Thursday, August 2, 2012

Balance Rotation: Take 3

The end is near, and that's not a big surprise. People have been guessing that it would be a late summer early fall release for months. With the expansions cinematic being released on August 16th and the new novel Tides of War coming out on August 28th, it wasn't hard to guess that the expansion things were almost done. Then Blizzard made it official by announcing the release date of Sept 25th.

In years past, I would have taken this as a sign that almost everything was done. The rotation was the rotation and it was up to us to make the best of the tools we have been given. However, that is not the case this time with Blizzard making significant changes to the rotation just two months before MoP's release. It's hard to say how much time we have to influence these the design at this point but lets take a look.

Fae Empowerment: We Barely Knew You

The first big change is that Fae Empowerment has been removed and rightfully so. Fae Empowerment was a failure on almost all fronts. At first, it provided little or no DPS boost even when used perfectly and lowered DPS when used incorrectly. Blizzard did fix the DPS problem by increasing the duration on the buffs, increasing the buffs, and adding Astral Empowerment, but that wasn't the ability's only problem.

The original goal was to spice the moonkin rotation up a little bit by making it more dynamic and adding a little more choice. Unfortunately Fae Empowerment did neither. Instead of making the rotation more dynamic it added another predictable button to an already very predictable rotation. In the end, FE became a very complicated ability with three different buffs that didn't add anything significant to the rotation.

Given how close we are to the release of Mists, I think it is important to give Blizzard some credit for scrapping this change and trying something new. It would have been easy for Blizzard to move forward with Fae Empowerment despite it's problems, and force us to deal with it. No matter how the new changes turn out, we have to give Blizzard some credit for trying something new this late in the game.

The Big Changes:

MMO Champion has datamined most of the changes and you can find them here on their web site,  but there are three main changes that I want to focus on.

  • Moonfire and Sunfire are now two different buttons. Plus, both spells have been buffed and the duration of both spells has been increased to 18 seconds.
  • Critical Strikes from the nukes will now extend the duration of one or both DoTs by 2 seconds depending on the school of the nuke and DoT. Nukes no longer refresh the DoT for the full duration during Eclipse.
  • Euphoria doubles the amount of Eclipse energy generated outside of Eclipse.

There are a lot of other little changes as well but these are the main three that will significantly change how a Moonkin is played in MoP. Don't be at all surprised if we see more balancing changes to other spells and probably the Mastery bonus as well. As a result, it's hard to say definitively how this will affect the rotation since we don't know what the final numbers on all of these spells will be, but there are really only two ways it can go: DoT Clipping and No DoT Clipping.

No DoT Clipping: If the moonkin rotation returns to the traditional strategy of letting the DoTs run out before you cast them again, this could be a pretty interesting rotation mechanically. The core of the rotation would be the nukes generating and consuming Eclipse energy, but the DoT casts would vary because we can't predict how often DoTs would be extended. Not only that, since Starfire only extends Moonfire and Wrath only extends Sunfire, then the two DoTs will quickly fall out of sync and would make tracking the DoTs more challenging.

In theory this sounds great because it would create a fairly dynamic rotation that I think most players would enjoy. However, I'm not sure this play style would generate the best DPS give what we've seen in Cataclysm.

DoT Clipping: What we've learned from Cataclysm is that Clipping a DoT is sometimes the best option if it causes toe spell to receive a significant buff that it wouldn't have otherwise. This is why it is often a good idea to recast Sunfire and Insect Swarm as we leave Solar Eclipse even if they have just 5 or 6 seconds left. Having the Eclipsed DoT up for an additional 12 seconds over the normal version could be worth the GCD it costs to cast, and that could be a big problem for the "fun" of this rotation. It would also completely undermine the benefit of having Crits extend the DoTs.

The problem is that Eclipse could act like a timer and determine when it is best to recast DoT rather then the duration of the DoTs themselves. For example, if Sunfire has just 5 seconds left as you leave Solar Eclipse Clipping the last couple of ticks off the spell is likely a good idea if the Eclipse buff is strong enough. What if you are entering Lunar Eclipse but you already have a normal Moonfire on the target with 10 seconds left? If you clip the moonfire then you are cutting off more then half of the DoT. If you don't clip, then your Starfire Crits would be extending an inferior version of the DoT.

I don't know the answers to these questions, but if Clipping is frequently the best option in MoP, then the rotation will be very static and boring rather then the dynamic rotation I illustrated in the No DoT Clipping strategy.

How to Avoid Clipping:

The primary cause of the clipping problem is the strength of Eclipse. If the Eclipse buff is too strong then the DoTs hit like a wet noodle without it. Therefore, we clip some DoTs to maintain the stronger version on the target. If Eclipse was weaker, the difference between a normal DoT and an Eclipsed DoT would be smaller. As a result Clipping would be less desirable because the benefit to clipping would be smaller as well.

My toon Graylow on the live servers currently gets more then a 50% buff from Eclipse, which I think most would agree is a very significant buff. That is why it often makes a lot of sense to clip in Cata. In the current Beta version of MoP Eclipse is is even more significant. Not only has the base buff been increased, but the per point increase has been buffed and a new raid buff was added that gives every raid member 5 points of mastery. If I copied Graylow over to the Beta servers today, Eclipse would buff his spells by 76% with full raid buffs.

Blizzard knows that the MoP Eclipse numbers are too high and I know that they plan on nerfing Eclipse at some point in the future. The changes to Euphoria alone made a nerf to Eclipse necessary since it significantly increases the uptime of Eclipse as a whole. However, I would argue that a more significant nerf is necessary then one that would be based solely upon the increased uptime of Eclipse. In fact, I think that the Eclipse buff needs to be weaker then what we currently see on Live. The base buff needs to be less then the current base of 16%. Total Eclipse should provide fewer then 8 points of Mastery when it is trained at level 80. And the Mastery per point increase should be less then the current 2.0%.

I don't know what those numbers should be. It's actually more complicated then it sounds, because you have to adjust it in such a way that doesn't destroy the value of the Mastery Stat. However, if Blizzard wants to avoid widespread clipping in MoP then they need to nerf mastery more then they are probalby planning./

Other Thoughts:

A Movement DPS Improvement: I'm not sure if this is an intended outcome of the change or a happy accident but this actually helps a Moonkins movement DPS a little. As you guys know, Lunar Shower and spamming Moonfire was intended to be the primary source of movement DPS for moonkin in Cata. Unfortunately, there are several problems with that strategy and it didn't work out as in tended. One of the big problems is that spamming an uneclipsed Moonfire would overwrite an Eclipsed version of Sunfire or Moonfire.

Changing Moonfire and Sunfire to be two separate buttons eliminates that problem. With a little heads up play a player can realize that he Sunfire on the target is buffed with Eclipse and therefore they should spam Moonfire if they have to move and not worry about overwriting the Eclipsed spell. Granted this isn't that big of an improvement because Lunar Shower still has other problems like ramp up time, but it does help a little.

Is Crit the Stat of the Future? Crit has been the least favorable DPS stat for most of my WoW career. There was a brief period in WotLK were it was better then Haste, but that was the exception rather then the rule. However, Blizzard has made three changes that should make Crit Rating a much more desirable stat in MoP.

  1. No Crit from Int: In case you didn't know (and I assume some of you don't) Intellect no longer increases you a players Critical Strike chance in MoP. While I know some of you will bemoan the loss, but this is actually good for the value of crit since Crit Rating has very high diminishing returns. This will lower a players base crit chance and as a result increase the overall value of Crit Rating.

  2. Shooting Stars based on Crit: Having Shooting Stars procs based off of DoT crits rather then DoT ticks increases the value Crit rating for obvious reasons. Having two DoTs again actually makes this impact more significant now because more DoTs means more Crits. That said, I don't expect this impact to be huge since Shooting Stars has only a 30% chance to proc on the crits.

  3. Nuke Crits Extending DoTs: Again, if Nuke Crits extend the duration of the DoTs then that obviously improves the value of Crit rating. That said, this impact will be significantly diminished if Clipping happens frequently in MoP.

I don't know where crit will fall in the stat hierarchy in MoP, but I won't be surprised at all of Crit becomes more valuable then Mastery on a point for point basis. My guess is that Crit will end up falling between Haste and Mastery as the second best Secondary DPS stat. I will have to do some testing on this later.

The Lunar and Solar Similarities: In Cataclysm Lunar and Solar are quite different, and Solar ended up being the better buff because of the number of spells it affected. In MoP Blizzard has made several changes to correct that. Hurricane now has a arcane version called Astral Storm. Insect Swarm was eliminated balancing out the DoTs, and now that Moonfire and Sunfire are two separate buttons it is difficult not to notice how similar Lunar and Solar have become.

I know that number wise Lunar favors single target DPS and Solar favors multi target DPS, but in terms of structure there are really only two difference left. The first is that each has a unique spell that they buff. Lunar has Starfall and Solar has Wild Mushroom, but all of the other Druid DPS spells have version for both types of Eclipse or are buffed by both types of Eclpse.

The other difference is the cast time of Wrath and Starfire and as a result the amount of energy each of them generates. While they still have different cast times, that difference has been shrinking. After talents Wrath had a cast tome 1.5 seconds and Starfire was 3.0 seconds prior to Cataclysm. In Cataclysm the difference is much smaller. Wrath now has a base cast time of 2.0 seconds and Starfire's base cast time was lowered to 2.7 seconds. Wrath's cast time was increased to avoid creating a haste cap, and Starfire's cast time was lowered to make the Eclipse energy math work out better.

I agree with both of those changes, but two spells that used to be 1.5 seconds apart are now just 0.7 seconds apart. I now wonder why there is any difference at all. Is there any situation in PvE or PvP where Wrath has an advantage for being 0.7 seconds faster? Why not just increase Wrath's cast time and energy generation to match Starfire. This would make the Eclipse energy math much easier since all spells would then generate 20 energy.


Anonymous said...

> While I know some of you will bemoan the loss, but this is actually good for the value of crit since Crit Rating has very high diminishing returns

AFAIK there's no diminishing returns on Crit rating. The issue is that there are no _increasing_ returns on crit: therefore, when you get 'free' crit, you will be less likely to get additional crit because that 'free crit' increases the marginal value of all the other stats.

The point is not that crit gets less valuable the more crit you have, but that other stats get more valuable the more crit you have. That's generally true for all stats

Graylo said...


Yes and No.

Most of what you said is correct but your not accounting for other factors as well.

First off, The stat Crit Rating does not have any deminishing returns itself. For example in Cata 179 Crit Rating will always give you an additional 1% chance to crit if you start with a 10% chance or a 90% chance. However, the Value or Impact of Crit Rating has significant deminishing returns.

Part of it is as you say that when you get more Crit Rating that increases the value of Haste and Mastery because you are doing more damage and that makes a little more mastery or haste better, but that's not the whole story.

What you're missing is that Crit is a binary event. You either Crit or you don't. So gaining more Crit Rating only helps a portion of your spells. Here's an example.

Lets say you have 0.00% crit chance and you cast 100 spells that do 100 damage each. I will also assume that Crits only double the damage to make the math easier.

If you gain 1% of crit chance then 1 of the 100 spells would crit and your damage would go from 10,000 to 10,100 an increase of 1%.

Now lets say you start with 50% chance to crit and you gain an additional 1% chance to crit. With just 50% chance to crit you would do 15,000 damage. By adding an additional 1% chance to crit you would do 15,100 for just a 0.667% increase in damage.

The problem is that gaining more crit rating has no impact on the spells that would have crit without the additional crit rating.

So the deminishing returns isn't just that haste gets better when in increase Crit. It is also that Crit gets worse marginally when it is increased.

Haste and Mastery also get worse marginally as they are increased, but their rate of decline is slower then that of Crit. That is why Crit has higher deminishing returns then the others.

Andrewida said...

The difference between cast times can be useful if we have "on use" trinkets, that gives us +fix sp(instead of dmg%). We can earn(use the bonus dmg more often) if we are casting the faster spell than the slower one.

If u cast SF every 2.7secs and the bonus dmg is active on every hit, then u can have 3x bonus-dmg with SF, but during the wrath the bonus number is 4x.

Another reason - its more safe to use those trinkets with wrath in a "wasd" situations(pve/pvp).

The difficult thing is to timing those :(

Regards :)


hibbynana na said...

wow, this is a great post. the problem is that gaining more crit rafting has no impact on the spells that would have crit without the additional crit rating

hibbynana na said...

wow, this is a great post. the problem is that gaining more crit rafting has no impact on the spells that would have crit without the additional crit rating
where to sell world of warcraft accounts

Azrael said...

Int still gives crit. I just checked it myself on the beta server.

It only gives an extremely small amount of crit, though. You need over 2000 int per 1% at level 90.

Lufitoom said...

Love your number crunching! Of course I am doing the PvP thing and loving every bit of it in beta.

Tagartou said...

Looking forward to what you have to say about the even more recent changes, mostly the DoT duration reduction again.

Related to the topics in this post, I think that it looks like we will be going back to no DoT clipping(especially with the nerf to eclipse a few builds ago) and I can't be happier. It was a pita mechanic and I'm glad to be done with it.

I also am fairly sure that crit will be our main stat in MoP. It's looking to me that we will be reforging to a haste breakpoint and then going all crit after that.

Always enjoyable to read your perspective on things.


Anonymous said...

Dear Graylo,