Jump to content


Photo

Moonkin PvE Discussion, 3.3.3


  • This topic is locked This topic is locked
328 replies to this topic

#1 Hamlet

Hamlet

    Mike Tyson

  • • Guide Author
  • 11,540 posts

Posted 05 March 2010 - 01:38 AM

Starting a new discussion thread that's patch-current (old one will be locked shortly). All topics related to current Moonkin raiding go here.

My Moonkin raiding guide is maintained here:
http://elitistjerks....ated_3_3_3_ptr/

Anyone posting should have good familiarity with the current state of Moonkin theory, either from this guide or from anywhere else. At the very least, do not ask questions that are directly answered in the guide; they will be infracted.

------

To get started, the old thread was talking about the incoming 3.3.3 changes (I should have started this one a little earlier to avoid breaking it up). The main issues are:
--Which Glyphs should be used in addition to Starfall--Starfire/Moonfire/Insect Swarm/Focus?
--When in the cycle should DoT's be refreshed?

#2 Guest_aceofsween_*

Guest_aceofsween_*
  • Guests

Posted 05 March 2010 - 03:24 AM

These questions have several underlying questions that need to be answered first.

What is your goal?
Is highest possible DPS your priority?
Are you required to keep unglyphed IS up?
Should we use 1 DoT or 2 (unless moving)?
Is it feasible to stand within 18 yards of the boss?

That being said, I think striking the best balance comes with the Starfall/Starfire /Moonfire combination. It provides the highest DPS combination without resorting to the Focus Glyph. Starfall/Focus/Insect Swarm will provide you higher DPS, but it comes at a pretty stiff cost. It effectively limits you to 18 yards and prevents you from making use of the IS miss debuff, although that may be completely inconsequential if you are only applying a non-glyphed DoT while moving (using a 1 DoT rotation).

As far as DoT refreshing is concerned, I believe that is a simple case of measuring the DPET of your DoTs vs. the DPET of your Nukes and seeing which one comes out higher. Insect Swarm is infinitely easier to figure out as it's duration is constant while if you are glyphed for Moonfire, you must cast 3 Starfires to get the full benefit of the DoT. This is going to vary for everyone. With enough crit, Wrath will outweigh a Glyphed IS during Solar Eclipses, although speccing into Improved Insect Swarm will enable you to offset this minor damage loss, provided there is enough time left on Eclipse. The same is also true for Moonfire: with enough Crit, Wrath will outweigh Moonfire, unless it is fully extended. Therefore, refreshing Moonfire is dependent on when you will cast your next Starfire and you must properly time your refresh so that you can be sure it will be extended. During Lunar Eclipses however, a glyphed Insect Swarm never has a higher DPET than Starfire while a glyphed Moonfire (extended) almost always does, although with enough Haste, even Moonfire could be surpassed (during Bloodlust/Heroism, for example).


Finally, there is the idol proc to consider and I believe the consensus is that it should never fall off for any reason. Therefore, if you are using a 1 DoT rotation, you must reapply IS at least once every 27 seconds (it has to tick once to refresh the Idol) or MF once every 27 to 36 seconds (depending on if it is extended or not). Most of the time, I don't think this will be a problem. The biggest concern would be with Moonfire, but normally when first applied it lasts into a Solar Eclipse, giving you more than enough time to reapply Moonfire, even should you choose to refresh after the Eclipse ends. If you refresh during a Solar Eclipse, but time it well enough so that you will be able to extend it by proccing a Lunar Eclipse, then you will have to refresh it again during your Eclipse cycle, and it will again last into the Solar Eclipse phase, repeating the cycle. It's hard to predict this with complete accuracy though because we all know how finicky proccing Eclipse can be at times. In any event, the answer remains fairly straight forward: if ever there comes a time where your Idol proc is about to wear off, reapply your highest DPET DoT.

#3 Valardruid

Valardruid

    Glass Joe

  • Members
  • 13 posts

Posted 05 March 2010 - 08:18 AM

DoT refreshing is not as simple as just comparing DPeTs. Plausible numbers for the sake of illustrating the point:

IS: 9000
wrath under eclipse: 8900
wrath under eclipse/is: 9100
wrath under IS:7100
wrath: 7000

Even though insect swarm has a higher DPeT than an eclipsed wrath, choosing to apply insect swarm over an eclipsed wrath could be the wrong decision. (Not considering the hit debuff) Say you have time for three globals left in your eclipse, and assuming gcd-capped wraths.

Option 1: cast three eclipsed wraths, then refresh IS, cast 3 wraths to proc lunar
Total damage: (3*8900)+(9000)+(3*7100)=57000
57000/7=8142.9 dps

Option 2: refresh IS, cast two eclipsed wraths, cast 3 wraths to proc lunar
Total damage: (9000)+(2*9100)+(3*7100)=48500
48500/6=8083.3 dps

This is still an over-simplification, but the underlying cause of a dps gain by choosing a lower DPeT spell is due to the concept of opportunity cost. By refreshing insect swarm out of eclipse, you give up a lot less (a non-eclipsed wrath) than if you refresh it during eclipse (an eclipsed wrath). This is one of the reasons DoT refreshing rules are not rigid, and modeling the top DPS refreshing pattern is impossible, it depends on too many things. This is why only general refreshing rules are used.

#4 Guest_aceofsween_*

Guest_aceofsween_*
  • Guests

Posted 05 March 2010 - 09:36 AM

Perhaps I didn't explain that point as thoroughly as I have in the past thread, but this was something I have already illustrated, and in fact mentioned in my response.

With enough crit, Wrath will outweigh a Glyphed IS during Solar Eclipses, although speccing into Improved Insect Swarm will enable you to offset this minor damage loss, provided there is enough time left on Eclipse.


Basically, 3 GCDs left is not enough time nor have I ever stated that it would be. Let me illustrate my point more thoroughly (again... since it'll be lost amidst the locked thread). However, let's do this using more accurate numbers, and again I'm going to use the values from WrathCalcs (with my gear) as my example.

Average GCD: 1.1
IS: 9646
Wrath: 8290
Wrath under Eclipse: 10971
Wrath+IS: 8511
Wrath under Eclipse+IS: 11273

Let's assume Eclipse has enough time remaining to cast 7 GCDs and can either refresh it now, or wait:

7 Wraths under Eclipse, Refresh, 3 Wraths to proc Lunar:
(7*10971) + 9646 + (3*8511) = 111976 damage, 9254.21 dps

Refresh, 6 Wraths under Eclipse, 3 Wraths to proc Lunar:
9646 + (6*11273) + (3*8511) = 102817 damage, 9347 dps

Net difference for immediately refreshing: +92.79 dps

Conveniently, 3 GCDs is the breaking point (was that a coincidence?). If I have 4 GCDs or more, it is better to refresh IS immediately. If I have 3 GCDs or less, then it is better to wait until Eclipse expires. I want to stress this is only for me, personally. It will probably be a little different based on various gear setups. Also, a quick note... The damage added by using 3 Wraths to proc a Lunar Eclipse is irrelevant, because it's the same in every situation. It only serves to normalize the DPS. If you drop that damage off, the DPS fluctuates even more.

EDIT: I also wanted to add that not only does this setup already provide higher DPS, but it also reduces Eclipse downtime since you don't waste a GCD refreshing a DoT outside of Eclipse.

#5 Valardruid

Valardruid

    Glass Joe

  • Members
  • 13 posts

Posted 05 March 2010 - 10:42 AM

Conveniently, 3 GCDs is the breaking point.


This was the exact point I was trying to make. A simple DPeT comparison is not sufficient in evaluating the best use of a global. There is a point where the increased opportunity cost of an eclipsed wrath outweighs the gain of having insect swarm up for the remainder of that eclipse.

As a side note: It seems my post was viewed as some sort of attack on or counter to the previous post, which was not intended.

#6 Valardruid

Valardruid

    Glass Joe

  • Members
  • 13 posts

Posted 05 March 2010 - 06:45 PM

That is still simple DPeT comparison. It's just a broader scope than looking at just Eclipsed Wraths and just Insect Swarm's values.

To maintain clarity and transparency throughout this thread, one term should only mean one thing. Using this broader scope, the bonus from NG uptime due to starfall should be considered in starfall's DPeT, tarfire's DPeT increases if moonfire is able to be extended, and the DPeT of IS increases if there are wraths to follow. In this manner, a significant amount of damage will be counted twice, and average DPeT retains no relation to dps.

#7 Pokerkin

Pokerkin

    Glass Joe

  • Members
  • 8 posts

Posted 05 March 2010 - 07:21 PM

I've been playing around with Focus/Starfall/IS the last week in raids to be better ready for when the patch rolls around as that is the glyph setup I am planning on running. I've been trying to experiment with different rotations to see what works best in practice as well as how the focus glyph range reduction impacts how I move, react and use starfall and how much that changes the rest of our abilities.

I got on the PTR the other day and tested the new starfall change. On live servers, my starfall is doing anywhere from 15-19k per one cast, while on the PTR, starfall is doing 55-60k per one cast. This is an increase of roughly 40-45k per cast which should equate to a dps increase of 750 dps or so (45k/60=750). This information is already known for the most part, so I'm kind of just reinforcing the fact here.

IMO the range reduction on starfall is actually very nice. Yes, it does require us to stand closer which is the direct opposite of one of a caster character's main advantages; however, I feel that it gives me more control over what I want to hit with the starfall so I can put more concentrated damage on a specific target(s) if I want to.

The rotation I have settled on is to start with FF, IS, Starfall, then the normal eclipse rotation. I've found that throwing both DoTs(to get the idol rolling quicker) on the target is actually less dps than just starting in on some wraths and hopefully getting an eclipse within the first few seconds of the fight(1 wrath > 1 MF). I refresh IS at the end of Lunars (normally during the last 1.5-2 seconds of lunar eclipse when the eclipse is effectively over). I also refresh IS at any point during the solar eclipse(more on this later). My crit is more than over cap without MF up so that MF really has little effect on proc times of my next eclipse or on the lunar eclipse phase. I do have IIS (although I've been fidgeting with either 2 or 3 points in it and 2 or 3 points in OkF) so my wrath damage does go up with IS on the target.

I have had no problems with idol buff falling and I watch powerauras (which I have set to showing how many seconds until the idol buff drops) in case I get REALLY REALLY terrible rng and have to reapply at an awkward time.

The math in the last few posts is a little sketchy. I ran some numbers this morning to prove that IS refreshing during a solar is actually better than 1 wrath cast IF you have IIS. These numbers are with 3 pt IIS and are weighted averages of wrath depending on crit%. I pulled the data from my last WoL parse for average hit and crit numbers and percentages on bosses fights.

1 IS cast = 7 ticks*1780 damage = 12460 damage = 1 IS cast
Wrath Crit average = 12965, Crit % = 71.2%, Weighted Wrath crit damage contribution = 9231 damage
Wrath Hit average = 6642, Hit % = 28.8%, Weighted Wrath hit damage contribution = 1913 damage
Overall Wrath Weighted average = 9231+1913 = 11144 damage <= This is the overall wrath number usd for comparison.
I weighted the IIS additions to see what wrath would hit for without IS up for the majority of IS casting(using crit% and the multipliers for IIS[1.03 for hits and 1.06 for crits respectively])
Results : Wrath weighted average (w/o IS) = 10600 damage meaning on average 544 extra damage per wrath occurs with IS up.

With the rotation stated at the beginning of the post, IS is up nearly 100% of the time I cast wraths as I cast it whenever there is a heavy solar part of the rotation. The only factors not considered is time spent casting wrath inside solar eclipse versus outside solar eclipse. Even with very pessimistic values, IS is worth more than Wrath is on a strictly DPCT perspective.

#8 feior

feior

    Glass Joe

  • Members
  • 24 posts

Posted 05 March 2010 - 08:30 PM

I believe that is a simple case of measuring the DPET of your DoTs vs. the DPET of your Nukes and seeing which one comes out higher


To be perfectly honest, I don't really know what your purpose or intent was for posting.


The intent of Valardruid's post was to say that this is not always true and he constructed a plausible example to show you that this isn't always true.

When in the cycle should DoT's be refreshed?

I tend to use MF at about 6 seconds before solar eclipse finishes if MF isn't up near this time. It takes me about 4'ish casts to proc lunar eclipse (assuming 30% chance to proc per Wrath for 3.33 casts on average). Supposing none of those wraths were NG'd, then I'd still have about 15 -4.5 (Non-NG'd Non-eclipsed Wraths) -6 (time casted before eclipse ended) = average of 4.5 seconds left of moonfire during lunar eclipse. Extending the 4.5 seconds worth of moonfire gives me 13.5 seconds during lunar eclipse (forcing it to fall off before eclipse is over). This means I can reapply it while trying to proc solar eclipse.

Obviously this won't work if it takes more than 6 wrath casts to proc lunar. I'm not sure if the gcds are better placed elsewhere, but this seems to be a good solution.

I've found that throwing both DoTs(to get the idol rolling quicker) on the target is actually less dps than just starting in on some wraths

I personally throw out both dots so I can swap between SF and Wrath and use whichever eclipse procs first. Throwing out IS first to get the idol rolling faster. With really bad RNG, I tend to do this again if both eclipse are off CD. I heard swapping of Wrath/SF is better in this scenario, but who knows... maybe things have changed.

#9 Erdluf

Erdluf

    Great Tiger

  • Members
  • 960 posts

Posted 05 March 2010 - 10:26 PM

Pokerkin,

You said that during Lunar, you would cast IS when there isn't enough time to complete another SF within Lunar. I argue that you should usually cast two more Starfires before casting IS (this assumes you are in no danger of losing the Idol bonus). By delaying IS, the Nature's Grace "left over" from Lunar is spent on another SF (where it does more good), and IS lasts longer into Solar (avoiding refreshes during Solar).

In general, you might have to move in the next four or five seconds. If it happens, saving your DoT for movement was probably a good thing. The rest of this post assumes you don't end up having to move.

Looking at the next several casts (including the last Lunar SF).
SF = crit
sf = non-crit
Sf = don't know.
IS = insect swarm
Red is a spell gaining a NG benefit. (2s or 1.1s)
Black is a spell without NG (2.4s or 1.2s)
I'm going to assume the spell before these sequences was a Lunar SF that crit.

For each sequence I will list:
probability%
Time till start of first Solar Wrath.
Time in Solar when IS fades.
Time "wasted" at beginning of Solar before starting first Wrath

with your strategy (SF crit at 90% Lunar or 50% otherwise):

SF IS sf sf Sf : 22.5%. 11.9s+x, 7.1-x, 2s.
SF IS sf SF Sf : 22.5% 9.5s. 9.5s, 2s.
SF IS SF Sf Wr : 45%. 8.2s. 11.9s, 2s
sf IS sf sf Sf : 2.5%. 12.3s+x. 6.7-x, 2s
sf IS sf SF Sf: 2.5%. 9.9s. 9.1s. 2s
sf IS SF Sf Wr: 5%. 7.5s. 11.5s. 2s

New strategy: If the last Lunar SF crits, cast two more SF before IS. If the last Lunar SF does not crit, cast one more SF before IS.

SF sf sf IS Sf: 22.5%: 11.6s+x. 11.4s-x. 2s
SF sf SF IS Wr: 22.5% 7.1s. 14s. 1.1s.
SF SF Sf IS Wr: 45% 7.1s. 16s. 3.1s.
sf sf IS sf Sf: 2.5% 12s+x. 7.1s-x, 2s
sf sf IS SF Sf: 2.5% 9.6s. 11.4s. 2s.
sf SF IS Wr Wr: 5% 5.1s. 14s. 1.1s

If I got those numbers right, on average:
You begin casting Solar Wrath about 1.2s earlier
IS above lasts about 4s longer into solar. 70% of the time, IS lasts into the last second of Solar (or goes beyond Solar).
You do lose an extra 1/4 s to non-wrath spells (on average) at the start of every Solar, but I think the earlier proc, and reduced need to refresh IS during Solar makes up for that.

Edit: You also lose some DoT uptime. I think you're still better off, but I haven't crunched the numbers to be sure.

#10 Pokerkin

Pokerkin

    Glass Joe

  • Members
  • 8 posts

Posted 06 March 2010 - 02:08 AM

I've thought about this a few times too, even with latency, my haste is high enough that I have time to cast the IS, a SF, and then another SF while still under the effects of NG.

For example: Lets say my last lunar SF lands with 1.5s left on the eclipse. I'm massively over the crit cap, so this is guaranteed to crit. I then cast IS and start another SF. This is a total of ~2.8 seconds which with my latency (which is usually pretty low) allows me to still get the second SF in at NG speed. It's still the same amount of SF casts at the end of the lunar eclipse as in your example, except that once the solar starts, I can start spamming wrath without needing to apply IS during the solar (which I proved that IS is worth refreshing during a solar; however, it is always better to have it rolling at any point you are not in the lunar eclipse phase). With my current state of crit, 2 SF is usually more than enough to proc solar, though we all know the sad state of moonkin rng. Theoretically, I believe that both of our situations would lead to similar dps. My method allows a little more uptime for IS which is going to be the strongest DPCT any time youre not in lunar eclipse. I also understand that your method allows you to see if you got a NG proc while you are casting IS, but the dps difference between starting the next cast and possibly waiting on latency or other factors to see if NG procced is so minimal that it's almost non-existant.

One small thing I noticed about your numbers: if I understand correctly, you assume that SF has a 90% crit chance in lunar and 50% otherwise. I'm not sure if thats your crit percentage, but I'm easily over 100% crit inside lunar (I'm actually somewhere around 105% with idol rolling, buffs, etc, and much greater than 50% outside lunar, so that would be a factor to think about in our differing rotations.

I'll have to play around with your theory though as it is a slight variation that I did not test

With respect to the movement and saving the IS cast, I plan ahead and know when I'm going to be moving and change my rotation accordingly, so the rotation I said I use is a general outline for stand still fights, not really meant to be utmost rotation that you have to stick to no matter what. I know how to plan ahead and I know the encounters very well, it's not like it's my first time around the block in a raid :P. <=not meant to be a jackass statement, but I couldn't think of a better way to put it.

#11 Knofle

Knofle

    Glass Joe

  • Members
  • 14 posts

Posted 07 March 2010 - 09:41 PM

I've been playing around with various rotations in the 3.3.3 beta version of the fairly new tool "Team Robot", found here: Team Robot Wow Simulator Beta, since this is a fairly simple tool to play around with stuff like rotations.

I personally don't know the accuracy level of this simulation tool to its full extents, but I did however find some interesting results in the various rotations and glyphing.

Please note that the equipment I used is my own, with 3028 Bonus Damage (3350~ counting item procs), 1017 crit rating, 652 haste rating, 274 spirit and 1308 intellect. All calculations are with full raid buffs (Totem of wrath, no demonology warlock). I do have 4setT10, but this is currently bugged with this tool, as it provides insufficient amounts of damage.

The "normal" cookie cutter rotation would be Wrath until lunar eclipse, starfire until wrath eclipse, and keep the dots up.
For this rotation i found this result with the different glyphs.

[TABLE]1: swarm,sfall,focus 10573
2: mfire,sfall,focus 10489
3: mfire,sfall,sfire 10607
4: mfire,swarm,sfire 10491[/TABLE]

The rotation I present is however a bit different, and started off as a thought experiment.
Eclipses go as normal, but to only apply insect swarm when solar eclipses pops, and moonfire only applied when lunar eclipse pops, or whenever insect swarm is not present.

[TABLE]1: swarm,sfall,focus 10614
2: mfire,sfall,focus 10634
3: mfire,sfall,sfire 10576
4: mfire,swarm,sfire 10356[/TABLE]

Secondly, we have had discussions about whether it just pays off to drop one of our dots, and I tried those too.
Dropping Insect Swarm:
[TABLE]1: mfire,sfall,focus 10563
2: mfire,sfall,sfire 10593[/TABLE]

Dropping Moonfire:
[TABLE]1: swarm,sfall,focus 10765[/TABLE]


This is all just data from a simulation model that hasn't been thoroughly enough tested, but it does give some food for thought, I guess.



edit: I guess my personal opinion would be to take any rotation without glyph of focus, at least for non-static fights, since I like to be as mobile as possible and keep my distance to the boss. This is a personal opinion though, and it might well change once more facts are presented.

#12 Hamlet

Hamlet

    Mike Tyson

  • • Guide Author
  • 11,540 posts

Posted 07 March 2010 - 10:34 PM

Robot is even nice enough to give confidence intervals, it's a huge waste to not include them in the results. They go a long way toward showing how meaningful any tests are.

#13 Jivv

Jivv

    Glass Joe

  • Members
  • 6 posts

Posted 07 March 2010 - 11:47 PM

The rotation I present is however a bit different, and started off as a thought experiment.
Eclipses go as normal, but to only apply insect swarm when solar eclipses pops, and moonfire only applied when lunar eclipse pops, or whenever insect swarm is not present.


So, are you suggesting that you use it RIGHT when the eclipse buff comes up, for either eclipse? So, the second one sees Lunar, you start the starfire spam right after throwing up moonfire, and vice versa? Also, so if there are no dots up, we refresh moonfire? Just wanted to test your rotation, but wanted to make sure that I knew exactly what you were saying before I tried it. I'm assuming that IS would fall with solar eclipse falling, so since there was no dot, and you said if there is no dot up, then to refresh MF, that means that we use MF right before the solar? If this is the case, say that moonfire falls mid solar eclipse, what would we do then? Thanks in advance!

#14 Hamlet

Hamlet

    Mike Tyson

  • • Guide Author
  • 11,540 posts

Posted 07 March 2010 - 11:50 PM

Also, why would refreshing a DoT right after Eclipse procs be better than refreshing right before Eclipse starts (i.e. when the previous one ends)?

#15 Knofle

Knofle

    Glass Joe

  • Members
  • 14 posts

Posted 08 March 2010 - 01:18 AM

Well, the thought behind it is that moonfire is most likely up when you get lunar eclipse anyways, since one should keep a dot up for the idol buff. When solar eclipse pops and you apply insect swarm, you get all of your eclipsed wraths within the swarm. That was the theory behind it, at least. As I said, it started off as a thought experiment, and I haven't tried it out properly in a raiding environment, but it looked quite promising in the simulations. I'm happy if someone proves me wrong, or finds a better rotation. Food for thought and thinking outside the box is needed to make final conclusions. :)

Robot is even nice enough to give confidence intervals, it's a huge waste to not include them in the results. They go a long way toward showing how meaningful any tests are.

I'm not quite sure about what the confidence intervals are in this simulator, but if you mean the +/- number behind the actual dps, then I can say that it was quite stable at around +/- 20-30.
The simulations were 7 minute fights, and a sample of 700 fights per test.


If anyone want to experiment for themselves, and improve/tweak the possible rotation, the way I set up the dot rotations in robot was:
Moonfire:
Only execute this action if ((NOT(Target has Insect Swarm) AND NOT(Target has Moonfire)
OR ((Have >=1 Stacks of Lunar Eclipse) AND NOT(Target has Moonfire)

Insect Swarm:
Only execute this action if: ((Have >=1 Stacks of Solar Eclipse)
AND NOT(Target has Insect swarm)

#16 Videl

Videl

    Piston Honda

  • Members
  • 111 posts

Posted 08 March 2010 - 02:42 AM

In the case of IS doing that gets you IIS for the entire solar eclipse exactly. Additionally, an extendable moonfire cast right at the end of wrath eclipse will pretty reliably last through the end of lunar eclipse. As you get back into the pre-solar phase there's a few situations that can come up:
If your eclipses both procced pretty quickly (4.5 seconds average) you only cast one moonfire for that cycle; it gets fully extended and it lasts through lunar eclipse and pre-solar.
If they proc slowly you cast a second moonfire during your pre-solar phase. If the starfire right before that moonfire procs eclipse in which case the moonfire doesn't get extended, but replaces a starfire that you would have cast during your solar eclipse.
If they proc slowly and the starfire before your moonfire doesn't crit the moonfire gets extended atleast twice and is up for the small crit buff to help proc solar.
All these situations seem pretty good if it's the case that the only dot worth casting for its own damage is an extended moonfire and otherwise you need to maximize their IIS effect to make them worthwhile. In the case of moonfire casting whenever IS wears off is very close to casting right when solar ends under these conditions though, so there's not much difference there.

If you cast IS right at the end of lunar you either proc solar quickly, in which case the IS doesn't wear off until nearly the end of solar and is not worth recasting (meaning less than 6 seconds left on your eclipse if I remember right from earlier posts). In that case you trade off the IIS boost to any wraths left in that eclipse and the first couple pre solar wraths for another eclipsed wrath (a very small gain at that 6 second mark, growing with faster eclipse procs).
If you proc slowly and IS wears off in time to recast it you gain the IIS bonus on any wraths in your next pre-solar phase after the first couple and the damage of another IS, but lose the IIS bonus on one eclipsed wrath and extend your whole cycle by a GCD. I assume that's a net loss since it is slightly less than you'd gain by casting IS right when solar procs then again right when that wears off which we know isn't desirable with an unglyphed IS.

For what its worth, when i first got 4T10 I set up some simucraft comparing a couple one dot rotations with full dot uptime and different glyph sets and this rotation was the best without glyph of IS (this was before starfall changes so full uptime with all glyphed dots was still best).

I really think this stuff has to be looked at with simulations over formulation tools because the small changes and variations in time to proc eclipse have pretty significant effects on dot uptime with this priority.

#17 Guest_aceofsween_*

Guest_aceofsween_*
  • Guests

Posted 08 March 2010 - 04:00 PM

Something about the Team Robot simulator bothers me...

According to everything that I know, have personally tested, and have witnessed for myself, casting an unglyphed IS results in a fairly noticable DPS drop, no matter when you cast it. I see a difference of about ~75 DPS on average while the WrathCalcs value shows a difference of 90 DPS. However, when testing this rotation out in the simulator, the results don't match up. Removing IS from the rotation in the simulator consistently results in a slight drop in DPS, although both results remain close to the margin of error. I consistently get values that range from 10730 to 10770 for a rotation without IS and 10770 to 10800 for those that include IS.

I don't believe that the WrathCalcs models are wrong, seeing as how I've noticed similar results on my own, but this is in direct contradiction to those results. Is the TRS engine wrong then?

#18 Erdluf

Erdluf

    Great Tiger

  • Members
  • 960 posts

Posted 08 March 2010 - 04:18 PM

One fairly significant difference is that WrathCalcs says IS never benefits from NG. That is too pessimistic. If you had 100% crit, every IS would benefit from NG, and IS would never cause you to lose NG.

Earlier versions of WrathCalcs assumed that IS behaved the same as Wrath for NG. That was too optimistic. IS can benefit from NG, but can't generate it. Casting IS increases the chance that a later cast will occur without NG.

I also believe Robot's default 'Lag' settings are more optimistic than WrathCalc's default 'Delay' settings. Higher delay favors long casts (SF). I'm not sure I know what Robot's default Lag settings are though. I think it remembers my last setting, and I don't know how to make it go back to the default.

Someone else posted that Mr. Robot undervalues the 4t10 bonus. If correct, that bug makes your nukes look worse (and IS better, by comparison).

#19 Ettarean

Ettarean

    Glass Joe

  • Members
  • 4 posts

Posted 08 March 2010 - 04:41 PM

I personally throw out both dots so I can swap between SF and Wrath and use whichever eclipse procs first. Throwing out IS first to get the idol rolling faster. With really bad RNG, I tend to do this again if both eclipse are off CD. I heard swapping of Wrath/SF is better in this scenario, but who knows... maybe things have changed.


Whether a 'twisting' rotation between Eclipses is optimal depends very much on your personal haste rating, glyph choices, and the distance between you and the boss. Assuming a case in which you run with SF/MF glyphs, if you are standing in melee range, the travel time of Wrath is negligible and thus your first SF after a Wrath has procced will benefit from Eclipse immediately, ensuring maximum use of the Eclipse period; there is the added benefit of ensuring that in the case of Eclipse being slow to proc, your MF will have been fully extended, which is not the case with a non-twisting rotation. A 'twisting' rotation is likely to be optimal in this situation.

The issue becomes more complicated as the travel time of wrath increases, and there is a point at which the combination of your haste and the travel time of Wrath causes your SF in the 'twisting' rotation to land fractionally before Lunar Eclipse procs. You will have queued your next spell to be a Wrath, and you lose some benefit of the 'twist' which you might previously have had. If you don't run with MF/SF as your glyphs, there might be no benefit received. Lunar Eclipse is preferred to Solar eclipse, too, and procs faster on average: it is rarely better to hope to use SF to proc a Solar Eclipse, unless there are peculiarities related to the fight mechanics that would make it so (you know you will have to move soon, for example, and thus a Lunar Eclipse will be wasted while you may gain some use from Solar)--and in those cases, you should be aiming to proc Solar anyway.

The above assumes that if you are not using a 'twisting' rotation, you are also not timing DoT refreshes to gain an advantage in seeing whether a cast has procced Eclipse to smooth your transition, nor that you are switching to a SF on a Nature's Grace proc from a Wrath. I argue that 'twisting', in real raiding terms (i.e. you are probably not standing in melee range, and post 3.3 you might well not have both MF and SF glyphs) has negligible benefits over a simple rotation (no clever timings on refreshes or switches)--if you compare it to a 'smart' rotation (using your DoT refreshes to enable your travelling Wraths to hit the target, or to break between Starfires in a post-Lunar phase to ensure that the next spell you queue is the optimal one, or even switching to SF on a Wrath NG proc, or after a fixed number of Wraths which are all in transit to the boss) then it is likely to fall behind.

#20 Guest_aceofsween_*

Guest_aceofsween_*
  • Guests

Posted 08 March 2010 - 08:14 PM

One fairly significant difference is that WrathCalcs says IS never benefits from NG. That is too pessimistic. If you had 100% crit, every IS would benefit from NG, and IS would never cause you to lose NG.

Earlier versions of WrathCalcs assumed that IS behaved the same as Wrath for NG. That was too optimistic. IS can benefit from NG, but can't generate it. Casting IS increases the chance that a later cast will occur without NG.

I also believe Robot's default 'Lag' settings are more optimistic than WrathCalc's default 'Delay' settings. Higher delay favors long casts (SF). I'm not sure I know what Robot's default Lag settings are though. I think it remembers my last setting, and I don't know how to make it go back to the default.

Someone else posted that Mr. Robot undervalues the 4t10 bonus. If correct, that bug makes your nukes look worse (and IS better, by comparison).


The default Lag setting for Robot is 150 (if memory serves). My latency tends to be in the 75 to 50 range, so in both Wrathcalcs and Robot, I've modified the value to reflect this. I'm aware of the 4t10 bug (as I'm the one who brought it up), but I wasn't aware of the way NG worked with IS as far as WrathCalcs goes. Even still, I don't believe the difference is so great that it would generate a swing of nearly 200 dps between the two models.

As far as the simulator goes, the best results that I have gotten are (avoiding Focus glyph) with actually using the glyphs of Starfall, Starfire, and Insect Swarm, but setting Moonfire not to refresh if Solar Eclipse is up and setting Insect Swarm not to refresh if Lunar Eclipse is up. Essentially, it reapplies the DoTs only if the appropriate effecting Nuke is in rotation currently. I haven't been successful in noting any fringe benefits for these refreshes (such as casting IS as Lunar Eclipse is ending or only refreshing IS if there are X seconds left on Solar Eclipse). It may be that the DPS gains for these tactics are so minimal that basic RNG wipes them out...

In any event, I was able to consistently achieve about 150 to 200 more DPS using this method than other glyph/rotation combinations.

EDIT: It should be noted that is you do use the Focus/Starfall/Insect Swarm setups, you will gain more DPS by dropping Moonfire from your rotation entirely. The difference seems to be ~60 DPS on average. More interestingly though, the difference between Focus/Starfall/IS without MF vs Starfire/Starfall/IS (using the rotation I mentioned above) is very minimal, a difference of ~30 DPS. I'm not sure if the 18 yards you lose from Focus is worth such a small difference in damage, but I'll keep tweaking things as I haven't really looked much into rotations using Focus yet.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users