaged 484.
Tooltip says RI does "2% of weapon damage"
My weapon damage is listed on the weapon as 7334-13623, which averages at 10478.5 damage. 2% of this is 209 damage.
My weapon damage is listed on the character screen as 14025-20313 on the mainhand, and 7012-10157 on the offhand. This averages at 17169 and 8584.5, respectively. 2% of which is 172 and 343 dmg, respectively.
Now, being frost damage, RI is buffed by frost's mastery, which for me is 41.41%, bringing it up to an expected 485 damage, right where it actually is.
Interestingly, RI does not seem to be buffed by RI's debuff on the training dummy.


RI in mainhand:
Details for Mendenbarr - 19-03 00:41 - Divide - World of Logs
758 melee attacks, 149 miss, 609 melee attacks that actually hit.
604 RI hits
This is odd

Assuming a perfect ratio of offhand to mainhand attacks, we would expect 305 RI, we get 302*2. Assuming 302 to 305 is reporting or more misses on mainhand, the number of RI has DOUBLED.
RI damage averaged 485.
Tooltip says RI does "2% of weapon damage"
My weapon damage is still listed on the weapon as 7334-13623, which averages at 10478.5 damage. 2% of this is 209 damage.
My weapon damage is still listed on the character screen as 14025-20313 on the mainhand, and 7012-10157 on the offhand. This averages at 17169 and 8584.5, respectively. 2% of which is 172 and 343 dmg, respectively.
Now, being frost damage, RI is buffed by frost's mastery, which for me is 41.41%, bringing it up to an expected 485 damage, right where it actually is.


Additional data:
RI does not change damage despite swapping to normal rotation, while it procs of every weapon based ability, it remains 485 on average.
RI seems to proc once per non-melee weapon based attack regardless of which weapon has the enchant, with the exception of soul reaper, which I was unable to test on the training dummy, but presume it will only proc when it's on the main-hand.

Conclusion:
RI procs on every weapon based hit when the enchant is on the main-hand weapon, but only procs on either mainhand or offhand (not sure which), when RI is on the offhand. This is not reflected accurately in simC, and napkin maths out to be a 566 dps loss.

If you are concerned about this, please go to the following bug report on battle.net, and state "I am experiencing this as well", and anything else you want to add in.
Unexpected RI behavior - Forums - World of Warcraft



Edit: In the mean time, RI on mainhand is ahead by about 100 dps, based of my above napkin math, so I's suggest swapping over to it. Not going to update the OP unless this continues to be an issue for at least a week, I hope it will get fixed before then. You can help by going to the above link.

#34 Mørtisant

Mørtisant

    Glass Joe

  • Members
  • 1 posts

Posted 28 April 2013 - 06:17 AM

I'd like to suggest adding IceHUD + OmniCC to your rune section of mods. I personally feel like its the best because it shows you your runes in the default fashion, except in the lower center of your screen. OmniCC adds visible numbers to the runes and gives off a little blink when it's ready to use. IceHUD is also really good just to have an easy way to see track certain buffs you're interested in tracking, your health, targets health, target-of-target health, debuffs applied to the target, target's cast time and things like that. and if you don't like any of that other stuff, you can just disable it all and use only the rune aspect of it. Also, you could add these macro to your macro section (for those who dual wield)

#showtooltip
/cast Frost Strike
/cast Blood Tap

#showtooltip
/cast Death Coil
/cast Blood Tap

These macros make it so that every time you cast frost strike, it uses any blood charges u have to give you runes in much the same fashion as RE does, except you get death runes instead of those pesky unholy runes. The Death Coil macro is just for those few times that you might have to be out of range and you're spamming death coil and you don't want to cap on blood charges. For those who insist on micro managing their blood charges themselves, a very useful macro would be

#showtooltip
/cast Blood Tap
/cast Blood Tap

This one allows you to get 2 blood taps in with the same keystroke if you have enough blood charges. This works because Blood Tap is off of the GCD.

#35 zombeez

zombeez

    Glass Joe

  • Members
  • 1 posts

Posted 03 June 2013 - 08:15 AM

These macros make it so that every time you cast frost strike, it uses any blood charges u have to give you runes in much the same fashion as RE does, except you get death runes instead of those pesky unholy runes.


Doing that makes you miss out on one of the biggest advantages of blood tap: saving up charges for when you need them most (jikuns primal nutriment for example)

#36 Mendenbarr

Mendenbarr

    Oh, I never leave home without my party cannon.

  • • Guide Author
  • 181 posts

Posted 12 June 2013 - 05:44 PM

Set bonus math:
Frost:
DW:
2-set sounds interesting, but hell if I know what "200% of hnd +4" means, so we won't be mathing that out quite yet. We can do some napkin math on the 4-set though. Pillar of frost increases rune regeneration speed by 100% while active, which is 20 seconds of every 60 seconds, or an average increase of 33% to rune regen. With my modest 5k haste as DW, that leads to a base regen time of 8.88 seconds. Over the course of a minute I would normally regenerate 6.756 runes of each type, ignore outside haste buffs or t5 procs. This change brings that up to 9.006 runes of each type, an increase of 2.250 runes, or 6.75 more runes total, every minute. Over the course of a 10 minute fight, this leads to an extra 67.5 more runes. 6.75 more runes basically means 7 more HBs and 4 more FSs, speaking as DW, per minute. That would be a 22k dps increase, or about 10-11%. But, DW is already very close to GCD capping, and the fact that this extra rune regen is in small bursts of 100% means the majority of this boost will be no-doubt wasted. So how many of the 11 extra GCDs could we fit into our rotation? There is room for 1.72, according to the sim, with perfect play, which makes this about a 1.5% dps increase total. But, we can also cut down on lower priority abilities, like HoW. You can skip one HoW during the course of pillar, so let's say one extra GCD. That brings us up to being worth about ~2.5% for DW frost with a low 5k haste, and without AMS soaking, and without lust or encounter based haste boosts. More haste or AMS soaking will quickly drop this bonus even lower, making it a very meh bonus from a power perspective. On another note, we'll be wasting about 75% of the runes we regen, which is going to suck from dk mindset; don't waste resources. I hope this one receives changes in terms of DW frost.

2-hand:
Running with a nice 11.5k haste build on simc for numbers. That gives us a base rune regen of 7.87. That leads to 7.62 runes per minute. With the 4-set, that becomes 10.16 runes per minute. That's a gain of 2.54 runes, or 7.62 runes total. This basically means 4 more obliterates and 4 more frost strikes. About 15k more dps, or 7.5% more. How much room does 2-hand have for waiting? A little more than DW, even with double the haste, expected because of using obliterate instead of HB as the primary attack, 2-hand has about 1.89 free GCDs per sitting, and only needs to fill 8. Adding another 1 for HoW puts us up to 2.89/8, or about ~2.7% increase. 65% of the regen is still wasted, and remember, the gain will only get lower with more haste, lust, or AMS soaking.

<Midwinter> - 14/14 HM -25m - LF exceptional raiders
http://www.mwguild.net/
DK Theorycrafting/Game Dev Blog
DK Gear/log reviews/audits


#37 Jeges

Jeges

    Von Kaiser

  • Members
  • 59 posts

Posted 12 June 2013 - 06:16 PM

Let's look at 2-handed Frost again, since I think you're undervaluing the bonus, mendenbarr: by prioritizing those extra Obliterates above Frost Strikes and Rime procs, and the extra Frost Strikes above Rime procs, you can use all of them. Each Obliterate (215k) that replaces a FS (106k) adds about 109K damage, and each FS that replaces a HB (91k) adds about 15K damage. In the 20s of Pillar, you'd ordinarily have 7.38 FS, 2.48 HB, 0.83 HoW, 4.57 Ob, and 1.36s free (strictly based on intervals from simc; I'm not sure where you got 1.89s free? I did 30.6/450*20). All of that assuming no forethought given to planning resources around Pillar, and glossing over the other things that the average intervals include (disease management, soul reaper). We can replace the 2.48 HB, 0.83 HoW, and 1.36s free (total 4.67s) with 7.62/2 = 3.81 Ob and 0.86 FS. That's a win of 819k from Ob and 91k from FS vs. a loss of 226k from HB, for a realized gain of 11.4k DPS. This doesn't include the drop in Ob values from fewer of them being KM procs, nor the fact that the positive effects of your resource glut will extend past the 20s window of Pillar, nor the fact that the glut comes when Pillar is up and so it'll actually be even more of a bonus. Disregarding all that, it looks like about a 6% gain (with potentially more upside than downside due to the bias of my glossing, and the possibility that it might be enough of a game changer to shift reforge priorities).

#38 Mendenbarr

Mendenbarr

    Oh, I never leave home without my party cannon.

  • • Guide Author
  • 181 posts

Posted 12 June 2013 - 10:00 PM

Right you are Jeges, you can simply spam oblit as 2-hand frost with pillar active. It doesn't make much of a difference to DW as HB and FS, and oblit are all petty close in DPET, but it will make a difference for 2-hand frost. I got 1.89 seconds from the 14.2 wait time/7.5 minutes in the 2-hand 541 sim profile. To re-do my calculations:

2-hand:
Running with a nice 11.5k haste build on simc for numbers. That gives us a base rune regen of 7.87. That leads to 7.62 runes per minute. With the 4-set, that becomes 10.16 runes per minute. That's a gain of 2.54 runes, or 7.62 runes total. This basically means 4 more obliterates and 4 more frost strikes. About 15k more dps, or 7.5% more. How much room does 2-hand have for waiting? A little more than DW, even with double the haste, expected because of using obliterate instead of HB as the primary attack, 2-hand has about 1.89 free GCDs per sitting, and only needs to fill 8. Adding another 1 for HoW puts us up to 2.89/8. Existing rune overflow is low for 2-hand frost, and therefore we can gain all 4 obliterates, still wasting no runes, at the expected cost of 1 rime proc, our lowest DPET after HoW. We will not gain any km procs though, so taking non-crit oblits at 151k, and multiply by base crit chance of 22% gives us 184k per oblit. 737k total. The one rime proc is a loss of about 95k, bringing us down to 642k, or about 10.7k dps. This is boosted about 10% stronger than the average obliterate damage by occurring when pillar of frost is active, for a total of 11.77k dps, or 5.85% dps boost.


50% of the regen is still wasted, but the gain will arguably get higher with more haste/lust, as it let's you get off more oblits during the time window. You will waste 100s of runic though, and as a dk, it sucks to waste so many resources.

<Midwinter> - 14/14 HM -25m - LF exceptional raiders
http://www.mwguild.net/
DK Theorycrafting/Game Dev Blog
DK Gear/log reviews/audits


#39 blessed

blessed

    Glass Joe

  • Members
  • 7 posts

Posted 14 June 2013 - 02:23 AM

Frozen Power only last for 10 seconds, don't know if it's intended or will be fixed to full duration of PoF, but for now - it's only halftime uptime on it, so I guess the bonus is even smaller :(

#40 Italiandk

Italiandk

    Glass Joe

  • Members
  • 11 posts

Posted 18 June 2013 - 09:24 AM

DW
What about changing the rotation while under PoF? I mean if you loose a lot of the 4p T16 potential dps doing the traditional rotation, what about using more OB to open some GCD?

2H
You take 11.5k haste as a base number to make sims, but if a good amount of resources under PoF is going to be wasted what about dropping a bit of haste?




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users