My turn: Not having Imp. Faerie Fire! (3% crit for 3 points is a pretty sweet deal)
Id take a close look on your actual mana needs - Is 3/3 Intensity really needed? 1x Moonglow->1x Genesis is a small dps boost aswell but only if you dont need the mana.


While your conclusion is right, its an odd parse. He procced Eclipse with Wrath all 3 times though, so 49% critrate on Starfire is very bad luck & his IS uptime is costing him dps aswell.

I tested it with SimulationCraft.
2t7+4t7, IS/MF/SF glyph
Druid1: only cast IS when your next spell would be wrath
Druid2: do not cast IS at all
Druid3: cast IS if you do not have eclispse procced
Fightlength: ~280s (set to 300 and the simulator calculates all dmg done at the 150s mark and implies that this is exactly 50% of the HP, bloodlust at 35% then)
The above was also the dps ranking, but Druid3 was only about 1% less dps then Druid1 with Druid2 in the middle.
And Rokaz had only 14 IS dots, 2 casts, which he probably did while trying to procc eclipse.
He is just maxing it out, but yea, the RNG owned him with his SF critrate

Wow Web Stats - Patchwerk

Now I know that I don't have the perfect dream set yet, and I am wasting some possible spellpower because I am waiting on a few drops to get myself down to the hit cap, but here is my concern...

The WWS I see moonkins doing very well in all have them in the 4k-5k range, which is not bad. But the other end I see Hunters/Mages/Locks etc, all coming close to breaking 6-7k. So if your raid has good mages for example, they will be pushing 6k on a patch while a moonkin appears to be on top of his game at around 5k DPS. Maybe it is just my worrying too much, but this kinda scares me in an environment where we should be pretty close to other pure DPS classes.


I only put up IS if I am refreshing MF(and eclipse is not activated), or I am about to cast wrathh to proc an eclipse.

Finally got a decent lag free Patchwerk kill:
Wow Web Stats

Comments:
- My second eclipse took nearly 30 seconds to proc after ICD ran out even with 60% wrath crit rate (go to buffs/gains on my char, then eclipse -> browse log file and you can see the gap) so I lost quite a lot of damage there. Looking at the log, it's actually 8 crit wraths before eclipse procs.
- I was only casting IS when eclipse wasn't active, and even then wasn't keeping full track of when it ran out.
- Starfire crit rate was ~= to wrath crit rate, even with using a Starfire benefitting eclipse. I think it's mostly a good wrath crit rate rather than bad starfire crit rate though (~27.5% crit buffed, works out to about 52% expected crit rate with SF/Wrath).
- Bloodlust was cast at 35%, used treants about 10 seconds too early. No DPS warrior so I couldn't ask for BS either.
- I used a wild magic potion instead of a haste potion as I didn't realise I had some haste potions on me at the time.


As to scaling with Bloodlust - we do scale pretty well in general, the problem is cooldowns. Most other classes have some significant cooldown that can gain them a lot of damage from bloodlust (e.g. icy veins/combustion for mages) - the only thing we have is Treants which only gain 30% haste themselves. We have no 'personal' cooldowns to use to buff our damage. Not to mention that on a fight that short, randomness will affect your damage more than playstyle by quite a long way, although that's true of just about any fight length (such as my minute-long gap between eclipses). Randomness affects other classes more in some cases though, such as Mages who gain a lot from crit streaks but can have (relatively) terrible DPS if only every other spell is a crit due to the mechanics of Hot Streak.

@Dukes: How many Starfires do you cast outside of eclipse procs+cooldown? Do you focus on getting 3 Starfires in during every MF regardless of eclipse status?

@Ashaera: I can probably skip some mana talents. I haven't really bothered maximizing for that since Stormscale usually is unplayable, not unlike Al'akir I guess. I have that one try in the parse I posted because we scheduled a raid for 2200. We did the same this week, and it was unplayable again. Our kill this week took a minute longer, and I managed to cast ~10 spells less. I was mostly looking for things in my playstyle that aren't optimal, such as when to use pets like you commented earlier. So thanks for that

I use starfire as my main nuke, so I'm always casting it unless the eclipse ICD is <4 seconds, in which case I switch to casting wrath. I never try and make sure I cast 3 starfires after a moonfire, as the vast majority of the time wrath will always proc an eclipse prior to needing to (as you can see from the log, the average proc rate was 2 wrath crits ignoring the 30 second one), or the moonfire will be extended due to casting starfire under eclipse / within the ICD period.

Damage itself is not registered until Wrath hits. Most procs function this way as well. NG is a specific exception they coded in so that it would affect the next cast correctly. This leads me to believe that all of the relevant information is determined at that time, but doesn't take effect until it hits, but that's purely academic.

Wrath damage is done as the actual wrath 'ball' hits the target. The damage is calculated as the spell is released. Natures grace procs at the time of release, as this is when all calculations to do with the personal effects of the spell are done (crit or not crit / trinket procs), while debuffs on the target (such as E+M) are calculated at the point of impact. This is easily provable through the use of clickable trinkets and by standing at max range chaincasting wrath in order to prove the effect of E+M (or similar debuffs on the target).

As starfire is an instant spell, all calculations are done at the time the spell finishes casting.

Going by this assumption, if i am on a IIS-MF-Starfire-Wrath(on Eclipse) rotation that would mean that the closest i am to a target, DPS should increase. I definetly have to try this on Patchwerk on two resets, with WWS to show for, standing as close as i can, and at max range with this rotation.

Not really. You're using SF to proc Eclipse, so there's no discrepancy between your cast finishing and it proccing to worry about. After that, if the bonus to Wrath is calculated at cast finish, it doesn't matter how far you are from Patch, you'll still get the same number of casts with benefit.

The only way standing close would increase your damage is if that lets you get in one extra wrath that hits the boss right before he dies, as opposed to him dying while it's still in the air. Not worthwhile.

Standing close to a boss can increase your damage if you manage to convince a rogue to use tricks of the trade on you.

Barkskin is being taken off the global cooldown next patch.

This should allow us to use it pre-emptively against pushback without losing any dps time, and makes nature's focus even less necessary. This will be nice if we can anticipate a shadow bolt volley or whatever.

Normally, I'd say that spamming WWS reports isn't an incredibly useful thing here. There's usually mitigating circumstances, gear differences, etc. that cause disparities between reports, so comparing them isn't going to tell you a whole lot. There are two reasons I'm not taking that stance today.

1.) I also have a parse! Wow Web Stats is my raid from Tuesday. We're nowhere near the top end (just under 4 minutes for a kill), so it gives a look at that perspective for those who're in my situation where gear/guild/etc. are still holding you back.

2.) With so many of you having posted them already, I decided to do some comparisons anyway. Late last week I wrote another spreadsheet with a very specific purpose in mind: Understanding WWS parses. I may drop this in the thread for WC at some point once I get it polished a bit, but the functionality's pretty simple. You input your stats from a Patchwerk parse (actually, any parse works, but for obvious reasons Patch is best), and it gives you a listing of information including %time spent casting, average latency, and IS/MF uptime%. It takes into account haste on gear (not procs), Lust, and Haste talents/WoA. It doesn't care about damage at all, in fact. After running all the parses on this page through (including looking up everyone's armory for haste/talent numbers and using wow-heroes for a sense of gear), here are all the results:

Name	Gear	DPS	Cast%	Avg. Lat.	IS%	MF%
Hamlet	2282	5034	79.5%	.396s	57.5%	82.6%
Dukes	2256	4913	88.0%	.210s	50.1%	90.9%
Eilt	2130	4114	83.5%	.319s	43.8%	90%
Rokaz	2293	4784	84.9%	.289s	24.6%	89.5%
Superku	2234	4508	80.7%	.365s	79.5%	85.7%
Adoriele	1898	3650	96.3%	.084s	80.7%	85.7%
Randomsmo	2312	5868	91.2%	.165s	53.0%	90.9%

Now, %cast adds a default .2s to your cast times for latency, so it's a little incongruous, hence why I added Average Latency to the equation (basically, time you spent not casting/number of casts). And the gear scores are, I believe, just a sum of iLevels, plus a set amount per enchant or so, so it's not really going to check how optimal the gear actually is. But, for example, take Randomsmo vs. Rokaz. Similar gear level, about 20 ilvls difference, but due to Rokaz' abysmal IS uptime and about 5% less time spent casting, there's about 1000 DPS difference (yes, that doesn't factor in luck with crits, eclipse, etc. though if I remember, neither had particularly outstanding luck in either direction).

The sheet's attached if you'd like to play with it yourself.

Also take into account the boss timer. What I mean by this is, as it is usually a patchwerk link, how long it took them to kill the boss. The shorter the fight, typically the higher the dps because of bloodlust/heroism having a higher percentage of uptime during the fight.

There are parses out there of 2'17" patch kills and also I'm told though not seen, one under 2mins. Thus about a 3rd of the fight with BL active. With those, I've seen Moonkins with 5.8k dps. Which, I'm assuming would mean about 5k with your typical 3-4 min fight.

Lastly, I'd also work on training your guild and your guild shamans to pre-announce bloodlust, thus giving you time to get your trees out just ahead of time. It may seem trivial, but if you think about it, they do about 6-8% of our damage on a boss fight like patch, and they are only up for 30secs. So in a true min/max setting, getting optimal use out of them is key.

Pretty sure the 5.8k dps moonkin got like 85% starfire crit rate, aka super lucky rng,.

Here is my crappy crit rate and pretty good dps

Wow Web Stats

The idea to compare parses in order to find out how different rotations perform under non-ideal conditions is interesting. However, a few more common variables should be included in any such table, because they influence caster dps just as much as gear level:

- presence / uptime of 5% crit debuff
- presence / uptime of 3% crit debuff
- presence / uptime of 3% damage multiplier
- presence of Totem of Wrath (or lesser spell power alternatives)
- presence of Wrath of Air
- FF use yes/no

A few of these variables will most likely be the same in every parse, but it's best to be sure, so the comparison is accurate. In Adoriele's parse, for example, there was no Totem of Wrath, and possibly not even a Wrath of Air (only one shaman and no death knight with Imp. Icy Talons, so shaman may have opted for Windfury Totem).

Patchwerk WWS
Here's my latest parse without obscene lag (this week's kill was avg of 500ms+. Using the spreadsheet I came up with the following numbers:

Name	Gear	DPS	Cast%	Avg. Lat.	IS%	MF%
Eothlorien	????	4434	85.8%	.260s	63.8%	77.5%

Lagged WWS
I added this WWS to just to show how lag hinders our rotation. I proc'd eclipse with Starfire so many times in this fight because the timers were off.

Name	Gear	DPS	Cast%	Avg. Lat.	IS%	MF%
Eothlorien	????	3824	72.0%	.506s	71.1%	75.8%

EDIT: grammar and latest (lagged WWS)

in fact, the starfire crit was 71%, thus not highly crazy lucky rng, but still good.

this is the WWS I found on MMO. Wow Web Stats

I'de like to see some more parses from him, I noticed they have 3 healers for their patch, so their fight is generally shorter, where as we have 7 healers in ours, not much longer of a fight, as our dps per player is higher, just less dps in total.

anyways ya, 5800 is very impressive, I would like to see some more parses to see how consistent his patch dps is

Well on such a short fight you are very RNG reliant -- Im sure 6k dps is very doable in a 2min kill, but at the same time you can play flawless & end up going 5k, if eclipse refuses to proc or your starfires arent critting as expected thats just too bad.

Over a 6minute Brutallus this somewhat evens out, but fights from 1.45->2.15 are not a good measuring point.

It's his only parse above 5.4k. 6k is definitely doable, I don't even have nearly ideal gear and I did 5992 dps on Noth(with low treant damage and obviously a high starfire crit %).

Yeah, I completely agree, I'm missing quite a few pieces, mostly haste wise, and am managing to do 52-5400 on patch consistently, so 6k is very doable. We'll see