I am talking about real world dps situations. I've run this spec without problems through all of BWD (sans Nef), all of BoT, and through Council in 4W. Using 2/3 epidemic still gives you 2 full seconds of wiggle room in your rotations, and you're not clipping any disease cooldowns. Yes, the timing is slightly tighter than 3/3 epidemic, but it's not a machine-perfect sim-only timing situation; this is very doable. Once you've gotten used to the thirty second refresh rhythm, I think you'll find that keeping BP up is easy and reflexive with 2/3 epidemic, and you'll enjoy the extra dps that free talent point will get you. 3/3 epidemic has odd intervals of disease refresh and leads to lots of clipping in my experience - 2/3 is standard 30 second intervals which match much of the timing in WoW boss fights.

All in all I'm seeing more benefit from 2/2 butchery, 2/3 epidemic than from 1/2 butchery, 3/3 epidemic.


Last note, I see that you're DW, while I'm 2H - having faster GCDs due to unholy pres probably makes it easier for 2H frost to refresh BP as needed than the more GCD constrained DW spec. Perhaps 2/3 is only worth it for 2H spec?

So the question is what is the PPM of killing machine or is it percentage based? If it's percentage then maybe we would want more hit. I guess it's time to play with gear again....

The tool-tip for the talent reads as follows: "Your melee attacks have a chance to make your next Obliterate or Frost Strike a guaranteed critical strike. Effect occurs more often than Killing Machine (Rank x).

This sounds like it would have to be percentage based because more points in the talent means more chance to proc.

I am in the process of trying to gather some real hard data that I can post as follow-up to my original post. I should have it posted later this evening.

It's 5 PPM, it has always been 5 PPM and always will be 5 PPM. Yes it will proc more often the more haste you haste you have, which one of the reasons it's such a good stat.

I am posting this only because this information is not in the OP right now (to my knowledge), so please don't launch us into unnecessary pages of discussion for a mechanic that we already understand.

1) Off hand weapons do not trigger KM procs, have not since around patch 3.2.
2) You are not counting the 20% haste bonus of Frost Spec and the 10% from IIT.
3) With this in mind your data supports the already known value of 5 PPM.

Thanks for the response Darkside. I did indeed find an old blue post (2009) regarding off hand weapons not procing KM. I would imagine that this has not changed.

I'm just back to square one on my DPS increase theory. My DPS is consistently higher with higher hit rating. KM was the only way I could explain it.

Has anyone actually tried this themselves? I noticed that one of my posts had gotten copied over to the official WoW forums and it had quite a few of comments. I'm just curious now if anyone has had the same experience.

KM can explain it. Whether it does explain the difference is another question, but let's clear something up first. You seem to misunderstand exactly what is meant when someone refers to PPM, which is presumably why Darkside stepped into the debate in the first place.

PPM doesn't refer the number of times a minute a proc would happen. It's the number of times a proc would happen without haste, misses, or special attacks (in the case of those PPM-based procs which can proc off of special attacks). Haste and specials will increase the number of procs, and misses will decrease them. For example, a rogue with an item a 1 PPM proc chance and no cooldown (like a DMC) would likely expect 3 procs per minute on average, between haste, SnD, and special attacks (hat tip to Adriana's example in the Rogue forums).

In the case of Killing Machine, only hit and haste will come into play. From your examples, you're increasing from an approximate total hit rate of 86% in the mastery example to about 91% in the hit example. This leads to about 5.6% more MH hits, thus 5.6% more chances for KM to proc. This may or may not be enough to explain the seeming increase in damage (10 short tests doesn't really mean anything, particularly with a spec as RNG dependent as frost. That you've received similar simulation results is potentially interesting, however).

Based on your tests, the observed number of procs was about 7.7% higher in the hit-based than the mastery based fights, which is statistically indistinguishable from the 5.6% you'd expect from the hit rating change alone due to the very small number of trials. Two points of note. First, I didn't strip out any of the tests, since we have no reason not to. That a five-minute test with the lower-hit reforge yielded the single highest number of procs of KM between all your tests is a completely reasonable result, and we ought not prune it as an outlier, particularly not in a sample size of 5. This is important because it points out the underlying variability of the spec. Second, it's also worth noting that not all additional KM procs will be spent on Obliterate, and some number will likely be spent on Frost Strikes, which are now dealing a significantly lower amount of damage due to the lack of mastery. Following from this, it's also likely that player variables (most notably the efficiency of KM proc use) will factor into this hugely.

Pretty clearly with napkin math, the hit increase causes a smaller white damage increase than the loss of mastery causes in frost damage increase. Hit may have a better conversion rate, but Frost DK do a good deal more damage with frost attacks than with melee swings. The increase in KM procs might be enough to counteract that loss, but napkin math or dummy tests can't seal the deal. The complexity of the relationship will require simulation.

Things to keep in mind: [1] Don't just report the number of procs, but the percentage of total damage done by each ability. The value of mastery and of white swings can be pretty clearly assessed from how large a portion of a DK's total damage they are. [2] Try to be better in reforging. The mastery-focused version is still significantly over the 5% breakpoint and the fact that the crit rate is significantly different in the two examples blurs the waters when trying to isolate the hit/mastery trade-off. [3] It isn't much harder to run numerous sims, so presenting intermediate stages between minimal hit and higher hit ratings would be a good idea. [4] To that end, checking different gear sets in both high hit and high mastery situations. There are several class/spec combination for which stat weights are very flexible and slight increases in one stat can radically increase the value of another, causing the relative stat weights to flip.

edit: completeness

I'm a little confused, if you are at 0% miss with main hand weapon, how can adding more hit increase the number of procs of Killing Machine? Aren't you already not getting any misses on the attacks that can trigger KM at that point? I was under the impression that above 8% hit the only thing gaining benefit from further hit is offhand auto-attacks.

While dual wielding, both hands have a 24% chance to miss (vs a level 85 mob) with 0% hit.

At 8% hit, both hands have a 16% chance to miss, and like a 19% chance to miss vs a boss mob (27.2%, I think?)

Hit applies evenly to both weapons. When most speak of hit cap here, they are speaking of "yellow" hit cap, or the point at which special attacks will not miss. This is 8%.

White attacks, however, when dual wielding have a miss rate of 27%. With 8% hit, you still miss auto attacks 19% of the time. This means nearly a fifth of your white attacks will not connect, and assuming that KM only procs upon hit, that you will lose a full fifth of KM procs triggered by your auto attacks. The current discussion is exploring if the additional DPS gain from recovering those "lost" KM procs, combined with increased white damage is enough to make hit more valuable then mastery. While I haven't seen anything definitive, initial tests show there is some merit to this claim. Many of us are interested in more rigorous testing and simulation on the subject, as the reported gains, if real, could significantly improve performance.

Mynou, The list looks pretty solid. I would replace Crushing Weight for Impatience of Youth/Mirror trinket from Baradin rep. Impatience has an amazing amount of mastery on it that can easily yield a great reforge if needed plus on on-use strength boost to line up with a CD. Mirror is the same, only granting 500 resist to all schools which is definitely a great situation trink to pop for extra survivability.

I'd definitely rate Impatience of Youth higher. DMC:Hurricane is amazing as well, but with as much RNG that frost brings, I like the prospect of a trink I can manually proc so I know I'm at least getting extra OOMPH with a CD macro.

Having a guaranteed proc like that is also good for fights where you might spec into Chillblains or something else situational that may take points out of DPS-gain talents.

That's my two cents.