Pages

Monday, September 15, 2014

The Power to Choose




My last post illustrated all of my reasons why a playstyle with downtime benefits from being played skillfully, and why it can be a fun and interesting playstyle. However, the fast-paced style of today also benefits from being played skillfully, and it also can be a fun and interesting playstyle. While I personally prefer downtime, the message I seem to have failed to convey in my previous post was that neither downtime nor no-downtime are objectively bad.

Subjectivity is a whole other matter. Many players picked DWF for it's current fast-paced playstyle, which they have enjoyed for a long time, and is familiar to them (while mastersimple is a recent invention, low downtime has been a spec feature for as long as I can remember). To lose that would be to surrender an important, subjective reason for playing the class. On the other hand, countless others dislike the current playstyle, and see the potential of downtime to be the playstyle they have always wanted. Both groups of players feel like they have something to lose, and are fighting to defend their playstyle of choice.



Choice is at the heart of this matter. The choice of playstyles is nothing more than that, a choice. A choice is a powerful thing, and perhaps an active choice will save half of us from losing our playstyle of choice. Recently, Celestalon and whomever writes the patch notes has been promoting rotational choices as a good gameplay option, and I believe this certainly qualifies.

"For Energy-based gameplay to function well, the primary limitation on ability usage should be Energy, not time. That breaks down when a rotation becomes limited by available GCDs, rather than by Energy. Windwalkers were hitting this “GCD-cap” too easily, causing scaling problems as they geared up, and removing rotational choices." - http://eu.battle.net/wow/en/blog/13423478/warlords-of-draenor%E2%84%A2-beta-patch-notes-august-19-19-08-2014#monk (Windwalker changes)

"any reason ret loses INQ and gain massive dps increase from Seraphim but ferals still have to SR in their rotation? Seems weird"
"Seraphim is a choice. Savage Roar can be glyphed, another choice." - https://twitter.com/Celestalon/status/505373892287160320

"Now then... Yes, the Marksmanship and Survival rotations are simpler than they were before. The goal is to have a simpler baseline rotation, and have meaningful choices that allow you to layer more depth onto that in talents, if you choose.
Hunters have a ton of talents which grant rotational or short CD abilities, to support this design. However, they were lacking in rotational depth. The gameplay of most of them boiled down to "use on cooldown". No need to think about preparing for when you're going to use it, or how it interacts with your rotation, etc. So, we're going to make some tweaks to try to add some rotational depth to several of the talent choices." - http://us.battle.net/wow/en/forum/topic/13978297949?page=3#60

Adapting a low-downtime and high(er) downtime talent seems like a reasonable solution to give death knights a rotational choice between the 2 distinct playstyles.

Here is an example of some talents that would do the job:


Plague Leech:
Draw forth the infection from an enemy, consuming your Blood Plague and Frost Fever diseases on the target to activate two random fully-depleted runes as Death Runes. No Cooldown. Also increases rune regen by x%.

Creeping Infection:
Increases all damage done by x% while your diseases are active on at least 1 target, further increase your disease damage by x%, and allows your diseases to benefit from haste.

Soul Pox:
Victims of your frost fever or blood blague are also afflicted with soul pox. Soul pox deals X shadowfrost damage every 3 seconds and increases the damage the target takes below 35% by x%.


There are many other possible solutions, such as locking a playstyle into a weapon choice, (DW remains fast paced, 2H becomes the downtime spec), any number of talent choices I haven't considered, or simply the choice between gearing towards haste vs. another stat. All of these choices have a problem though, as does the choice philosophy in general. 

One choice will be mathematically better than the others. I feel pretty comfortable saying that it would be impossible to balance the choice to be exactly equal, and that even getting close would require a significant amount of developer time, which in turn might get other things cut. 



There is also the issue that each potential solution would lock you into other effects of a talent talent you dislike, for example, hating PL and still wanting no downtime. The problem with making a weapon division is weapon availability is outside of most player's control, at least at the start of progression when it matters, so it's not really a choice at all. Leaving the problem alone would leave a lot of death knights unhappy, whether they leave downtime or no-downtime as the optimal choice. 

If they could balance 2 (or, titans forbid, 3) talents close enough together, then you could choose to either gain or lose a miniscule amount of damage to go with your prefered playsytle, and I believe that many players would happily make that choice. 

So my question to you is, do you see a solution I've overlooked? If not, how would you solve this problem? If they were to provide a talent option, how close would it have to be to work for you? Would you give up 5% DPS to play your preferred style? 1%? 0.05%? How close does it have to be. 

Or share your talent ideas! 

Monday, September 8, 2014

Downtime in Warlords of Draenor

What is downtime?




Downtime is time during your rotation during which there are no buttons to press. You don't have the runic for Frost Strike or Death Coil, you don't have any runes at all you want to spend, no killing machine or sudden doom procs, no blood tap charges, and even horn of winter is on cooldown. You can't do anything.

Downtime is commonly represented with a percentile value. So, a downtime of 10% would mean that 9 times out of 10, you are pressing a button, and the rest of the time you're not pressing a button. The sample profiles for 588 gear have 0.14% downtime, 0.44% downtime, and 13.52% downtime, for DWF, 2HF, and unholy, respectively. For frost, that means you are nearly never out of things to do, and it promotes a spammy play style, in which the goal is to always be hitting something, and hitting the wrong thing is better than hitting nothing at all. Unholy has a bit more downtime, but is still active a vast majority of the time. This is what we are used to, and what most of us are comfortable with.



Warlords


With 6.0, everything we know is changing. The sky is falling! Here are the changes that affect downtime:

Frost:

  • Raise Dead removed, adding 1 free GCD every 2 minutes
  • Horn of Winter no longer generates runic power, adding a free GCD every 40 seconds
  • Frost Strike now costs 25 runic instead of 20, significantly increasing downtime
  • Frost now gets 5% passive haste from icy talons, reducing downtime
  • Anti-Magic Shell soaking has been reduced by about 90% compared to live
  • Despite costing more, Frost Strike now procs rune regen talents less. 
  • Pillar of Frost no longer costs a rune, taking away a GCD every minute to spend that rune


Unholy:

  • Unholy Frenzy has been removed, significantly increasing downtime
  • Horn of Winter no longer generates runic power, adding a free GCD every 40 seconds
  • Death Coil now costs 30 runic instead of 32, slightly decreasing downtime
  • Anti-Magic Shell soaking has been reduced by about 90% compared to live
  • Dark Transformation no longer costs a rune, taking away a GCD every 30 seconds to spend that rune
These changes, along with the gear reset, ends up being about 20% downtime for all 3 specs, in the 660 pvp gear. 


Is downtime a good thing?



All good things in moderation. Downtime can be good, as it has many benefits attached to it. For example, downtime rewards pooling behavior, as you can control what portions of any particular encounter have downtime, and which portions don't. You might hold onto some runes prior to the ball lightnings on Lei Shen, instead of letting the downtime happen while the ball lightnings are up. Perhaps you need to run around a bomb on Blackhand, or dodge a fire on the Twin Ogrons, your runes keep recharging during those periods of time. In those ways, downtime both allows you to hold onto damage for periods of time in which you need it (adds, damage taken buff, priority targets), and let's you turn short periods of inactivity into a DPS gain when you do get back in range.


Downtime also has particular significance for frost, due to killing machine. You can choose when you want the downtime to happen, and you can use the downtime as a resource to wait for killing machine, with the resources to use either Frost Strike or Obliterate, depending on your weapons, when that killing machine does proc.


Here is an example of the playstyle of heavy downtime 2H Frost: You open up with 3 obliterates and 2 frost strikes, and suddenly you have nothing to do. You can twiddle your thumbs, wait for the next set of runes, and use them up instantly, or you can use this new downtime to your advantage. As the first new set of runes comes up, you don't spend it, but instead keep one eye on your melee swing timer, and wait for the opportune moment to strike. Killing machine lights up, and you react by hitting your banked obliterate, instantly using up the proc and scoring a sizeable hit on your foe. You hit a frost strike immediately after with the new runic power, to minimize the chances of that ability using up one of your precious killing machine procs, and then you use up the rime proc that was generated from your last obliterate. Another set of runes is already up, and you wait for your next killing machine. Bam Bam Bam, 3 killing machine procs in a row, luckily you have 3 sets of runes. You use them all up, doing a massive damage spike to your target. A fourth killing machine procs, and you activate plague leech, despite there still being 6 seconds left on your diseases, and hit obliterate once more. This lucky streak is followed by a drought of procs, and as you watch your runes bar fill up more and more, eventually you can wait no longer, and spend the runes immediately prior to them capping. After you're no longer in danger of capping, you keep holding on to the extra runes, waiting for more opportune moments.



I've been testing this playstyle for every raid testing boss available, and for several hours on dummies, and I'm finding it both more skillful and more fun than the current spammy rotation. The same principles described above apply to DWF, spending most of your time at 70-90 runic, waiting for that inevitable lucky streak of killing machine procs to eat up all of your runic, and reward you for your patience.

However, the outcry of players claiming that they hated "watching movies while playing", has put a very negative connotation on downtime. The 5% haste added to frost since the first iteration of the frost strike cost change which made this possible has forced us into playing spammy more often than not. Whatever some people might say, the developers do listen to us. I encourage you to try out the downtime playstyle for yourself, and report your findings in the beta forums or to Celestalon over on twitter. Even if you don't have beta, let blizzard know how you feel about the idea. Even if you disagree with me. It's my hope that the downtime on launch is closer to 30-40%, as 20% downtime on launch will end with the same downtime as we have now by the end of warlords.

And that makes me sad.


Future posts


Some of you may have noticed a lack of recent activity on my part. This is because there really isn't that much to say yet, with the constant new builds everything could change again in an instant. As we approach launch, you should see more and more content on this website, and I hope to have all of the relevant information available a week or 2 prior to launch. I still hope to go to blizzcon(the week OF launch), and if I do, I have a special surprise for everyone attending or watching from home, but I still have to rely on your generosity to get there.

And if you have an idea for something you'd like to see a post on while we wait for the numbers to settle, feel free to drop a comment or send me a tweet!

Saturday, June 14, 2014

How Theorycrafting Works; A look at the Warlords of Draenor level 100 talents

I've been asked a few times how to math, and how to get started with theorycrafting, and so today we'll be taking an in-depth look at the process. The goal today is to answer the question, "Which level 100 talent should I pick?"

Step 1:
Let's begin with finding out all the information we can. An excellent starting point is the spell data, which we can review over at the wowdb database. Let's copy and paste the 3 talent choices:


Breath of Sindragosa Death Knight - LvL 100 Talent
15 Runic Power, plus 15 per sec
Instant 60 sec cooldown
Requires Death Knight

Continuously deals [ 1 + 39.96% of AP ] Shadowfrost damage every 1 sec to enemies in a cone in front of you within 12 yards, and affects them with Mark of Sindragosa for 6 sec. You will continue breathing until canceled or Runic Power is exhausted.

Mark of Sindragosa causes you to be healed for 5% of spell damage dealt by afflicted enemies.



Defile Death Knight - LvL 100 Talent
1 Unholy 30 yd range
Instant 30 sec cooldown
Requires Death Knight

Defiles the ground targeted by the Death Knight. Every 1 sec, if there are any enemies standing in the Defile, it deals [ 1 + 17.32% of AP ] Shadowfrost damage to them, and grows in area and damage by 2.5%.

Enemies standing in your Defile deal 10% less damage to you.

Replaces Death and Decay.



Necrotic Plague Death Knight - LvL 100 Talent
Requires Death Knight

A powerful disease that deals [ 1 + 4.66% of AP ] Shadowfrost damage per stack every 2 sec for 30 sec. Each time it deals damage, it gains 1 stack, and infects another nearby enemy within 8 yards if possible.

You gain 5 Runic Power whenever a target infected with Necrotic Plague attempts to attack you.

Replaces Blood Plague and Frost Fever, and is applied by any ability which applied either.



Some of this information we already know, but we'll leave it alone for now and deal with it in step 2. Now, in order to determine what the net gain of anything is, we'll need to know what the loss is, so let's go find the spells that these talents will compete with, and copy their spell data. Starting with Breath of Sindragosa, we see that it costs runic power, and will therefore stop us from being able to use our primary rune dump, so let's go find out the spell data for those dumps. 


Death Coil
30 Runic Power 30 yd range (40 for friends)
Instant
Requires Death Knight

Blood, Frost:

Fire a blast of unholy energy, causing [ 127.2% of AP ] Shadow damage to an enemy target or healing [ 269.9% of AP ] damage on a friendly Undead target.

Unholy:

Fire a blast of unholy energy, causing [ 152.6% of AP ] Shadow damage to an enemy target or healing [ 269.9% of AP ] damage on a friendly Undead target.



Frost Strike Death Knight - Frost Spec
35 Runic Power Melee range
Instant
Requires Melee WeaponRequires Death Knight (Frost)

Instantly strike the enemy, causing 75% weapon damage as Frost damage.


Next we'll see that defile replaces death and decay, and therefore competes with unholy rune spenders. Let's start by assuming that death and decay is worth using, and so all we really need to know to figure out how defile is worth it death and decay. If death and decay is not worth using and defile is, then we'd need to instead calculate the benefit of using defile over SS for unholy, either half a death strike (if it's worth spending runes on it) or a death and decay/pestilence(whichever is better, if it's worth using on a proc of CS). These are a lot of questions, and work that we can defer until later by assuming that death and decay is worth using, like it is now for the most part.


Death and Decay
1 Unholy 30 yd range
Instant 30 sec cooldown
Requires Death Knight

Corrupts the ground targeted by the Death Knight, causing [ 6.93% of AP ] Shadow damage every sec to targets that remain in the area for 10 sec.

Lastly, we have necrotic plague, which replaces diseases, so let's grab the spell data for diseases. The disease tooltip doesn't have an AP value, which we'll need, but the aura does, so let's copy and paste that instead. 


Blood Plague Disease

Suffering [ 19.7% of AP ] Shadow damage per 3 sec.
30 seconds remaining



Frost Fever Disease
Suffering [ 19.7% of AP ] Frost damage every 3 sec.
30 seconds remaining



Step 2:
We now have all the information we need to get our first answer. Theorycrafting and math in general involves a lot of numbers, and even more importantly involves a lot of choosing the right number for the right place. Let's grab all the important looking numbers and take a note of them, so we have them accessible for easy reference. Here is an example of the actual notes I took when doing this on my own:

Death and Decay: 6.93% AP as shadow damage x11. 30 second CD. I've added x11 here, as I happen to know that death and decay ticks 11 times. 
Defile: 17.32% AP as shadowfrost damage x11, multiplied by 5% per hit. 30 second CD. Despite the 2.5% value listed on the tooltip, my request for clarification in the theorycrafting thread led Celestalon, one of the lead wow devs, to share this tidbit of information. So the growth rate has been amended to 5%. I've also assumed that defile has the same number of ticks as death and decay, which may later be proved wrong. 

Hit Damage multiplier
1 1
2 1.05
3 1.1
4 1.15
5 1.2
6 1.25
7 1.3
8 1.35
9 1.4
10 1.45
11 1.5
total: 13.75
Here I list out the damage multiplier on each tick of defile, simply adding 5% each time, starting at 100% damage, followed by 105%, followed by 110%, ect. I then add them all together to get the effective total multiplier for a full defile, which is 13.75. 

Breath of Sindragosa deals 39.96% AP shadowfrost damage per second to a cone in front of you. Costs 15 runic, plus 15 per tick. Procs rune regen talents/SI at proportional rates to primary rune dump. This information is also courtesy of patch notes (can't find a link so have one of Magdalena quoting them) and Celestalon. It basically means that we don't need to worry about lost value from not getting rune regen procs when using BoS. 60 second CD
Death coil deals 152.6% AP as shadow damage to the primary target. 
Frost strike deals 75% weapon damage as frost

Necrotic Plague deals 4.66% of AP as shadowfrost per stack every 2 seconds for 30 seconds. Each time it deals damage, it gains 1 stack and jumps to a new target within 8 yards. Replaces both diseases.
Blood Plague deals 19.7% AP as shadow every 3 seconds
Frost fever deals 19.7% AP as frost every 3 seconds

Step 3:
We create the formulas, or equations, that we'll use to answer our question. 

Breath of Sindragosa:

Unholy: 
BoS competes with death coil for unholy. We can see looking at the spell data that BoS deals 39.96% (X) AP damage per tick, and each tick costs 15 runic, plus a getting started cost of 15 per use. We also know that death coil deals 152.6% (Y) AP damage for 30 runic, and dividing that by 2, we find the damage it deals for 15 runic, which is Y/2, or 76.3% AP. Wait a second, death coil deals more damage for the same amount of runic? Why should we pick BoS then? Let's find out exactly how much more damage it deals. 

Here is our general formula, where X is the APCE(attack power coefficient) of BoS, Y is the APCE of DC, T is the number of ticks, N is the number targets, and I is the interval between uses, which will be the cooldown when it's used on cooldown. 

((Breath of Sindragosa Attack Power Coefficient * Number of Targets - Death Coil Attack Power Coefficient/2) * Ticks - Death Coil Attack Power Coefficient/2)/ Time interval = Damage per second, in terms of Attack Power * Mastery

((X*N - Y/2) * T - Y/2)/I = DPS

Let's break it down. Y is the damage that DC does, in terms of AP. If we divide that by 2, we find out what the damage of half a DC is, which happens to cost exactly the same as a BoS tick, 15. If we then subtract that amount from the damage that BoS deals for 15 runic, X* H, we find that X - Y/2 is the damage gained(or lost) per tick. We then multiply this by ticks T to find out the total damage dealt per use. We also need to account for the cost of activating BoS, which is 15 runic, or half a DC, or Y/2 per use, so we subtract that. If we then divide by the interval, we go from damage per use to damage per second. That's really all there is too it. 

Frost: 
Very similar to unholy, but with a few major differences. We still want to take BoS's damage, which hasn't changed (much, I'll get to that in a second), and subtract the damage we would otherwise lose. However, in this case the damage is lost to frost strike instead of death coil, which means we need to do 2 more calculations. Firstly, frost strike damage is in weapon damage instead of AP, so we need to know how they convert. We can find our answers at the WoD patch notes, where we see that "Attack Power now increases Weapon Damage at a rate of 1 DPS per 3.5 Attack Power (up from 1 DPS per 14 Attack Power)." We'll also need to grab the old weapon normalization formula from wowpedia, which is as follows; normalized_damage = base_weapon_damage + (X * Attack Power / 14), where X is 2.4 for 1-handed weapons, and 3.3 for 2-handed weapons. So, let's see the updated formulas:
DW: normalized_damage = 1.5 * (base_weapon_damage + (2.4 * Attack Power / 3.5))
2H: normalized_damage = base_weapon_damage + (3.3 * Attack Power / 3.5)
This translates to a base APCE of 102.857% + 1.5*BWD(Base weapon damage) for all DW strikes, and an APCE of 94.286% + BDW for all 2H strikes. We then multiply that by the FS weapon damage %, which is 75%, to find out what the number is specifically for FS, which ends up being 77.143% + 1.125 * BWD for DW, and 70.714% + 0.75 * BWD for 2H. 

Now that we have translated weapon damage into something we can work with, we have another issue, and that is the cost of FS. Because of improved frost strike, the actual cost is lower than the base cost, which raises the problem we managed to skip over for DC, and that is rune regeneration talents. It seems that FS will have it's base cost changed to 30, so that it can generate rune regeneration talents at an equal rate to death coil, and improved frost presence will be changed to a 10 reduction instead of 15. This basically means that FS will proc rune regeneration talents 50% more "efficiently" than anything else. So for example, if you spent 100 runic on death coil, you would get 100 runic/30 cost *2 charges per use, or 6.6 BT charges. However, if you spent it on frost strike, you would get 100 runic/20 cost * 2 charges per use, or 10 BT charges, which is exactly 50% more. That includes 50% than BoS. So, how do we deal with defining the value of those talents? With a calculation a lot longer than this one is going to end up being. I'm therefore going to leave this up as a challenge to anyone interested in giving some theorycrafting of their own a shot, and for the time being I'll be ignoring the added rune regeneration from using FS over BoS. 

A third factor we need to take into account is rune of razorice, which will give a 15% frost damage bonus to the primary target, but no other targets (generally). This gives us enough to make a new formula!


((Breath of Sindragosa Attack Power Coefficient * (Number of Targets + Razorice bonus) - 4 *Frost Strike Attack Power Coefficient/3) * (Ticks) - 4 * Frost Strike Attack Power Coefficient/3)/ Time interval = Damage per second, in terms of Attack Power * Mastery

((X*(N + .15) - 4Y/3) * (T) - 4Y/3)/I = DPS

And yet another issue arises it's ugly head. We can SOLVE for DPS only if we know all the variables, and one we don't know is base weapon damage. We can update the equation to get around this by taking out the Ys, which contain the BWD, but it isn't pretty. For the time being, it will do. 

((X*N * (T + .15))/I = DPS + (1.15 * 4Y/3 *(T + 1))/I


Defile:

Unholy:
Luckily for us, defile is easy to understand compared to BoS. As long as we continue to assume you want to be using death and decay, even on single target, the formula for the gain is simply:

Number of Targets * (Defile Attack Power Coefficient * Number of Ticks - Death and Decay Attack Power Coefficient  * Number of Ticks)/ Time interval = Damage per second, in terms of Attack Power * Mastery

N*(X*T - Y*T)/I = DPS

Frost:
Slightly worse than unholy (Is that a thing? Seems to be), but still not that bad. The main difference for frost is that defile is affected by razorice on the primary target, DnD isn't, and that DnD doesn't benefit from mastery. So we get the slightly adjusted:

((Number of Targets + Razorice bonus) * Defile Attack Power Coefficient * Number of Ticks)/ Time interval = Damage per second, in terms of Attack Power * Mastery + Damage per second, in terms of Attack Power

((N+ 0.15) *X*T)/I = DPS + (Y * N * T)/I



Necrotic Plague:

Unholy:
Let's look at the tooltip for this one, one more time. "Necrotic Plague deals 4.66% of AP as shadowfrost per stack every 2 seconds for 30 seconds. Each time it deals damage, it gains 1 stack and jumps to a new target within 8 yards. Replaces both diseases." The damage it deals would be the following:

Necrotic Plague Attack Power Coefficient * Stack* Number of Targets/ Interval

(X*S*N)/I

It replaces base diseases, so we need to subtract the damage they would be doing, which is:

(Blood Plague Attack Power Coefficient * Mastery * Number of Targets + Frost Fever Attack Power Coefficient * Number of Targets)/I

(Y*M*N + Z*N)/I

Which gives us the finished formula of:

(X*S*N)/I - (Y*M*N + Z*N)/I = DPS 

We have another instance of part of the damage not being affected by mastery, so we need to juggle it around so that's outside of the first part. 

(X*S*N)/I - (Y*N)/I = DPS + (Z*N)/I

Frost:
For once, frost is actually the same as unholy. Oh, wait, nevermind. Razorice creating more work for us again! Man, I'm going to look like such a dumbass when it ends up being not worth using or removed or something. 

Anywho, 


(X*S*(N+0.15))/I - (Z*(N+0.15))/I = DPS + (Y*N)/I



Step 4:
The final step, we need to use the formulas to produce some numbers, and then compare numbers together to compare the relative strength of the talents. I've included things like number of targets, number of ticks, and average number of stacks into the formulas, so it should be very easy to tweak them to get whatever result you desire. The goal of this post is to share not the results, but the process. As we go further into warlords, it's inevitable that tuning and mechanics change, rendering any results I might post now irrlevent. However, barring massive reworks, those formulas will continue to work and continue to produce updated results whenever anything changes. If you are curious about the current balance, I challenge you to use them yourself and figure it out, and share your results in the comments! Now, I'm going back to working on Rotationcraft! 

Friday, May 9, 2014

Rotationcraft

Skill(PVE):

Players who put in the time and effort to find a theorycrafting blog are generally skilled players. But what is a skilled player? I like to divide skill in WoW into 2 things, both the ability to correctly deal with the various encounter mechanics ever present in raids, and the ability to choose which is the optimal ability for your rotation, whether it be for staying alive, saving someone else, or (more oft than not for us DPSers), which ability will lead to the most overall damage.



So, we have mechanical skill, and rotational skill. Mechanical skill is learned through practice, clever strategies devised by your guild or copied from others, and through research into boss mechanics, knowing what to do and when. There are a lot of useful tools to increase your mechanical skill, including the dungeon journal, datamining from MMOC, and video/text guides from sources such as Icy Veins, or Fatboss. Watching another guild kill a boss, or chatting with a player who has killed said boss and learning what you need to pay attention and react to is another source of insight. I'm not versed enough in mechanical skill myself to be able to school anyone in it.


(I don't really have a place for this but it's too good to leave out. Credit to Ketsuki in #acherus for linking it)


Which leaves rotational skill. While mechanical skill helps to keep you and your guild alive, it's rotational skill that beats the DPS checks, shortens the fight (lowering the chances you get to screw up a mechanic and wipe), and makes you look good on the meters. So, how do we increase our rotational skill? We can further divide rotational skill into knowledge and application. In a number based game like wow, there is always a "right" choice in the rotation, always an optimal ability to press (or in some cases not press) to give you the most possible overall, or burst, damage. Knowing which ability is correct is only part of the puzzle, you also need to be able to react to changes in your resource state, and in the number of targets, all while executing your rotation in a fast-paced raid encounter while paying enough attention to raid mechanics to avoid death. This is pretty hard.


While doing reviews for sentrytotem.com/help-me, and looking over various logs for research, I noticed a very interesting connection for every single DK log I looked at. Every single log, from the 10 man normal mode raider to the world 1 25 heroic parse, made mistakes. There was not a single perfectly executed rotation among all of the hundreds I looked at. There was, however, a very large gap between how well a high parse was doing vs. a low parse. While no one played 100%, plenty of people played 95%, and plenty more people played at 60-70%, and the difference showed. 



Improving Your Rotational Skill:


Right now, the way we learn our rotations involves looking at a (generally) text based guide for instructions and trying to replicate the rotation listed there. However, those guides are often oversimplified and lose something in the translation into text. So, those who want to be more skilled scour the internet for information, looking to #acherus, theorycrafting blogs, websites like EJ/IV, and forums like battle.net and mmoc, asking expereinced players questions or clarifications, and trying to get a good grasp on what ability to use and when. Then we put that into practice in a raid, and our main source of feedback on whether or not we're doing it correctly is a single number, our DPS. A number that might go down when we do something badly, up when we do something well, and overall stay the same. A number that is heavily influenced by RNG. A number that isn't that useful. 

With that in mind, I decided to develop a better way of learning both the knowledge and practicing the skill of an ideal rotation.



Rotationcraft:



Rotationcraft lets you practice your rotation in a controlled environment. When you use a suboptimal ability, it let's you know what the optimal ability is and why. That's the gist of things. Let's take a look at planned features! 

  • Allows you to practice rotations for multiple specs including Frost DW, Frost 2H, Traditional Unholy, Festerblight2.0, and Blood. 
  • Let's you practice a particular fight, helping you learn how to time AMS, how to time Dsim, swap from ST to multiple target sustained, and multiple target burst, and back as well. All SoO fights, as well as a target dummy will be available. 
  • Teacher feature watches you play, informing you of any missed opportunities for more DPS. Dynamic skill settings allow beginners and veterans alike to learn something new. Offers a score at the end of the fight, letting you know how much you missed out on from mistakes, and which mistakes are the most important to fix. 
  • Timescale slider lets you practice the rotation at full speed, half speed, 25% speed, or even turn based. 
  • Allows you to change trinkets, tier set bonuses, talents, and glyphs on the fly. 

Now, you're either like me, and really excited about this idea, or you don't see the point. If you're in latter tribe, I've got nothing for you, head back to camp.






Still here? Let's talk practicality. This is a ton of work. It may take quite a while for me to do, but I'm confident I can do it. I'll be spending a fair bit of my free time on this project in the coming months, but I'm already at the point where I feel feedback on current progress will be helpful, so at the time this post goes up at 6 PM EST on Friday, May 9th, 2014, I'll be looking in #acherus and my friends list for about 5-10 beta testers to give me feedback on the current version of Rotationcraft. Here is a list of all the currently finished and (hopefully) fully function features, as of the time of this post. 

  • Spell casting system complete
  • Spell tooltips complete
  • Rune/runic resource system, including regeneration, death runes, and correct spending complete
  • Action bar and spellbook GUI complete
  • GCD and CD system complete
  • All spells required for single target unholy rotation (other than army) complete
  • Buff and debuff system (diseases) complete
  • Pull button, and linear boss health drop for SR complete
  • Exit button complete (it's important!)

Updates on my progress will be posted here on my blog. The next expected feature to get coded is the teacher component, which will be bundled with the timescale feature and released as a public testing version in about a month. I plan to have the project finished a few months in advance of WoD, and changing the mechanics over will be a lot easier than coding them from scratch. By the time WoD does release, I'll hopefully have added in all of the bosses for the first raiding tier. 

So! Quick Recap and FAQ. Not that I've been asked any questions yet.... but maybe this will save you the trouble?

Q: Where can I get me a copy to try out? I want to give feedback on the program!
A: The project is currently in a limited testing state. When a public release is ready, it will be announced on the blog, so just keep your eyes on it.


Q: Where can I leave feedback on the idea?
A: Here in the comments! I'm curious as to any potential features you think are not worth it, or any ideas for improvement. 

Q: How many people are working on this project?
A: As of this post it's just little ol me. 

Q: I think the idea is dumb, and you're wasting your time. 
A: Well, that's not exactly a question now is it? Don't worry, it's my time to waste. If the project ends up being a total failure, that's the only real risk. 

Q: I think the idea is great! How can I help?
A: Leave feedback in the comments on the idea! When the project is ready for it's initial release, I'll need plenty of help to test it and fix any bugs. Also, it needs an icon (a program without an icon to click on looks silly), so if any artists more skilled than I (read: every artist, ever) wants to give it a shot, let me know. Reminder, Destiny Softworks is still looking to hire an artist! See the Destiny post. 

Q: We're Blizzard, and you've stolen our art assets. We're going to shut down your website and sue for you large amounts of money. 
A: Ahhhhhhh! Please reconsider. I've put in a © Blizzard for the art assets in the info of the program. If you have any objections to rotationcraft, I will shut down down the whole thing. Don't shut me down!

Q: Do you have any plans to support any other classes? This would be really cool for my alt!
A: It's a huge amount of work just to get DKs up and running. On top of that, I don't know enough about other classes yet to be able to fully implement their mechanics, and teach the program what the perfect rotation is. So, no current plans to support other classes. If you're interested in representing another class, and are versed in c# programming, send me a message somewhere or other. (Twitter, #acherus, comments, where-ever really). 

Q: What platforms will Rotationcraft run on?
A: Windows, mac, and linux. If it wasn't such a hassle to distribute a mobile app, I would put it on the Iphone/Ipad as well. 

Q: Ask me a question so I can put it here!
A:

Q: Ask me a question so I can put it here!
A:

Q: Ask me a question so I can put it here!
A:

(Disclaimer: It might come out in 3 days, it might take 3 years, though I sure hope not. I could die tomorrow and it would never get finished. This is a project which will take a lot of my free time and doesn't stand to reap me any profits. I'm doing it because I believe it will help people, and while that is all well and good, sometimes life takes a priority over such hobbies. So don't try to hold me to a release date. It'll come out when it comes out. )
(Second note: I'd love to have the opportunity to do this for a living. Blizzard, if you're reading, I'm looking for a job. On the one in a million chance this ends up happening you can disregard the above note about release dates, and feel assured that it will instead come out soon™)


Saturday, April 19, 2014

Haste, Breakpoints, and Dual Weaponry

Haste:

What do you think of when you think of haste? Do you look at haste with affectionate eyes?


When I think of haste, the cold gleam of hate in my eyes should remind all those around me that I can be a grumpy death knight. Haste is, by far, the most complicated stat to deal it. It looks simple enough, haste makes you do stuff faster. But the value of haste defies understanding, it seems to actively resist being quantified into exactly how much faster and why. Sure, it makes you do stuff faster, but it also does things in a bumpy way. You can't get half an extra hit in per fight, just like your army ghouls, gary, and timmy can't get in half of an extra hit. So, unlike all the other stats, haste is riddled with steps along the way, where someone or other gets in one more hit with that extra juicy point of haste.


Despite all this, haste is generally considered crap for DW frost dks. It's our worst stat, point for point, and this is for a couple of reasons. The biggest of all these reasons is that we don't have any time to do more stuff. DWF dks are already "GCD capped", or using an ability every GCD, with very little haste. This isn't an inherent design flaw, but rather stems from plague leech and AMS. Plague leech generates 2 runes every 30 seconds, and has the added benefit of being used only when we need it. It also takes up a GCD to cast, so it ends up spending 3 GCDs every 30 seconds, or 10% of our time, just on PL. It's worth about 8000 haste, in terms of rune regeneration, and is actually better because it's on call for whenever we want it. AMS, in a similar vein, generates on-demand runic. With the regenerative magic glyph, and a decent amount of magic damage in the fight, AMS generates a LOT of runic. So much, that it ends up eating a large number of our GCDs as we try to keep up with the runic gain, which leads to BT charges, which leads to HBs which leads to more FS. It takes a while for the flood of resources to die down after an AMS usage.


While not letting us use more stuff with all of the extra resources coming in from PL and AMS, haste still gives us more white hits, as well as proc uptime, km procs, pet swings, ect ect. The non-rune related benefits of haste seem to be worth about 50% of it's total value, with the other 50% coming from actually being able to use the runes. When we are GCD capped to the point we are now, and unable to use half of the value of haste, the other half simply isn't good enough to compete with better stats like mastery and crit. And so, haste is weak, and feels bad because it leads to wasting more resources. Most DWF dks avoid haste like the plague. 


Breakpoints:

There seems to be one breakpoint worth reaching. The T16 2-set bonus is as follows: Killing Machine grants 500 mastery for 6 seconds, stacking up to 10 times. Now, haste makes the uptime of "death shroud" go up in a pretty linear fashion. More haste means more melee swings means more KM procs, which in turn means more death shroud buffs. However, there is also a pretty big step hidden here. Death shroud lasts 6 seconds, and in that 6 seconds you do a certain number of melee swings. Going from 5 swings to 6 swings is actually a really big deal, as it's an extra chance to extend the duration of a proc instead of proccing a fresh one a half second later. 

The most important, and only really viable, breakpoint I could find was for 8 attacks in the 6 second period. It doubles the uptime you have on a 10-stack of the 2-set. It's worth a net increase of about 1400 dps, or 0.4%, which is amazing for a single point of haste. It's easily reachable and worth reaching in current gear. 




The breakpoint is at 3686 haste. Below is a list of all of the breakpoints. 

Attacks:   Interval Between Attacks          Haste required      Haste from gear required   Haste rating

5 attacks: 1.2 1.08333333333333 0
6 attacks: 1.0 1.3 0
7 attacks: 0.85714285714286 1.51666666666666 0
8 attacks: 0.75 1.73333333333333 1.08672936259143 3686
9 attacks: 0.66666666666667 1.95 1.22257053291536 9460
10 attacks: 0.6 2.16666666666667          1.35841170323929 15233
11 attacks: 0.54545454545455 2.38333333333331        1.4942528735632 21006
12 attacks: 0.5 2.6 1.63009404388715 26779



Dual Weaponry: 

After calculating the values, I tried them out in simC. At 3686, (nearly exactly), the sim reported a huge increase in both damage and 2-set 10 stack uptime. I was overjoyed to have everything work out, and thought I'd be wrapping up soon. Then I tried the next breakpoint at 9460. There was none. I tried 15233, and uptime went from 25% to 40% in a single point of haste. 21006 was missing as well. The culprit of the missing odd breakpoints is the way simC handled auto-attacks. It activated main-hand and offhand attacks at the same time, causing a super breakpoints, gaining 2 attacks half as often. This was quickly fixed, thanks to Collision in the simC IRC channel, and yet upon trying out the new version, I found that the odd breakpoints had moved from on top of the even breakpoints to about ~2000 haste before they were supposed to be. This had me perplexed, until I ran through a simC output log and noticed the problem. At the end of bloodlust, the swing timer remaining for each weapon was different, and the dynamic adjustment in haste caused the weapon swings to desync. This gap between weapon swings led to the odd breakpoints, or breakpoints with the offhand weapon, to be in an unexpected place. This could be fixed simply enough, but to what end? Review of in-game logs show that melee hits happen 0.2% more often than they should, and not in perfect intervals. That could just be logs misrepresenting the information, but in-game testing also points to desyncing issues. If you autoattack a dummy out of range, wait a second, and move in range, you can watch your toon attack with both weapons at the same time, and then seemingly sync back to alternations, and then back to at the same time, over the course of several seconds. 

I'm 95% sure that the autoattacks of the mainhand weapon are timed correctly, but I'm not sure what the hell is going on with the offhand. Luckily, the first odd breakpoint is at 9460, and that's not worth going for. Stick as low as possible haste, without going lower than 3686, and expect to see a little bit of extra damage!

Bonus: 

It's possible to force sync weapon swing timers by stepping out and in of range. That would bring the odd breakpoint right down to on top of the even breakpoint, which would be a DPS gain. Stepping in and out after every single melee swing..... (if you didn't lose any dps by stepping out of range)
Would make the death knight hokey pokey a way to increase DPS. 
Huge thanks to Arrak in #acherus and Collision/Reia in #simcraft for helping me get this all figured out. 
I'm trying out a new style with the google images because reasons. I like it, do you like it?

Friday, April 4, 2014

6.0 DK Changes, Part 1

Today is a good day. After months of drought, we finally have the first batch of changes for the 6.0 WoD expansion, and I'm going to take a look at the ones pertaining to DKs. Keep in mind all of these are, as always, subject to change, and the full list of changes can be found at battle.net, and mmoc.


Revised Level 100 Talents:

Necrotic Plague - A powerful disease that deals XXXX Shadowfrost damage per stack every 2 sec for 30 sec. Each time it deals damage, it gains 1 stack, and infects another nearby enemy within 8 yards if possible. You gain 5 Runic Power whenever a target infected with Necrotic Plague attempts to attack you. Replaces Blood Plague and Frost Fever, and is applied by any ability which applied either.

With a base damage of X, and gaining 1 stack every damage deal, the first tick will do X damage, and the final tick will do 15X damage. The total damage is X + 2X + 3X... or 120X. Under half of this damage (55/120) is done in the first 20 seconds, and the majority takes place in the last 10 seconds. The free spreading infection will be a very nice cleave boost, especially on streaming adds, without requiring constant reapplication of pestilence/BB. How high X is will determine whether or not this is worth sacking FF and BP and a talent slot for. From a DPS perspective, you will not often have targets trying to melee you, so the defensive benefit is likely to be moot.

Defile - Defiles the ground targeted by the Death Knight, causing XXXX Shadow damage every 1 sec for 30 sec. Each time Defile deals damage, it increases in area and damage by 5%. This growth can occur a maximum of 3 times per second. Enemies standing in your Defile deal 10% less damage to you. Replaces Death and Decay. 1 Unholy, 30 yd range, Instant, 30 sec cooldown

Defile has had it's duration increased from 10 seconds to 30 seconds, which is huge. Below are some tables of the overall damage with both 1 and 3 targets total. Assuming that the 10% less damage debuff applies to AE abilities, defile has the potential to be a great defensive CD as well, 10% less damage taken with 100% uptime is silly for both tanks and dps, and I expect to see that nerfed. 

With each defile tick hitting 3 targets:

Tick 1: X
Tick 2: 1.15X
Tick 3: 1.3X
Tick 4: 1.45X
Tick 5: 1.6X
Tick 6: 1.75X
Tick 7: 1.9X
Tick 8: 2.05X
Tick 9: 2.2X
Tick 10: 2.35X
Tick 11: 2.5X
Tick 12: 2.65X
Tick 13: 2.8X
Tick 14: 2.95X
Tick 15: 3.1X
Tick 16: 3.25X
Tick 17: 3.4X
Tick 18: 3.55X
Tick 19: 3.7X
Tick 20: 3.85X
Tick 21: 4.0X
Tick 22: 4.15X
Tick 23: 4.3X
Tick 24: 4.45X
Tick 25: 4.6X
Tick 26: 4.75X
Tick 27: 4.9X
Tick 28: 5.05X
Tick 29: 5.2X
Tick 30: 5.35X
Total: 92.25X

With each defile tick hitting 1 target:

Tick 1: X
Tick 2: 1.05X
Tick 3: 1.1X
Tick 4: 1.15X
Tick 5: 1.2X
Tick 6: 1.25X
Tick 7: 1.3X
Tick 8: 1.35X
Tick 9: 1.4X
Tick 10: 1.45X
Tick 11: 1.5X
Tick 12: 1.55X
Tick 13: 1.60X
Tick 14: 1.65X
Tick 15: 1.7X
Tick 16: 1.75X
Tick 17: 1.8X
Tick 18: 1.85X
Tick 19: 1.9X
Tick 20: 1.95X
Tick 21: 2X
Tick 22: 2.05X
Tick 23: 2.1X
Tick 24: 2.15X
Tick 25: 2.2X
Tick 26: 2.25X
Tick 27: 2.3X
Tick 28: 2.35X
Tick 29: 2.4X
Tick 30: 2.45X
Total: 51.75X

Breath of Sindragosa - Continuously deals XXXX Shadowfrost damage every 1 sec to enemies in a cone in front of you within 12 yards, and affects them with Mark of Sindragosa for 6 sec. You will continue breathing until canceled or Runic Power is exhausted. Mark of Sindragosa causes you to be healed for 5% of spell damage dealt by afflicted enemies. 15 Runic Power, plus 15 per sec, Instant, 1 min cooldown

It's possible to keep this up for a long period of time, starting with 100 runic gives you the 85/15, or 5.6 seconds, during which you can pretty easily generate another 50, which is another 3.3 seconds, after which you can AMS for 100-200 runic over 5 seconds, which is another 13 seconds, and you can continue generating runic as you go along. I'd imagine it's possible to keep it up for about 20-25 seconds max, but not much has changed about this talent since blizzcon, and I'll still need to know what X is before I can do much more with it. The defensive benefit is better than necrotic plague's, but worse than defile, IMO. 

LV 90-100 Perks:


All:
Improved Diseases - Increases damage done by your diseases by 20%.
Improved Soul Reaper - Increases Soul Reaper damage by 20%.

Blood:
Enhanced Will of the Necropolis - Allows Will of the Necropolis to be triggered again 30 sec sooner.
Enhanced Rune Tap - Reduces the cooldown on Rune Tap by 10 sec.
Enhanced Death Grip - Reduces the cooldown on Death Grip by 5 sec.
Enhanced Bone Shield - Bone Shield has 2 additional charges.
Improved Heart Strike - Increases Heart Strike damage by 30%.
Improved Death Strike - Increases Death Strike damage by 20%.
Enhanced Rune Strike - Rune Strike also creates a Rune Shield, absorbing the next XXXX Physical damage within 10 sec.

Frost:

Improved Razorice - Rune of Razorice deals an additional 12% weapon damage when it triggers.
Improved Cinderglacier - Rune of Cinderglacier has 3 additional charges when it triggers.
Improved Soul Reaper - Increases Soul Reaper's Shadowfrost damage by 40%.
Empowered Pillar of Frost - Pillar of Frost provides an additional 10% Strength and no longer costs any runes.
Empowered Icebound Fortitude - Icebound Fortitude reduces damage taken by an additional 10%.
Empowered Rime - Rime also increases the damage done by Howling Blast and Icy Touch by 20%.
Enhanced Death Strike - Increases Death Strike healing by 20%.
Unholy:
Enhanced Death Strike - Increases Death Strike healing by 20%.
Enhanced Dark Transformation - Dark Transformation no longer costs any runes.
Improved Scourge Strike - Increases Scourge Strike damage by 20%.
Improved Festering Strike - Increases Festering Strike damage by 20%.
Enhanced Fallen Crusader - Rune of the Fallen Crusader heals for an additional 7%.
Enhanced Death Coil - Increases Death Coil damage by 20% and range by 10 yards.
Empowered Gargoyle - Increases the duration of Summon Gargoyle by 10 sec.

Most of the perks are only number changes, which are irrlevent. The numbers only matter compared to each other, and don't make much of a difference in a vacuum, as balance is way too far off to even consider. I'll list the mechanical changes below, and talk about them a bit. 

Blood:
Enhanced Will of the Necropolis - Allows Will of the Necropolis to be triggered again 30 sec sooner.
This is.... insane. Will of the Necropolis has a 45 sec ICD right now, so this brings it down to only 15 seconds. True, the 25% damage reduction for 8 seconds is being nerfed down to 4 seconds, but it's still overall a massive buff, and makes blood a better tank at surviving more damage when already at low health. The free rune tap every 15 seconds is nothing to snuff at either. 

Enhanced Rune Tap - Reduces the cooldown on Rune Tap by 10 sec.
This brings rune tap down to 20 seconds, and when combined with the above change to Will of the Necropolis, it means you can rune tap 7 times a minute, (if Will of the Necropolis has no downtime)

Enhanced Death Grip - Reduces the cooldown on Death Grip by 5 sec.
Brings the CD down to 20 seconds, nice QoL change, will give blood a little more utility in certain raid encounters, and help the leveling process a bit. I wish this applied to DPS specs as well. 

Enhanced Bone Shield - Bone Shield has 2 additional charges.
Up to 8 charges baseline, nothing major, but a nice little defensive bonus for blood. 

Enhanced Rune Strike - Rune Strike also creates a Rune Shield, absorbing the next XXXX Physical damage within 10 sec.
This is a pretty major mechanical change. Now, both death strike and rune strike create a shield. Rune strike will no longer be a filler, used during low damage periods to ensure you have the runes available to survive burst damage, but instead be a part of the survival cycle, depending on how they balance the shields. Assuming they are at least somewhat similar, and taking into account the fact that death strike shield no longer scales with short term damage and rather long term vengeance, I believe DS timing will be a thing of the past, and be replaced with a constant flow of DSs and RSs. 

Frost:
Improved Razorice - Rune of Razorice deals an additional 12% weapon damage when it triggers.
This is a pretty significant change, RoRI only deals 2% weapon damage now, so this will increase it to 14%. Still only a numbers change though, so not much to say, other than it looks like they are giving runeforging a look like they said they would. 

Improved Cinderglacier - Rune of Cinderglacier has 3 additional charges when it triggers.
I expect this change to put cinderglacier on the map. Considering how hard I expect defile, necrotic plague, and sindy's breath to hit for, and the fact that 5 charges is a lot better than 2 charges, this could well be the de facto runeforge by the time WoD rolls around.

Empowered Pillar of Frost - Pillar of Frost provides an additional 10% Strength and no longer costs any runes.
Nice QoL change, you'll no longer need to bank a rune right before pillar's CD comes up, and this should overall increase our pillar uptime by a small amount (more for less skilled players). You also get a free rune! 

Empowered Icebound Fortitude - Icebound Fortitude reduces damage taken by an additional 10%.
Nice to see a 50% buff to our main defensive skill, which has always felt very weak-sauce. 

Empowered Rime - Rime also increases the damage done by Howling Blast and Icy Touch by 20%. 
Another significant mechanical change. This will probably kill mastersimple, as even DW's obliterates will give you free, juked up HBs. Only time and math will tell. 


Unholy:
Enhanced Dark Transformation - Dark Transformation no longer costs any runes.
Nice, small QoL change, similar to PoF. We no longer need to bank a rune, and get slightly higher on DT, AND a basically free rune.

Enhanced Death Coil - Increases Death Coil damage by 20% and range by 10 yards.
Death coil range is the interesting thing here, and even that's not very interesting. It's nice, but we out of range rarely enough that it won't matter much. 

Empowered Gargoyle - Increases the duration of Summon Gargoyle by 10 sec.
40 second gary is nice, but still basically just a numbers change. the longer solid block of casting won't cut off as many casts, on a per fight basis, which is nice. 



Megablog:

Each point of Agility or Strength now grants 1 Attack Power (down from 2). All other sources of Attack Power now grant half as much as before.
Weapon Damage values on all weapons have been reduced by 20%.
Attack Power now increases Weapon Damage at a rate of 1 DPS per 3.5 Attack Power (up from 1 DPS per 14 Attack Power).

Leaving this here for my reference more than anything else, this is the most numbers we've gotten in ages. This isn't that interesting now, but will be very useful later on.


Army of the Dead now deals 75% less damage and is only available to Blood Death Knights.
Dual Wield is now only available to Frost Death Knights.
Horn of Winter no longer generates Runic Power, has no cooldown, and lasts 1 hour.
Necrotic Strike is now learned only by Frost and Unholy Death Knights.
Raise Dead is now only available to Unholy Death Knights, summons a ghoul as a permanent pet, and has a 1-minute cooldown.
Master of Ghouls has been removed.
Rune Strike now replaces Death Coil for Blood Death Knights.
Frost Strike now replaces Death Coil for Frost Death Knights.
Obliterate now replaces Blood Strike for Frost Death Knights.
Unholy Frenzy has been removed.
Our CD pruning! Overall, I think we got off pretty easy. Frost/unholy lost army, frost/blood lost raise dead, and death coil, and unholy lost unholy frenzy. All specs also lost HoW as a runic generator, and frankly, it was probably one of least interesting abilities. I have no idea what the comment about frost being the only spec that can DW means, as frost is currently the only spec that can DW. Other than CDs and HoW, we didn't lose anything.

A few things have happened for Death Knights. As described above in Ability Pruning, several cooldowns were made spec-specific. Rotations remain unchanged for the most part. One thing we did polish up is the effect of Diseases on the damage of other abilities. Diseases now do enough damage on their own to warrant using them, so having them act as multipliers on the damage of other abilities was extraneous, and just cluttered up the tooltips of those abilities. We removed those multipliers, and just baked in their benefit to the corresponding base spell. This slightly reduces ramp-up time.
Blood Boil damage has been increased by 50%, but no longer deals additional damage to targets affected by Blood Plague or Frost Fever.
Heart Strike damage has been increased by 30%, but no longer deals additional damage for each disease present on the target.
Obliterate now deals 25% more damage on both main- and off-hand weapons, but no longer deals additional damage for each disease present on the target.
Scourge Strike no longer deals additional damage for each disease present on the target. The ability now deals 50% of the Physical damage dealt as additional Shadow damage; this effect can critically hit and no longer ignores modifiers.

This is old news from blizzcon, diseases no longer scale ability damage up. This means pretty much nothing for us, other than what is described above, that we have slightly less ramp-up time. Disease uptime will still probably be essential, if the application remains as easy and painless as it does today.

In order to better balance the scaling rates and value of secondary stats for Unholy Death Knights, we reduced the power of their passive Unholy Might ability.
Unholy Might (Unholy) bonus changed from +35% to +10%.
Numbers nerf! 50 gold says by the end of WoD it's back up to 50%.


Active Mitigation was a very successful design that was inspired by Death Knights' tanking style. However, it went beyond that, and Death Knights themselves were somewhat left behind in that regard. We made several changes to bring up the interactivity of Blood combat. This includes making Death Strike cause healing based on attack power, but be affected by the new Resolve passive (see Tank Vengeance and Resolve, above), which gives it the traditional increase from recent damage. Plus, Rune Tap is being significantly improved, to become a strong Active Mitigation button. To tie those together, Blood's Mastery is being changed to affect both the size of the Blood Shield absorb, and the new Rune Shield absorb. Additionally, we removed Dodge and Parry from gear, and expect Blood Death Knights to value Haste and Crit as important secondary stats. In order to achieve that, we made Riposte give defensive value to auto attack and Rune Strike crits. Finally, we tweaked the targeting AI of Dancing Rune Weapon, and fixed it up to use most Talents that you know.
Heart Strike now only cleaves 1 additional target (down from 2).
Death Strike now causes healing that scales in effectiveness with attack power, instead of based on damage taken in the last 5 sec. This healing is affected by Resolve.
Blood Rites now causes Death Strike to also cause an absorb for 50% of the amount healed.
Rune Tap now grants an absorption shield that scales in effectiveness with Attack Power, instead of a heal based on a % of maximum health. It also now has 2 charges, with a 30-second recharge time.
Will of the Necropolis now grants an immediate charge, and makes your next Rune Tap free. However, its damage reduction buff now only lasts 4 sec (down from 8sec).
Mastery: Bloody Shields now increases the effectiveness of Blood Shield and Rune Shield by 16% (percentage increased by mastery), instead of adding the Blood Shield effect.
Riposte has been redesigned.
Riposte: After getting a Critical Strike with an Auto Attack or Rune Strike, you gain +100% Parry chance until you Parry an attack. This can stack up to 2 times.
Dancing Rune Weapon’s summoned Rune Weapon now remains fixated on the Death Knight’s target at the time of summoning, and copies the effects of Talents that are tied to the Death Knight, such as Blood Boil, Frost Fever, or Asphyxiate. Should the original target be dead or otherwise unavailable, the Rune Weapon will switch to assist with the Death Knight’s current target.

Blood stuff! Not my specialty, but I'll go into these changes a little. The primary difference between blood and the other activate mitigation models for mop has been that blood scales with mitigated damage taken, while every other spec scales with vengeance, or unmitigated damage taken. That means that every healer absorb, defensive cooldown, or even our own blood shields reduced the value of our next blood shield, and basically led to us shooting ourselves in the foot. That, combined with the short 5 second window for DS timing, lead to some interesting gameplay to try to overcome our weakness. This was called DS timing, and is described a lot more masterfully by reniat and tyvi in their respective blood guides on mmoc and EJ. Changing blood shield over to the new resolve system removes that timing, and to some degree removes an aspect of skillful play for the blood spec. However, at the same time it removes a major weakness of the blood spec, so I believe it's for the best. Combined with rune strike and rune tap also creating a shield, I expect blood rotation will look more like a flowing chain of DS, RS, and RT, trying to get as many DS/RS out as possible in the given period of time. It's possible this will popularize RE as the most throughputty talent for rune regen. The new riposte looks to be basically making crit rating a direct conversion into parry rating, (when 1 crit gives you 1 parry, they are essentially one and the same, assuming you have as many chances to crit as you do to parry), though it will fall very short on AE tanking.

Since the Ghoul pet is now Unholy only, that presents a problem for Death Pact. We revised Death Pact to not require an undead minion, but work a little differently. We left it at 50% heal, which is effectively a 33% buff to it from before (see Healing and Player Health above), and added a heal absorb for half the amount healed, instead. It should now be a relatively more effective heal for staying alive immediately, but with the downside of needing to heal through the heal absorb before you can be healed any further.
Death Pact no longer requires an undead minion, and instead places a heal absorb on you for 50% of the amount healed.
I felt a little indifferent about this change at first, but I'm pretty sure I agree with it now. Death pact is a major defensive cooldown, and one of the abilities I'm very happy was not pruned. It loses a little flavor without the pet sacrificing, (I suppose that's the sacrifice we have to make, flavor to make the game mechanics work) yet remains one of our few strong survival CDs. The anti-heal was an..... interesting way to handle the health absorb, and I'm not sure if it was entirely necessary. They've doubled health pools, but they've also doubled the damage dealt by creatures, so the "33%" healing buff described is only true when compared to other healing spells, this is a nerf compared to existing mob damage. It reminds me a bit of dark bargain, prevent/heal the damage now, and pay for it later. We'll have to see what this actually looks like in a raid setting.


There were also a few other miscellaneous changes. The Runic Power generation of Anti-Magic Shell was standardized, to make it more understandable, and balanced.
Anti-Magic Shell now restores 2 Runic Power per 1% of max health absorbed.
Major AMS regen nerf. Right now, the conversion rate is 1 runic per 1000 health, and AMS absorbs up to 50% of your max health, which meant about 350 runic for a max absorb, in BiS gear. With such a short CD on AMS with the CDR trinket, AMS was a very important part of playing a DK correctly, and the difference between using AMS intelligently or not could amount to as high as 10%-20% dps on certain fights. With this nerf, a fully absorbed soak for 50% of your health (which isn't that hard to get to be fair), will only grant 100 runic. While most of the runic over 100 was wasted, this will make soaking runic from smaller damage sources much harder, and actually reduce your damage as you gain more stamina. I expect people who want to parse as high as possible to actively remove stamina buffs to soak just a little more runic per AMS. 
Overall, this can be considered both a negative change, because it lowers the skill ceiling on how well you can play a dk, and a positive change, because it helps (slightly) to address the fact that we are GCD locked with no haste and a 1 second GCD baseline.



Mastery: Blood Shield now also passively increases Attack Power by 12% (percentage increased by Mastery), in addition to its current effects.
This is a necessary change, now that vengeance no longer increases damage done, to help tank damage scale with gear.





And that's all I have to say about these patch notes in their current state. I'm sure we'll see more information released over the next few months, and as we incorporate this into simC, and start getting some hard numbers, we can begin to run balance checks on the new talents, specs, and rotations to see if anything has changed on that end.

Tuesday, April 1, 2014

Stat Weights; We've been doing it all wrong.

Today, I've made the biggest breakthrough of my already extremely profitable career of being a DK nerd. A thought came to me, a glorious thought, a crazy thought. When we sim our stat weights, we only sim a certain range of stats. Strength, crit, haste, mastery, sometimes hit and expertise if we're feeling peckish. But we leave all the other stats untested. No one has ever been daring enough to sim every stat, before today. As it turns out, we've been doing things wrong ever since DKs were released in wrath.

Not surprisingly, I found only minor gains for parry, dodge, armor, agility, and block rating, but several stat weights blew my mind. The value of intellect, and spirit has been so far underestimated that it's a wonder we're even remotely competitive at the moment. Of all our major class issues, our clunkyness (clunkiness? clunkatude? clunkylunkymonkey?), our scaling problems, and our wonky mechanics, the fact that we have so low DPS is the biggest challenge we face.

And now we have the answer. We must fight smarter, now harder. Our brute force approach, applying as much strength as possible to the problem, could only ever have been folly in hindsight. We are now death knight scholars, and will solve our problems not by hitting our foes really hard on their shields, but rather hard enough in the..... smart places. Attached are the newest stat weights for DW frost, I'll get to the other specs later, but I can only imagine similar results.


Disregard my clearly not photo-shopped colors.