Shifting Perspectives: Changing Eclipse
Every week, Shifting Perspectives explores issues affecting druids and those who group with them. This week, we're going to be looking at the most common gripe of the balance druid. That fickle little talent that every balance druid loves to hate: Eclipse. What's really wrong with it? How can we make it better? Why do we misname the procs?
Every time that I visit the Damage Dealing or Druid forums, I often feel like I am back in Wrath of the Lich King beta. During that time, every day there was some thread about some balance druid issue that needed to be addressed, usually it was Eclipse. For those that currently take up an issue with Eclipse, I want to personally apologize for the state that it has come to today. Although I am not a design for WoW, nor do I have any direct influence on the design philosophy of any class, I can't help but thinking every day that I, along with several other notable figures, paved the way for Eclipse to become what it is today.
For those that do not know what I am talking about, let me explain. During WotLK Beta, Eclipse was a terrible talent. By terrible, I don't mean that it was on par with Genesis. No, it was worse than Genesis. Eclipse was so bad during the early stages of Beta that there was more than several incarnations of the talent where it was a DPS loss to spend any talent points into it. Eclipse started out where you could only gain one proc every minute and the proc was either 10% increased damage to Wrath or 10% increased crit chance to Starfire. I shouldn't say it was all Eclipse's fault that it sucked, at the time the Glyph of Starfire had no extension limitation which certainly complicated the matter as well. Even without that, though, Eclipse simply wasn't worth it. There was no point in proccing Eclipse to buff Starfire at that point because 10% was too at the mercy of RNG to have a noticeable effect, and Wrath did not play well enough with Nature's Grace at the time to make its Eclipse proc viable.
Despite all of that, Ghostcrawler and Koraa said several times that they were bound and determined to make Eclipse work. Thus the balance theorycrafting team went to work. First, we set about proving that the bonus was not high enough, resulting in a buff to both effects to 15%. Then we set about to prove that the duration, which was only 10 seconds at the time, was not long enough, which resulted in a buff to 15 seconds. Next came showing that the cooldown was too prohibitive. This caused a lot of different tweaks. Initially the cooldown was lowered to 45 seconds, then we went to a 30 second cooldown, but back to a 10 second duration, and finally it settled on a 30 second cooldown with a 15 second duration. Finally, we went on to prove that a 15% increase to crit was still too RNG heavy, especially considering all of the other RNG factors of Eclipse, and thus got the Starfire portion of Eclipse buffed to 30%.
How much of Eclipse that eventually came out of Beta was a result of player input and theorycrafting/testing is up for debate. I am not so arrogant enough to believe that Blizzard simply made that changes because we said to make them, but I also don't feel it is enough of a coincidence that a vast majority of the changes were almost perfectly inline with the theorycrafting that we had done. To that end is why I apologize. Blizzard may have done the changes, but we certainly had something of a hand in pushing them towards the current design.
History aside, many people are still upset about how Eclipse functions right now, and with good reason. Eclipse is probably the single strongest talent in the game as it stands. I may be incorrect, but I don't know of any other talent in the game that accounts for almost a 3,000 increase in DPS. With how important Eclipse is to balance DPS, and how awkward some of the mechanics of the talent are, it comes as no surprise that it is often the target of many complaints and suggestions. I would like to take a look at those suggestions, to weigh in on them, and, yes, to provide a little touch of theorycrafting to them to show where they work, where they fail, and where they can be improved. Eclipse is going to be changed in Cataclysm, we already know that, but what we don't know is how. Blizzard most assuredly has their ideals, and I wouldn't doubt it if they are already in the process of testing them, but player input does matter too. So, let's give some input, shall we?
Eclipse: Changing the damage contribution:
Eclipse, as I said, accounts for far too much damage; even Ghostcrawler has admitted to this. We all know that the damage contribution for Eclipse needs to be lowered, the question thus comes down to how. There are several ways in which this could be done. Here are the few that I've come across:
Increase the cooldown for the effect:
Pros:
By reducing the number of possible procs within a given time frame, the damage contribution naturally decreases.
Limiting the 'chaining' effect of Eclipse procs means there is less pressure to instantly get your next Eclipse proc, this allows time to refresh DoTs, allow for repositioning, and the use of long-cooldown based abilities.
Cons:
Doesn't actually reduce the importance of making the most use of each proc. If anything, it increases it.
Movement is still heavily penalized during an Eclipse proc.
Since damage would have to be equalized, the added effect of Eclipse would create additional burst for Eclipse making it more essential in PvP.
Reduce the duration of the effect:
Pros:
By reducing the number of casts possible during Eclipse, the damage contribution naturally decreases.
Limiting the 'chaining' effect of Eclipse procs means there is less pressure to instantly get your next Eclipse proc, this allows time to refresh DoTs, allow for repositioning, and the use of long-cooldown based abilities.
Cons:
Doesn't actually reduce the importance of making the most use of each proc. If anything, it increases it.
Movement is still heavily penalized during an Eclipse proc.
Since damage would have to be equalized, the added effect of Eclipse would create additional burst for Eclipse making it more essential in PvP.
The lowered time frame of usage would further hinder the use of Eclipse in PvP
Reduce the damage modifier of the effect:
Pros:
Given that each proc increases damage by a lower amount, the damage contribution naturally decreases.
Reduces the impact of movement during Eclipse procs given that Eclipse accounts for a lower proportion of damage.
Reduces the reliance on Eclipse for burst in a PvP setting, which is unreliable and subject to heavy RNG.
Cons:
Increases the effects of RNG on Lunar procs, especially at lower gearing levels.
Increases the free-casting out-put of balance druids in PvP that can be difficult to balance
Conclusions:
I know some of those points may seem out of whack, but I assure you that they are all on point to a degree. Although much of that all depends on how you re-balance damage out-put after nerfing Eclipse. For example, if all we do is focus on Wrath and Starfire damage in an effort to keep them approximately the same before and after the change, then all changes mean that the base-line, non-Eclipse damage of Wrath and Starfire need to be increased. As a flat example, if you reduce Eclipse to only buff Wrath's damage by 15%, then you need to increase the base-line damage by 25%. This would mean that a free-casting balance druid would deal approximately 25% more damage (probably closer to 20%) in PvP. As much as I'm sure druids would love that fact, it could easily end up being way too much free-casting damage that requires no more set-up than merely finding time to stand still. To properly handle nerfing Eclipse's damage out-put requires adjusting balance's DPS in areas beyond Starfire and Wrath. Be that through DoTs, Starfall, Force of Nature, or additional spells is a different matter.
Essentially, the only real way in which to balance Eclipse's damage out-put is to reduce the damage modifier of the the effect, though other things about the talent must be changed as well. Eclipse should be nerfed to only provide a 10 - 15% damage increase to Wrath. In reality, Eclipse should probably boost Wrath's damage by 10% while it is active, however given that the Eclipse buff stacks additively with Moonfury, the buff itself should increase damage by 15%. Say Wrath deals 5,000 damage per cast. With Moonfury it now deals 5,500 damage per cast. You want the Eclipse buff to increase this to approximately 6,050 damage per cast - a 21% increase in damage over the non-Moonfury damage of Eclipse. If you change Eclipse to be a flat 10% increase, then it only increases Wrath's damage to 6,000 (5000 * 1.2 = 6000,) which is lower than the target mark. If you change Eclipse to be a flat 15% increase, then it increases Wrath's damage to 6,250 (5000 * 1.25 = 6250,) which is higher than the mark. Although 10% is closer to the mark than 15% is, being slightly over the mark is more preferable to being slightly under the mark.
Altering the damage potential from Eclipse as it relates to Wrath is very simple, however not so much for Starfire. If we follow the same pattern for Starfire as we did for Wrath and reduce it to only a 15% increase in critical strike chance, we run into some major issues. While the issue wouldn't be felt now in ICC due to most raiding balance druids running well over 60% chance to crit with Starfire, it would be felt once we hit the start of the next expansion and have, potentially, only a 25 - 30% chance to crit with Starfire. Going from 30% chance to crit to 45% chance to crit is a fairly nice DPS increase, the issue is that the DPS increase of crit is normalized against time. That is to say, the longer you spend casting, theoretically, the closer to your actual crit chance you are going to get. This becomes problematic for Eclipse since it is only a 15 second time frame in which we are attempting to equalize out RNG. RNG simply cannot be equalized against such a low target point. Throughout a 5 minute encounter, a druid is probably going to have 7 to 8 Lunar procs. Even attempting to balance around a bit shy of 2 minutes of casting is simply not going to happen. Even during a full 5 minutes, Crit RNG will generally not balance itself towards the target mark. Generally speaking, you need to look in terms of thousands of casts to get an accurate equalization.
Due to this, the bonus to Starfire cannot remain as an increase to crit chance if the effect is to be lowered. Instead, it would have to either be a flat modifier to an increase to haste. The latter is more preferable to keep the procs more unique. Assuming Wrath's and Starfire's damage is roughly equalized at a base-level, which they have to be for Eclipse to function, then 15% haste might not work. From base cast times, if Wrath is deal 5,000 damage every 1.5 seconds, then Starfire needs to deal 10,000 damage every 3 seconds. With Eclipse, and Moonfury since it interacts with Solar Eclipse, Wrath would be dealing 6,250 damage every 1.5 seconds, or 4166.666 DPS (6250 / 1.5 = 4166.666.) With Moonfury, Starfire would be dealing 11,000 damage every 3 seconds, or 3666.666 DPS (11000 / 3 = 3666.666.) This means that Eclipse needs to boost Starfire's DPS by approximately 500. At 15% haste, Starfire's cast time would be approximately 2.6087 (3 / 1.15 = 2.60869.) This would change Starfire's DPS to 4216.6596 (11000 / 2.6087 = 4216.6596.) While not excessively higher than the Wrath mark, also keep in mind that all haste effects are multiplicative instead of additive, meaning this effect would increase slightly in a raid setting.
We don't particularly want one Eclipse proc to be more powerful than the other. Although, with chaining procs, it isn't as important as before when druids could only use one proc; it can cause a few scaling issues. Let us lower the effect to Starfire down to 10%, then. At 10% haste, Starfire has a cast time of 2.7273 (3 / 1.1 = 2.72727.) This changes the DPS to 4033.293. This is a touch lower than the damage of an Eclipsed Wrath, however, keep in mind that all haste increasing effects are multiplicative. Unlike the increase to Wrath from Eclipse which, when combined with Moonfury, is slightly lowered in actual contribution, the increase to Starfire from Eclipse increases with other haste effects. If haste effects remain the same in Cataclysm as they are now, then druids would see 3% haste from Celestial Focus, 3% from Improved Moonkin Form, and 5% from Wrath of Air Totem. Instead of an 11% increase in haste, this is actually an 11.3945% increase in haste. A minor difference, but a difference all the same. When combined with a 10% increase from Eclipse, the total haste would be 22.53395% haste. This means that Eclipse is actually providing an increase of 11.13945% haste (22.53395 - 11.3945 = 11.13945.) With that change, Starfire's cast time due to Eclipse would be 2.6993 (3 / 1.1113945 = 2.69931,) which results in a DPS value of 4,075.1305. It isn't a major increase, and Starfire would still be behind Wrath by around 91 DPS. There is also Nature's Grace to consider, and how cast times interact with haste from gear, so it would be more preferable to keep the Starfire bonus slightly lower than the Wrath bonus. Either option is within a close enough margin to work though.
