TLDR
I’m DW Frost. Did one week of Chimaeron stacking Mastery (18.86 Mastery with 8.5% Haste from Rating) and one week stacking Haste (11.29 Mastery with 16.63% Haste from Rating). Three gear changes of equal ilvl were made to stack more of my desired stat.
Fortunately enough, both parses were 254 seconds in length. Mastery raid comp included a Destruction Lock, Haste had a Moonkin.
Mastery Parse
World of Logs - Real Time Raid Analysis
19972 DPS
Haste Parse
World of Logs - Real Time Raid Analysis
21580 DPS
After trying it out myself I am going with Haste > Crit > Mastery for DW frost as indicated by SimCraft.
Haste Versus Mastery
When comparing these two stats it really comes down to frequency of rune regeneration versus a raw damage modifier. As such we can ignore crit and analyze the frequency of attacks and the average non crit hit of each attack to determine which stat is better. When calculating frequency of attacks, I ignored misses for Rune Attacks since you get the runes back. For frost strikes and melee swings I included everything.
For AVG damage I added MH and OH attacks where applicable.
Mastery
| Avg Freq |
| OB 25805 6.20 |
| Melee 3539 0.87 |
| FS 17413 4.31 |
| HB 18882 13.37 |
| FF 3442 3.10 |
| BP 2194 3.10 |
| BS 8510 12.10 |
| PS 6725 50.80 |
| KMPPM NA 7.32 |
Haste
Avg Frequency
| OB 25334 5.18 |
| Melee 3393 0.78 |
| FS 15168 3.74 |
| HB 15705 12.10 |
| FF 2826 3.02 |
| BP 2236 3.06 |
| BS 8448 13.37 |
| PS 7023 63.50 |
| KMPPM 8.27 |
Any average differences in nonfrost attacks will be attributed to RNG. Keep in mind the loop between more rune attacks and more frost strikes/hb. Faster obliterates lead to more frost strikes which leads to more obliterates to more frost strikes etc. The actual interaction is more complex than I want to try to tackle with napkin math.
When moving to Haste from Mastery we observe:
Obliterate – 19.96% Increase in Frequency.
Melee - 11.53% Increase in Frequency
Frost Strike – 15.24% Increase in Frequency, 14.80% Decrease in Avg Damage
Howling Blast – 10.49% Increase in Frequency, 20.22% Decrease in Avg Damage
Frost Fever - 21.79% Decrease in Avg Damage
KM Procs – 12.97% Increase
Blood Plague, Blood Strike and Plague Strike – not a significant enough portion of total damage to worry about.
With the wash between Frost Strikes you are looking at +20% OB Damage and +11% Melee damage versus a loss of 9% Howling Blast Damage (1.2022 Dmg/1.1019 Freq) and a 21% increase in Frost Fever Damage.
Using the percentages of total damage from the Haste Log in the first column and modifying the second column to account for the above differences gives haste a clear advantage. I even gave frost strike a +20% damage modifier like HB and FF instead of the observed 15%
| Haste Mastery |
| OB 33.2 27.7 |
| Melee 17.0 15.3 |
| FS 25.7 25.7 |
| HB 8.5 9.4 |
| FF 5.0 6.0 |
| Sum 89.4 84.0 |
Crit Versus Mastery
I’ll be using my data from the Haste Heavy Parse and mathematically adding 5% crit and 5 mastery (10% Frost Damage).
Crit Rates will be treated as a percent damage modifier such as 38% crit = (0.38 * 1.03) + 1 = 1.3914 Damage Increase. The extra 3% is the crit meta.
TLDR:I’m only adding 4.75% crit to OB and FS to account for KM critsIn the cases of FS and OB I am applying crit at 95% effectiveness to account for KM. This is derived using the following logic. Splitting my 8 KM procs per minute evenly between OB and FS and taking into account around 11 OB Per Minute and 16 Frost Strikes a Minute we have OB with 36% crit and FS at 25% crit regardless of crit rating. Using 10% crit from rating we have Obliterate at 15% crit (LotP) and FS at 20% crit (LOTP + 2 PC Bonus). Using a one roll system you would expect to get waste a normal crit on OB 5.4% of the time and a normal FS Crit 5% of the time. As you increase your crit from crit rating you are more likely
/TLDR End
Doing the math with the actual results and adding 5% crit in the 2nd column and 10% Frost Damage in the 3rd column gives crit a very slight advantage on a single target fight as indicated by SimCraft. It is clear to see that given more baseline mastery on our gear/or the 4.1 changes could easily throw the balance toward mastery.
| AvgDPS 5% Crit 10% Frost |
| 6634 6873 6634 |
| 4754 4978 4754 |
| 5168 5367 5685 |
| 1681 1747 1849 |
| 1003 1051 1104 |
| 803 841 803 |
| 769 802 769 |
| 111 116 111 |
| 20922 21775 21707 |
It’s glaringly obvious that Consider didn’t take something into account or simply did some math wrong when he calculated the stat weights for DW. The assertion that you are haste capped at very low numbers is simply not true although it is easy to see how you could conclude this on paper.
For simplification assume the 4.1 setup of all death runes. Given 10% haste you should have your runes recharging in 9 seconds. Use three obliterates in 4.5 seconds and you have enough RP for three Frost strikes in the next 4.5 Seconds. 4.5+ 4.5 = 9 and you are magically GCD locked without Rime or Runic Empowerment procs. Something is going on, be it latency or Rune Pushback from HB/Outbreak that is allowing you to not be capped at very high haste levels.
I can tell you from experience that with 120 Max RP and 16.5% haste I am not GCD locked if I don’t miss any rune attacks. On heroic attempts prior to the normal kill, I had some runes sitting from missing some OBs, but on the actual normal mode kill I got lucky and didn’t miss any OBs. I will be taking a point from butchery and throwing it into 130 Max RP in the future though.
Once again I can see how Consider reached his GCD lock conclusion, but anyone in blue gear can run out and stack 1200 + haste and prove this point wrong.
I am sure the first thing anyone will throw out is how this is “anecdotal”. I am a statistician working in medical device manufacturing. I know better than just about anyone that 2 data points doesn’t mean squat statistically; however, ask yourself this. Did I really just get lucky and move from 27th on the charts (where my Mastery parse would be if I didn’t bump it off) to number 1 by stacking the worst rating for DW DKs. Also note that I beat the next DW parse (#8 at the time of this writing) by around 5% (1k DPS).
Consider’s calculated stat weights for DW simply don’t hold water in the real world while SimCraft’s suggestions netted me an 8% increase. In my opinion,the haste interaction with rune regeneration and Runic Empowerment in play is simply to complex to map out with simple mathmatics. I am sure Simcraft isn’t perfect but a program that actually does some interations is simply going to do a better job than a guy with a spreadsheet when doing things like this.
In any case, stack haste and enjoy the 5-10% buff while also getting the 10% from the PTR changes. Happy Killing!
P.S. If you do haste stack you need points in runic power mastery. I did around 10 minutes on target dummy and had 14700 DPS with 0/3 and 15900 with 3/3.