I was hoping for something more, and I really hope they ain't satisfied with fury rotation. Having Slam! on 4 sec timer and causes GCD is horrible.

Hopefully this was just a quick fix and more will come in the next major content patch.

I'd like a bit of confirmation on how wether or not this CD change would change our rotation:

As I see it currently our priority of using abilities remains:
WW > Slam > Bloodthirst (Heroic Strike spam when you have more than 50-60 rage)
(Although the reducement to 20 rage for BT might lower this to 40-50 rage.)

Right?


Then another question, how does losing T7 bonus and getting the T8 bonus work into this, according to my quick theorycrafting it would appear this would change our priority of abilities to: WW > Bloodthirst > Slam? (which would make the 'buff' to BT alot bigger than it might first look like compared to the old rotation).

Would be nice if I could get confirmation or correction on this from someone that is a bit more certain about how our rotation / priority of abilities should be like.

If they have exactly the same stats then the 3.6 speed is always better... however, I am pretty sure that no two weapons in game have "exactly" the same stats so you need to be careful with that. Run it through Landsouls spreadsheet and it should give you a very definate answer. I learnt to my suprise the other day that duel Rune Edge were actually better than the 3.6 speed earthshaper MH and Rune Edge OH

I can't even begin to understand the correct mathematical approach to this new rotation problem.

It is always a DPS increase if you are inserting a slam attack into the rotation even if you are pushing bloodthirst back yet keeping whirlwind fixed or delayed by at most .5 seconds, as long as you aren't doing it enough to where your slam GCD use is near equal to your BT GCD use.

The limit on this rotation system is the GCD spaces available.

What I am thinking it would look like this:

With Bloodsurge after 2nd BT, we know that losing an entire slam is worse than 1/16 of a WW

time	action
0.0	WW
1.5	BT
3.0	Slam/Free GCD
4.5	wait
5.5	BT
7.0	Slam/free GCD
8.5	WW

We delay the entire rotation only .5 seconds for the chance to get a slam after the 2nd BT which can be 20% or 36% if a heroic is immediately adjacent to the 2nd BT. This table can also be used if a GCD is needed for sunder or demo shout during the expanded 8.0 second slice. What could also occur alternatively is having two slams in between the bloodthirsts thus pushing the 2nd BT back, but immediately follow with the next whirlwind.

Without bloodsurge or GCD use after 2nd BT

time	action
0.0	WW
1.5	BT
3.0	Slam/FREE GCD
4.5	wait
5.5	BT
7.0	Wait
8.0	WW

This slice only occurs when we do not have a slam opportunity after the 2nd BT which will be 80% to 64% of the time, depending on our heroic use on the adjacent swing next to the 2nd BT.

There is no leway for latency for the expanded slice, and a whole 1.0 second leway for latency for the unexpanded version. I am also comparing a lot of using slam versus delaying both BT or WW to see if any other pushoffs are legitimate for the use of slam!

Slower and more expensive? HS is easier to sustain if your MH hits less often and your OH more often.


Edit: I didn't see Landsoul's post before posting.
But I don't understand why using WW before BT. BT costs less rage and has shorter cooldown, delaying it 1.5 sec is 37,5% of BT lost while you only lose the half (18,75%) for delaying WW. And WW doesnt do double damage than BT. Also if when you start the rotation you have 20 or more rage but less than 25, Bloodthirst will be the only option. The only thing I see in favor of WW first it's that the space for Slam looks more optimal.

And in your expanded slice you don't mention you have to delay BT 0.5 sec on the next cycle. By the way nice fancy tables I don't know how to post it like that.


Prioritizing Slam (Expanded cycle following Landsoul's terms)
0 BT
1.5 WW
3.0 Slam or free GCD
4.5 BT (0.5 sec wasted)
6.0 Slam or free GCD
7.5 Wait
8.5 BT
10 Whirlwind 0.5 sec wasted
11.5 Slam or free GCD

Prioritizing Cooldowns Unexpanded Cycle (In simple is BT+1 Attack repeating to oblivion. 50% of times the other attack is WW the other 50% Slam, Deathwish, Berserker Rage, Heroic Throw, Sunder, Shouts or nothing)
0 BT
1.5 WW
4 BT
5.5 Slam or free GCD
8 BT
9.5 WW


Is not a clean rotation in any way. 8 seconds cycle in a world of 1.5 sec GCDs and having to squeeze 3 attacks doesn't work. As you see the BT cooldown reduction will be wasted everytime you have bloodsurge proc. There is only one moment after 2nd BT when WW is on cooldown we can use one Slam without collisions but if you had bloodsurge proccing from a HS, BT or WW at the start of the cycle it would have expired by then.

What I mean with 1 extra BT every 10, is every 10 bloodthirsts (50 seconds) not every 10 seconds. Which is what will probably happen in the rotation prioritizing Slam once you add human and latency factors.
I still have to figure out which will be the best rotation but I doubt it will be the one using BT every 4 sec and WW every 8. That leaves time to Slam only once per cycle (at 5.5 after 2nd BT) and at a moment the probabilities of having it available are not the highest.
The other rotation can slam twice per cycle delaying BT and WW by 0.5 sec.

Talking about cleaner I hope someone comes with a cleaner view.
I guess we'll have to go with one or the another depending the expected difference of Slams with one rotation or another, the expanded or unexpanded.
And the other variation to check is to use BT after or before WW.

I really believe this is the case. All the changes listed in 3.1.3 are very simple changes that really only consist of static values being changed and therefore do not require testing on PTR.

I imagine (more like hope) that more significant changes to Fury (like moving talents around or making Bloodsurge procs stack) are more difficult to implement and therefore will be need to pushed to PTR for testing once implemented code-wise.

So ive seen many people saying different things for rotation. Since the glyph came out for WW ive been doing WW>BT>slam never trying to push back either with a slam. My friend tells me that is incorrect and the order is BT first then WW but its ok to push back BT so that the slam doesn't get over written.

Now i did some tests on target dummies trying BT first and WW first and both came back relatively the same, not to much of a dps increase on either side to be noticed just RNG stuff.

My question is is it ok to push back BT's for slams? so they dont get over written? and with the new change of BT becoming 4 seconds how would it effect this rotation? I also have 4 piece tier 8.5 assuming that will change some things in your analysis

Your question is answered not 5 posts above yours.

I thought it was shown to be something like Slam > WW > BT > Slam. Meaning you should delay WW for a slam, but not BT because WW is twice as likely to proc bloodsurge which would be lost. My quick napkin math shows that would have been slighly higher than other rotations, but there isnt a huge difference.

I don't think there will be a definitive answer for this until we figure out how to handle the GCD first.

The last testing done that I can recall showed the chance of WW to proc Bloodsurge on a single target was no greater than that of BT. As such, the reason why we are more apt to delay WW is more that the longer cooldown results is less lost DPS (and thus a bigger gain from Slam).

Do you have a link to this testing by chance? I just find it odd that the number of procs of something with a static % chance to happen on a hit would not increase with the number of hits. Would seem like a bug.

Can you elaborate more on how latency will always do that to a rotational period?

I only partially understand the generalization you made about it.

I'm fairly sure both whirlwind hits are counted as one attack for most procs. I know this is the case for [Fury of the Five Flights].

Okay ... I got a few questions that I at first wanted to post in the Simple Questions/Simple Answers thread, but then figured they might be a bit too complex to be posted there, so I'll post them here.

Question 1: Precision vs. Enrage
Our BiS gear has pretty high hitrating, I'm currently wearing 5/5 T8.5 along with grim toll and several other BiS armorpieces and weapons and have 336 hitrating, which is about 6,5% above the yellow cap. My question was wether or not it would be a good idea to drop all 3 points in precision and put them in enrage instead. My guild is working on hardmodes, and raiddamage (from random aura or boss AoE attacks) is pretty high and constant in most hardmode fights, which would give Enrage a much bigger uptime than it has on normal modes. I'd like a few second opinions on this matter before I do make the switch.

Edit: It appears that alot of raid-AoE effects (like High Voltage on Iron Council, or XT-002's tantrums, etc.) don't even proc enrage. Making it seem like the answer on this question would be: "not worth it".

Question 2: Improved Berserker Rage
I've always been an advocate for taking 2/2 Improved Berserker Rage and weaving it into your DPS rotation. My concern however is wether or not it's still worth it once your gear gets unto a higher level and (perhaps) rage generation would start becoming a much lesser issue. (Also taking the rage cost reduction for BT into account in 3.1.3)

There is a certain amount of human error involved in including IBR into your rotation, it would be a bit optimistic to claim that it sometimes doesn't push back your WW or BT useage with 0,5-1 seconds. Perhaps it would be wiser to not include this in the DPS rotation anymore once your rage generation is sufficient, and to put a point in something else instead.

My idea, in combination with the explanation given after question1 above is to put 1 of these points in Enrage instead and the other one in Anger Management. This way you'd have 5/5 enrage, which at times, if the encounter allows high enough uptime, would be a considerable DPS boost.

Question 3: Rage Generation & Hit rating
There was a post a while ago in the fury warrior forum that listed all the "rage generating" talents (like UW, IBR, AM etc.) and how much rage/second these talents would give. I was wondering how hitrating fits into this as well, considering that capping hitrating always is the biggest DPS boost for most classes, I'd assume that hitrating also would have a considerable impact and boost on our overall rage generation. I know this basically depends on your other stats, but if anyone, much stronger at warrior theorycraft than me could do the quick napkin math of how much rage/second you'd have with the following stats, I'd be very grateful:

MH weapon 3.6 speed, 233 DPS
OH weapon 3.4 speed, 233 DPS
6500 AP
50% ArP
Improved windfury totem

(I've excluded critrating as to simplify the math).

Well thats from my experience (and I guess I might not be the best example since my latency is usually high) - but BT cooldown is realistically 4 sec + latency+ possible reaction time slip. While WW is 8 sec + latency + possible reaction time slip. So BT are affected by latency and slip twice as much, and since they are a non-zero factor, BT are the element that slows down the rotation. If the shift is 0.25 sec every "round" (my latency and no human error, or typical US latency but some imperfection), then we will go from WW,BT,BT cycle (with BT at 5.5 sec) to BT,WW,BT one (with BT at 0.0/8.0 sec) in around 30 seconds (delaying BT by 1 sec makes them "connect" with WW from "other side").

post removed (Shha replied already)

I dont feel entirely confident with the BT, WW, BT, GCD rotation, which is what people here seem to suggest. Using the sheet listed above I can get out 2 different lists by switching the parameters on how long you wait for a CD. One is the BT, WW, BT, GCD and the other is BT, WW, GCD, BT, GCD, GCD.

The second one seems to give rise to a lot of lost time of cooldowns. My suggestion would be to use a rotation like this: BT, WW, GCD, BT, GCD. Which is basically use slam if it only delays BT or WW, but not if it delays both BT and WW.

Wouldn't this lead to higher dps than the two cycles above?

It would lead to higher DPS than listed above, if you get the slams to fill the GCD's otherwise, you wouldnt be spending the GCD at all I assume and the normal rotation would go. My impression of lag-expanded rotations is that depending on what skill you hit first, the 2nd skill will always follow the first skill, and BT and WW will never overlap more than your total latency

I want to add a question regarding the new BT changes that i didn't see anywhere:

Could it change our rotation when the enemy has 20% or less health? Will it lower the minimum AP range when it's better to spend our rage on BT + WW instead of spamming Execute + WW? If yes, will it make execute totally useless for Fury and open a glyph slot?

I remember seeing somewhere that BT's are better then Execute + Glyph + talent when you have 6500+ or 8500+ AP, can't find where it was written ...

Well one thing wrong about the fury rotation page. Gcd isnt only thing affected by latency. BT cooldown is 4 sec+latency too basically.

Isn't there a client queue of sorts to help against this sort of thing though? Like how stop casting macros used to work because you could send the next ability slightly before the previous one had ended?

I think that was only to force the interface to send the input before it was "unlocked" from spell cast state. Gcd isn't interruptable in any way afaik and your interface doesn't wait for a server acknowledgement at the end of the cooldown. It just tries to send the input as soon as the clientside exits global cooldown.

Unfortunately BT is not only delayed by 4s+latency cause this could happen only if you wait for it to come out of cd.

latency 100ms
0.0 BT
1.6 WW
3.2 Slam
4.8 BT
---------
0.0 BT
1.6 WW
3.2 --
4.1 BT

First case has 3xlatency delay (and 0.5s due to gcd). Second case only has 1x. What matters is the gcds just before your spell. the more used the more delay due to latency.
.