thedopefishlives

I'm testing all sorts of variables with latency, talent builds, and glyphs. Rawr should be able to answer all of these questions for you when you get time to play around with it. One thing I haven't done yet, and I mean to as soon as I can get a solid answer, is to model the inability to queue spells that are GCD-clipped.

 

Erdluf

To find your "best case" lag for instants (and presumably NG'd Wrath) and non-instants, I'd recommend timing 50 FF casts, and 50 Roots casts (spells chosen because of low mana cost).

It would probably be bad do this while your server is heavily loaded (midnight tonight, through early Monday morning).

 

dedmonwakeen

Someone with a G15 volunteered to do some testing for me on beta: He tuned it to attempt to cast a low-mana instant spell every 10ms.

He ended up with an average gap between spells of around 10ms. This means that there is definitely no longer the old "latency penalty" of attempting to spam while under the GCD.

So.... while there are no queuing benefits, there are also no penalties for an "early" guess. The idea of spamming (even just while under the GCD) makes my stomach churn.....

For the purposes of modeling this in SimulationCraft I added an additional 100ms (optional) penalty on top of the specified lag used for spell queuing. Obviously, someone willing to use ScrollMouseDown (with modifiers for different spells) would probably be able to reach lag values on par with queuing...... but I'm not sure how many people are willing to go that far, hence my 100ms penalty for people just hitting the key 4-5 times right near the end of the GCD.

http://code.google.com/p/simulationcraft

 

• Adoriele

I haven't run anything for 80 yet. Figured that, while the window was small, a lot of people were asking about 70 so I should focus my efforts there. It also affords the ability to play with pitting talents against each other, as we're able to get a lot more of what we want at 80, only having very minor decisions to make.

A general word on Eclipse: It's definitely an increase. No doubt about it, that first point will grant you extra DPS, as long as you can take advantage of it. That said, it's very likely that you'll need an addon to do that, or a very good internal clock. You'll also need to be attentive. If worrying about Eclipse isn't your way of playing, be it because you can't take advantage of it, or because you don't enjoy the inconsistency of it, well, I'm sorry to say that you won't be topping damage meters and impressing your raid. Personally, I enjoy the RNGness of it, because I'll forget the bad days and remember the good. Playing a Moonkin takes skill. Playing any class and topping the meters takes skill (excepting Demo locks pre-3.0). Eclipse is just a very very obvious showing of it.

 

dukes

Surely Eclipse is a DPS increase regardless of if you pay attention to the internal CD? As long as you swap to the other spell once it procs, you'll see a DPS increase, although not as much as if you use it "properly" and make sure you always proc the correct version of it.

 

thedopefishlives

Rawr.Moonkin does not model proccing the "correct" version; that is to say, when you proc Eclipse, you switch from Starfire to Wrath for the duration and then switch back. It is definitely a DPS increase if you do this. I may consider adding an option later to attempt to model proccing the "correct" Eclipse buff, but for now, consider Rawr's numbers on Eclipse a conservative estimate.

[edit] Some further discoveries: I was incorrect in my conclusion that MF/SFx3/WxN was not viable at 80. It isn't viable at 80 _in level 70 gear_. After loading up a decent T7-level set, the Starfire glyph comes out as a DPS upgrade for both rotations. MF/SFx3/WxN is a clear winner at this gear level, also; I'm leaning toward the jump in spell hit and the almost absurd levels of crit and haste as factors. IS seems to be worth casting in almost every situation, save one: You're using Starfire as a main nuke (as under heavy amounts of haste), you have the SF and MF glyphs but no IS glyph. Lastly, it appears that spamming Wrath and switching to Starfire on an Eclipse proc may actually _not_ be higher DPS, as I indicated earlier. It appears to be situational. I haven't yet implemented the switch in Rawr that allows you to choose "smart" Eclipse procs.

 

yuppidoooooo

Has there been any effort in determining whether or not it's feasible to throw only 1 point in Eclipse to maximize the DPS benefit of other talents? It seems that with a decent amount of crit and 30 second downtime that there's not too much of a premium in filling the other 2 points, which lets you max out talents like Improved IS, FF and CF. I could be wrong though.

This is my planned build for moonkinning at level 80. One point in Eclipse seems to be enough, although I guess I could also switch 1 point from Furor into Eclipse if the rotation is lagging.

 

◊ Lord BEEF

Improved insect swarm is probably the weakest dps talent we have. It's hard to imagine it beating a second or even third point of eclipse.

 

• Adoriele

Well, it beats the third, at least (a lot of tests I've run show that the second is in fact very viable). Base talent tree looks like: Talent Calculator - World of Warcraft, final point placed in either iIS or Eclipse. Using a W-SF-SF rotation, triggering IS only when Wrath is ready. Everything else is as included in the basic raid_80.txt.



I ran a second test, assuming only one point in Eclipse prior to the final point decision, and Eclipse won out by about the same margin. Only 1000 iterations, blah blah, but the results seem solid enough.

 

lissanna

If your wrath procs Eclipse, and you switch to starfire - and starfire does not crit at all (or show an increased crit chance) during the proc, then Eclipse wasn't a DPS increase. Depending on how it's used, it could even be a DPS decrease, if it's doing something like causing your reaction times to be slower, or if you are using a really inefficient cast sequence with the talent, based on some of the cast sequences we did on testing dummies during Beta. At this point, it should be a DPS increase, but how much of an increase depends on a lot of variables (including player skill/latency).

 

lissanna

what if you dropped Brambles and picked up both Eclipse and Improved Insect Swarm? Also, why would you pick up Improved Faerie Fire, when most of the math I've seen shows it as a significant DPS decrease to cast faerie fire in your rotation? Even if you were trying to milk the 3% crit out of feral druids theoretically putting FF up for you (if we actually do get the bonus anyway), would that 3% crit be better than picking up Improved Insect Swarm or Eclipse?

 

• Adoriele

First, for every time you get no benefit from the SF Eclipse proc, there are dozens where you get a greatly-increased effect. I think I posted on the official forums that Eclipse makes it twice as likely to make every single SF crit, and 1/20 as likely to get no crits at all, or some similar numbers. There's simply on way for you to consistently not notice the benefit of Eclipse. So you might be down on the meters for a given boss fight. That's the breaks. Console yourself with the fact that you'll be much higher on all of the rest.

Second, the talent spec I used for that simulation wasn't anywhere close to what I would actually use. The entire point of it was to find a way to show the difference between a point in iIS and a point in Eclipse. Personally, I'm pretty sure I'll be taking both in my final raiding build.

As for iFF, it's going to be more beneficial than iIS in almost all cases. iFF affects all of your spells: Moonfire, Starfire, Wrath, Starfall, Hurricane, possibly FoN, Typhoon, etc. iIS only affects Starfire and Wrath, and only when your debuff is applied. You'll notice a slight increase on your Wraths from the difference between 3% crit and 3% damage, but that's likely going to be outweighed by the benefit to other spells.

 

lissanna

Yes, but Faerie Fire has to be on the mob to get the 3% crit from iFF, and applying Faerie Fire is a DPS loss for moonkin.

 

• Adoriele

People always forget the -armor for FF. If your Ferals or restos aren't keeping up FF, your raid's DPS will go up more than yours will go down. If you have a Spriest and someone keeping up whatever competes with FF's -armor, then don't spec iFF.

 

heptadragon

The MMO-Champion RaidComp lists FF as an "Armor (minor)" debuff, which is in the same category as locks' CoR and hunter pets' Sting. Considering iFF gives benefits to any caster DPS in your raid, I'd say it's worth the point investment over iIS, the more so in a 10-man environment. It's a different story if you're in a 25-man environment where some other sources of the "Armor (minor)" debuff are more likely present.

 

lissanna

In any 25-man setting, you are likely to have a shadow priest and a hunter/Feral druid to make imp FF completely obsolete. In general, I'm not going to recommend the talent until they increase the duration to 2 minutes in PvE (one of these expansions, *grumbles*).

It's possible that in some situations, it will be a worthwhile talent. Perhaps it will be more useful when we have dualspec options, so we can have one spec with the talent and one without it depending on what the raid makeup is.