Beast Mastery Bible - Page 124

DevEight · 03/12/08, 8:38 AM

Hey guys.

My guild has recently started MH and we got Kaz'rogal and I was thinking about switching out my old T5 shoulders with the [Beast-tamer's Shoulders].

I also have T5 legs so if I switch out my shoulders I'd lose my 2-piece set bonus.

Would be nice if you guys could give me some feedback on this.

xEw · 03/12/08, 9:46 AM

They explicitely said that lowering the GCD via haste rating will not be the case for hunters.

@DevEight: I also have beasttamers but i dont put them in for t5 because losing t5 will make your pet die more often and therefore the petbonusdmg is for nothing. Maybe i'm overreacting but i intend not to lose the 2piecet5bonus ever and so i'm heading for 4t6 2t5(shoulder+glove, since this means the least statloss) with 2.4.

and since this is a new page i would like to point to the hit-problem i posted last page. I'm really interested in an answer :-)

Nandei · 03/12/08, 9:46 AM

With 2.4 coming up with lots of haste gear and that amazing crossbow, I have been wondering how useful is Rapid Killing for a BM hunter?

I am cookie-cutter BM specced and raiding MH/BT (4 bosses down in each instance). I am the IHM-slave of my guild, so I don't have any points in Efficiency at all. I have been thinking about putting my two points in Rapid Killing into Efficiency instead. I am using 1:1 priority rotation with Arcane when I have mana and just Steady-Auto when I am low. As I have never had Efficiency I don't have a feel on how much of an effect those two points would have.

At the moment I have Sunfury and one haste item, a cloak from ZA timed chest. With Rapid Fire my attack speed with these is already going below gcd. Quite often I am also getting other haste effects on top of Rapid Fire: IAotH procs, shaman uses Bloodlust or someone uses Drums of Battle. With multiple hastes on, I find I cannot hold a rotation, I am missing some specials here and there. I think it is partly because of my playstyle (hand-weaving) and partly because of my rather high latency of 300ms. Usually in a boss fight I am using Rapid Fire three times. Without Rapid Killing that would be twice per fight.

I am just looking for other's opinions on this, is it worth getting some more mana efficiency and loosing one Rapid Fire per boss fight? Especially when already going below gcd with it and that getting only worse in the future.

Ragnar · 03/12/08, 11:31 AM

Originally Posted by DevEight

Hey guys.

My guild has recently started MH and we got Kaz'rogal and I was thinking about switching out my old T5 shoulders with the [Beast-tamer's Shoulders].

I also have T5 legs so if I switch out my shoulders I'd lose my 2-piece set bonus.

Would be nice if you guys could give me some feedback on this.

They're fantastic. The answer is simple: Use them for fights where your pet doesn't take a lot of damage (Rage, Kaz'rogal, prolly Anatheron too). Use the T5 2 piece for where your pet does take a lot of damage (Azgalor). Basically, if you're having a problem keeping your pet up on a fight while using Beast-Tamer's, then swap to the T5.

alienangel · 03/12/08, 11:54 AM

Originally Posted by DevEight

Hey guys.

My guild has recently started MH and we got Kaz'rogal and I was thinking about switching out my old T5 shoulders with the [Beast-tamer's Shoulders].

I also have T5 legs so if I switch out my shoulders I'd lose my 2-piece set bonus.

Would be nice if you guys could give me some feedback on this.

The spreadsheet will tell you the dps difference between the two, so the healing is the only factor to consider I imagine. I never had T5 2pc, and while my pet died a huge number of times the first few times on each encounter (something like 12 revive pets on early azgalor >.>) I found it dying less and less on later attempts. That was probably party from getting gear with more stam, but mostly it's from just learning to keep a mend up if pet is under full, and keeping it away from predictable damage.

So I'd say go with the upgraded stats, and see which fights you can keep the pet up on without too much trouble.

I wish T5 had belt/bracers/boots/ring so it was easier to get 2pc t5 without losing stats so much :S I managed to get a couple pieces recently and the healing output is insane.

Ragnar · 03/12/08, 12:02 PM

Originally Posted by Nandei

With 2.4 coming up with lots of haste gear and that amazing crossbow, I have been wondering how useful is Rapid Killing for a BM hunter?

My gut feeling is that Rapid Fire is still worth it, but it's tough to say since your latency / reaction time comes into play.

2 points in Efficiency is roughly 12 mp5 for an auto/steady cycle, and 14 mp5 for 1:1 priority using arcane.

I've got a faster weapon, so Rapid puts me below the GCD (1.4 with no other haste buffs / procs), but I feel it's still worth it. If you're going to get group buffs such as Heroism or Drums of haste, plan out when you'll get them so you can use Rapid before or after them. Likewise, I always use Rapid when I don't have a haste buff up, and it's even more beneficial when you're low on mana and using Viper.

As far as having it mess up your rotation, just focus on doing your 1:1 thing. You'll push back Auto-shot every time, but that's ok. Just focus on casting your next special after the GCD is up. Then again, I generally have 60-70 latency, so I can usually get the two shots out in ~1.65 sec. Going from 2.1 to 1.65 is a pretty big DPS boost. Even with more latency, I think you'd still be going from ~2.07 to ~1.8, and that's not bad.

Sure, sometimes I get a Hawk + DST proc immidiately after popping Rapid. Other times, I get the full effect with nothing overlapping it. Having both Hawk and DST procs up only happens 14% of the time, while having no haste proc happens about 37% of the time. I feel the haste boost from Rapid is worth it to hopefully speed up some of that 37%, and is worth the loss of ~13 mp5 even if I'm wasting some of the haste. If you don't have a DST, Imp Hawk is probably only up ~44% of the time.

Ragnar · 03/12/08, 12:22 PM

So, I've been playing around with Cheeky's spreadsheet and haste raiting, and I'm surprised to find sweet spots of haste, pockets where the benefit of haste raiting is disproportionately better than outside the pocket.

For example, using my gear (includes Legionkiller, DST, and Imp Hawk), I'm finding the following:

Haste Raiting -- DPS Increase compared to previous level
00 -- 0.00
10 -- 7.12
20 -- 7.12
30 -- 8.45
40 -- 7.08
50 -- 7.07
60 -- 7.07
70 -- 5.92
80 -- 8.58
90 -- 5.58
100-- 5.56

So going from 20 to 30 haste, I get a considerably larger benefit than going from 0 to 10, or 30 to 40. I've been trying to figure out why that is, and I'm just not seeing it. Any help explaining this would be greatly appreciated.

What's also odd is the drop in effectiveness at 70, the increase at 80, and drop back down at 90. I'd say that 70 is where it bumps hard against the GCD limit, but the increase at 80 has me confused.

If I repeat this with the Crossbow of Relentless Strikes, .1 speed faster than Legionkiller, I get:

00 -- 0.00
10 -- 6.50
20 -- 5.78
30 -- 8.76
40 -- 5.76
50 -- 8.11
60 -- 5.74
70 -- 5.72
80 -- 5.71
90 -- 5.69
100-- 2.94

Again a slight drop at 20, increase at 30, drop at 40, increase at 50, then steady decrease until 100. Is this a bug in the spreadsheet, or is there something I'm missing?

Tenkai · 03/12/08, 5:15 PM

I'm also confused as to how Cheeky's handles haste in conjunction with shot rotations. That's why personally, I would say find the perfect paperdoll speed for you and your connection/computer (for me it's about 1.75 on a 1:1 rotation) and then gear haste to achieve that.

grymwish · 03/12/08, 5:27 PM

Originally Posted by Ragnar

So, I've been playing around with Cheeky's spreadsheet and haste raiting, and I'm surprised to find sweet spots of haste, pockets where the benefit of haste raiting is disproportionately better than outside the pocket.

For example, using my gear (includes Legionkiller, DST, and Imp Hawk), I'm finding the following:

Haste Raiting -- DPS Increase compared to previous level
00 -- 0.00
10 -- 7.12
20 -- 7.12
30 -- 8.45
40 -- 7.08
50 -- 7.07
60 -- 7.07
70 -- 5.92
80 -- 8.58
90 -- 5.58
100-- 5.56

So going from 20 to 30 haste, I get a considerably larger benefit than going from 0 to 10, or 30 to 40. I've been trying to figure out why that is, and I'm just not seeing it. Any help explaining this would be greatly appreciated.

What's also odd is the drop in effectiveness at 70, the increase at 80, and drop back down at 90. I'd say that 70 is where it bumps hard against the GCD limit, but the increase at 80 has me confused.

If I repeat this with the Crossbow of Relentless Strikes, .1 speed faster than Legionkiller, I get:

00 -- 0.00
10 -- 6.50
20 -- 5.78
30 -- 8.76
40 -- 5.76
50 -- 8.11
60 -- 5.74
70 -- 5.72
80 -- 5.71
90 -- 5.69
100-- 2.94

Again a slight drop at 20, increase at 30, drop at 40, increase at 50, then steady decrease until 100. Is this a bug in the spreadsheet, or is there something I'm missing?

From the way I understood what Cheeky has previously said many pages back (forgive me if Ive misquoted you Cheeky), he doesnt full understand the FULL in's and out's of how and why 3:2 rotations work. None of us do, all we know for sure is that it works for some set ups and that it is a mana hog.

So knowing that, I wouldnt trust the spreadsheet completely when it comes to random outlyers like those times you listed the dps dropping. It would be too tough to say if its a problem in the spreadsheet or if its a true occurance (at least until the time we fully understood the 3:2 rotation......or Blizz changes it again).

Ivaldi · 03/12/08, 5:42 PM

I bet you it's not linear between those numbers either. There may be spikes or plateaus you're not seeing. Shadow priests have a similar issue with something called wait. Can't really tell you what it's about but at a certain point too much haste is a bad thing.

alienangel · 03/12/08, 5:58 PM

Benefit from haste being a non-continuous curve doesn't seem that strange to me, since our damage comes in discrete chunks (attacks), and the different types of attacks we do are affected differently by haste - each level of passive haste changes the proportion of each type of attack from our total, hence some values of haste are better than others. This would be true even if we perfectly understood how "3:2" macros work, since by their nature they trigger attacks based on what can be fired.

Kamma · 03/12/08, 10:55 PM

EDIT: Modified so that "busy" time does not include lag time (auto shots can sneak in during lag and the rotation continues without a clip).

Nosti, thank you for your post. I've been trying to get a firm understanding of the steady spam mechanic, and you post made me (I think, maybe, hopefully) see the big picture.

I'm sticking my neck out on this one, particularly since I'm a noob around these parts.

I did some math and developed a algorithm for determining the theoretical auto:steady ratio for given weapon speed, total effective haste, and lag. This is really just a reiteration (I think) of your post but with some funny symbols, so I take no credit.

I'll first present the algorithm, then some pretty pictures showing how things change with different variables, and then finally I will attempt to justify the algorithm.

Algorithm

given variables: $W_s, L, \lambda$ (respectively weapon speed, lag/reaction time,and effective haste)
for $m=1$

Find $n\in\mathbb{N}$ such that

$(1.5+L)n-\frac{1.5}{1+\lambda}<\frac{mW_s}{1+\lambda}\le (1.5+L)(n+1)-\frac{1.5}{1+\lambda}$
If

$(1.5+L)n-\frac{1.5}{1+\lambda}<\frac{mW_s}{1+\lambda}<(1.5+L)n$

then return (# of auto shots, # of steady shots)=(m,n)
else $m=m+1$ and go to 1.

EDIT: Note that in step one above (for all examples I can think of at least):

$n=\left\lfloor\frac{1.5+mW_s}{(1+\lambda)(1.5+L)}\right\rfloor$

Moreover, the rotation is done over a period of $n$ steady shots each taking a time of $1.5+L$ . So, the shots per second is $\frac{m+n}{(1.5+L)n}$ .

Pretty Pictures
Note: Made by evaluating 100 points for each graph. The resolution could probably be better, but since reality has randomness thrown in, these should get the idea across sufficiently.

Here's a graphical result of haste vs shots per second for $W_s=3, \lambda=.15 ... 0.9, L=0, .1, .2$ (order red, blue, green)

Here's a graphical result of haste vs shots per second for $W_s=3, 2.8, 2.6, \lambda=.15 ... 0.9, L=0$ (order red, blue, green)

Keep in mind, I don't know if this is right - it's merely the mathematical result if the assumptions I made represent reality.

Justification

First, ask the question: what marks the behinning and end of a cycle in the steady spam rotation? The answer is, auto shot clipping. Whenever an auto shot clips, due to the way that the .5 sec cast of auto shot occurs during the last .5 seconds of the steady shot cast which clipped the auto shot, the auto shot and steady shot line up and cast simultaneously. Since the entire rotation is essentially an inductive process, if any "condition" occurs more than once, there will be a repeating cycle. Thus, auto shot clipping marks the beginning and ending of the spam cycle.

Ok, so consider the make shift diagram below,
where [ to ] mark the gcd, S marks the beginning of a steady shot cast, --- marks casting, ) marks the end of a steady cast, and L marks a lag/reaction time of length L

L[S----)   ]L[S----)   ]L[S----)   ]L[S----)   ]L[S----)   ]L[S----)   ]

Now suppose we have an auto shot which wants to cast during the first steady shot, marked by an X. Then it fires with the steady shot at ), marked by x.

L[S----)   ]L[S----)   ]L[S----)   ]L[S----)   ]L[S----)   ]L[S----)   ]
    X--x

The next auto shot will occur $\frac{W_s}{1+\lambda}$ seconds follow x. This of course can either come up during a blank space where nothing is being cast and the GCD is active, or it can become available during a busy time. So, all the potential space falls into two categories: (a) where nothing is occurring and the auto shot fires, and (b) where the auto cannot fire and will clip. If it clips, we have the end of the cycle, and if it doesn't then we look at the next possible auto shot.

Now, to simplify things define:

$R=1.5+L$
$S=\frac{1.5}{1+\lambda}+L$

Using these we can label all the critical points in our diagram:

 LL[S--------)   ]LL[S--------)   ]LL[S--------)...
^  ^         ^   ^  ^         ^   ^
0  L         S   R  R+L     R+S  2R... etc.

Which stretched out to the nth steady shot we have:

  LL[S--------)      ]LL[S--------)      ]LL[S--------)...
^             ^      ^  ^         ^      ^
(n-1)R  (n-1)R+S    nR nR+L    nR+S (n+1)R... etc.

So, suppose that some auto shot following a clipped auto shot and before the subsequent clipped auto shot (marking the end of the cycle) occurs somewhere during the nth steady shot, its GCD, or lag that occurs before the next steady shot.

IE, if $t_a$ is the time that auto shot tries to fire, we have the inequality:

$nR+L<t_a\le (n+1)R+L$

However, since auto shot becomes available some multiple of $\frac{W_s}{1+\lambda}$ seconds following the clipped auto shot, and if we count time 0 as before the steady cast (and its preceding lag), then we have:

$t_a=\frac{mW_s}{1+\lambda}+S$ where $m\in\mathbb{N}$ .

Substitution into the previous inequality, some algebra, and substituting for $R$ and $S$ gives:

$(1.5+L)n-\frac{1.5}{1+\lambda}<\frac{mW_s}{1+\lambda}\le (1.5+L)(n+1)-\frac{1.5}{1+\lambda}$

This is the inequality found in step 1 of the algorithm. So, what step 1 does is (for a given m) finds which steady shot period the mth auto shot will attempt to fire during.

Then step 2 determines whether not the auto shot will try to fire during a "busy" time. If so, we've found the end of the cycle and m is the number of auto shots, and n is the number of steady shots. If not, add one more to m and go back through from step 1.

The inequality in step 2 comes from the following inequality representing the "busy" time (derived from the diagram above):

$nR+L<t_a<nR+S$

$nR+L-S<\frac{mW_s}{1+\lambda}<nR$

$(1.5+L)n-\frac{1.5}{1+\lambda}<\frac{mW_s}{1+\lambda}<(1.5+L)n$

Well, there's my neck. Please tell me if and how I messed up. I have not checked it to actual data, although since I modified the "busy" time to not include the lag time, it coincides with Nosti's numbers. Of course, this is all merely a brief theoretical construction. Although I may have gotten some specifics wrong, I think the general approach could potentially provide a way to model this chaotic rotation.

And of course this would be modeling without the wonder that is kill command causing random lags.

I'm going to try to come up with a nice interpretation of what happens if the effective haste changes at some time, so we could potentially model the rotation including haste procs.

Zurgat · 03/13/08, 4:26 AM

Originally Posted by Phanuel

Buffed up with ~1950 AP, Lightning Breath is hitting for around 180-190, so it's not terrible. But it's no 2.0 310+

Lightning breath : 99 to 113 Damage : 50 Energy
Plus 12.87% Dmg extra based on your RAP.

So, with 1950 Ranged Attack Power, your pet would gain 250,965 spell damage.
However, as it's an instant spell the added spell damage is only 30-33% correct?
75,2895 (30%) - 82,81845 (33%)

Wowwiki says it's Spell damage and healing - WoWWiki - Your guide to the World of Warcraft 42.86% however. Which would mean it could add 107,563599, which seems a bit high.
But some spells are 30%, and it's not listed on the hunter wiki whether lightning breath or scorpid poison follow any different rules.

From your values it looks like it's ~33%
Can anyone verify this percentage?

Also, scorpid poison is Scorpid Poison - WoWWiki - Your guide to the World of Warcraft
55 damage over 8 seconds at the highest rank.
55 / 8 sec * 5 stacks = 34,375 damage per second. Not including the spell damage bonus.
Doesn't seem to compare too well to lightning breath. Unless the scorpid poison works on a different coefficient than LB?

And another quick question, physical pet attacks such as gore, bite and claw. Are these considered yellow attacks or white attacks?
Meaning glancing blows would not occur.

Lactose · 03/13/08, 6:19 AM

Originally Posted by Zurgat

From your values it looks like it's ~33%
Can anyone verify this percentage?

I'm pretty certain it's not that high. I've got a few test results (not yet finished), so I should be able to get a definitive answer in not too long (barring work).
Things to consider: BM Talents (+20% damage), Loyalty (+25% damage), Wind Serpent bonus (+7% damage), race (Orc: +5% damage).
There's also some strangeness regarding the tooltip, which I might have tracked down what's the cause of, but not whether is a display bug only.

And another quick question, physical pet attacks such as gore, bite and claw. Are these considered yellow attacks or white attacks?
Meaning glancing blows would not occur.

Pet special attacks are yellow. They do not glance.

TranquilRage · 03/13/08, 6:52 AM

Firstly My apologies to Nosti for missing the GCD part of his explanation.

In a 3:2 rotation its my understanding that the Auto Shot will be delayed until the second Steady fires (Assuming no KC) effectively increasing your Auto Shot speed to GCD + Delay Triggering Macro + Steady Shot Speed. As shown here.

This diagram assumes a high speed of macro execution and thus no delay between GCD end and Steady cast.

So if we use the following macro, We could trigger KC (if available) on every other Steady Shot. So we would alternate between an Auto Shot or a KC after every Steady Shot. Thus not clipping Auto any more. Further, if Auto got delayed by an additional Steady cycle, it would effectively resync itself.

#showtooltip
/console Sound_EnableSFX 0
/cast !Auto shot
/castsequence [target=pettarget, exists] Steady Shot, Kill Command, Steady Shot; Steady Shot
/script UIErrorsFrame:Clear()
/console Sound_EnableSFX 1

Please note I can't test this as I am at work. If I am missing something please let me know.

Zurgat · 03/13/08, 7:29 AM

Originally Posted by Lactose

I'm pretty certain it's not that high. I've got a few test results (not yet finished), so I should be able to get a definitive answer in not too long (barring work).
Things to consider: BM Talents (+20% damage), Loyalty (+25% damage), Wind Serpent bonus (+7% damage), race (Orc: +5% damage).
There's also some strangeness regarding the tooltip, which I might have tracked down what's the cause of, but not whether is a display bug only.

Pet special attacks are yellow. They do not glance.

Alright, it's going to be tricky then.
But, as i'd like to find out more exact values, i was wondering if anyone had done more detailed research on the spell damage coefficient yet?
Surely it's been theory crafted at some point in time? Post 2.2 at any rate.
If not, i guess we got ourselves a nice new project!

Dodge, block and parry can still happen then, but pets automatically move to the back of a mob, so dodge and armor should be only factors after that.

Lactose · 03/13/08, 8:10 AM

Originally Posted by Zurgat

Alright, it's going to be tricky then.
But, as i'd like to find out more exact values, i was wondering if anyone had done more detailed research on the spell damage coefficient yet?
Surely it's been theory crafted at some point in time? Post 2.2 at any rate.
If not, i guess we got ourselves a nice new project!

I'm in the middle of it right now. Hopefully I won't have to work too much this weekend, and will have time to finish the tests, and be back with some answers & results.

Zurgat · 03/13/08, 8:57 AM

Originally Posted by Lactose

I'm in the middle of it right now. Hopefully I won't have to work too much this weekend, and will have time to finish the tests, and be back with some answers & results.

Looking forward to it.

It could also be worth investigating whether caster pets : Petopia: "Caster" Pets
Take any difference in the spell coefficient compared to normal pets. There aren't a lot of caster pets, but as most pets are used for melee purposes, their melee DPS will be lower, while their spells "might" get a boost.
I think it's doubtful however, and suspect that caster pets are simply a flaw in the pet system which blizzard hasn't fixed yet. But one can never be sure until it's tested.

Originally Posted by TranquilRage

Please note I can't test this as I am at work. If I am missing something please let me know.

It is my understanding that autoshot is unable to fire until 0.5 seconds after steady shot finishes. So you may want to look into that.
Your chart seems to assume it can fire immediately after.

Nosti · 03/13/08, 10:08 AM

Originally Posted by Kamma

Pretty Pictures
Note: Made by evaluating 100 points for each graph. The resolution could probably be better, but since reality has randomness thrown in, these should get the idea across sufficiently.

Here's a graphical result of haste vs shots per second for $W_s=3, \lambda=.15 ... 0.9, L=0, .1, .2$ (order red, blue, green)

Here's a graphical result of haste vs shots per second for $W_s=3, 2.8, 2.6, \lambda=.15 ... 0.9, L=0$ (order red, blue, green)

Keep in mind, I don't know if this is right - it's merely the mathematical result if the assumptions I made represent reality.

Hehe, you seem to be a very mathematically educated person, capable of latex

(I studied Physics myself).

Your results seem to be correct and are indeed just my results formulated in a more general and theoretically correct way. Nice work!

Some thing I was wondering, how do you determine the /lambda? Effective haste? What is the formula for it?

Some results I wanted to emphasize:

1) Lag can be benefitial! (LOL)

2) The amount shots/time clearly increase with haste in a stepwise way 3:2->4:3->5:4->6:5->....
You could indicate these x:y numbers on the graph for even more understanding (1.111 <-> 3:2 | 1.166 <-> 4:3). Eventuelly they will converge to a shot rotation where you fire 1 Steady and 1 Auto every GCD, that would require alot of haste rating, but with some Rapid Fire + procs it can be achieved. You might even be able to push it further so that there's sometimes 2 Auto's between Steadies, but those are really extreme cases.

3) I'm not sure how the hidden 0.5 Auto Shot casting time is taken into all of this. If this casting time is true I don't believe a 3:2 rotation could be obtained in some cases where I have been able to obtain it, OR I must been underestimating the lag/reaction time duration (which gives a little more leway to obtain 3:2).

Would 0.5s delay for autoshot after steady result in something like?

0.00 Auto Shot
0.00 Steady Shot starts casting
1.04 Steady Shot fires
1.50 GCD ends
1.60 Steady Shot starts casting
2.64 Steady Shot fires
3.10 GCD ends
3.14 Auto Shot fires (next Auto Shot will fire at 5.22s)
3.20 Steady Shot starts casting
4.24 Steady Shot fires
4.70 GCD ends
4.80 Steady Shot starts casting
5.84 Steady Shot fires
6.30 GCD ends
6.34 Auto Shot fires

Which is a 2:1 rotation.

Praxx · 03/13/08, 10:40 AM

Originally Posted by TranquilRage

#showtooltip
/console Sound_EnableSFX 0
/cast !Auto shot
/castsequence [target=pettarget, exists] Steady Shot, Kill Command, Steady Shot; Steady Shot
/script UIErrorsFrame:Clear()
/console Sound_EnableSFX 1

Please note I can't test this as I am at work. If I am missing something please let me know.

I believe your macro would force three steady shots between each auto because of the use of the /castsequence. Further, I believe it would hang in the /castsequence until kill command could be fired. Note: I am at work so I cannot test this in game to be certain.

Originally Posted by Zurgat

It is my understanding that autoshot is unable to fire until 0.5 seconds after steady shot finishes. So you may want to look into that.
Your chart seems to assume it can fire immediately after.

Actually not entirely true. Part of the reason the 3:2 macro is working at all is because the auto shot timer seems to start before the steady shot is fired. I have seen delays of less than a tenth of a second between a steady shot and an auto shot.

Here are three examples from one of my combat logs:

16:18:32.921 Steady Shot hits
16:18:32.984 Auto Shot crits

16:01:16.515 Steady Shot hits
16:01:16.609 Auto Shot hits

16:14:55.531 Steady Shot hits
16:14:55.640 Auto Shot crits

Each of these occurred on the first auto shot of a 3:2 cycle. This is the point where one auto shot has been delayed and the timer catches back up to the rotation.

TranquilRage · 03/13/08, 10:48 AM

Originally Posted by Zurgat

It is my understanding that autoshot is unable to fire until 0.5 seconds after steady shot finishes. So you may want to look into that.
Your chart seems to assume it can fire immediately after.

Thats a very good point.

So the following needs to be true

GCD - SteadySpeed + Delay > 0.5 seconds. This could be why faster spamming macros clip more shots. Because the Delay is reduced. You need an additional haste of >9% after SS and a Quiver to bring Steady to a 1 second cast. So the more haste you have the faster you can have the macro spammed to reduce the delay needed to not clip shots.

Ragnar · 03/13/08, 11:35 AM

Originally Posted by grymwish

From the way I understood what Cheeky has previously said many pages back (forgive me if Ive misquoted you Cheeky), he doesnt full understand the FULL in's and out's of how and why 3:2 rotations work. None of us do, all we know for sure is that it works for some set ups and that it is a mana hog.

So knowing that, I wouldnt trust the spreadsheet completely when it comes to random outlyers like those times you listed the dps dropping. It would be too tough to say if its a problem in the spreadsheet or if its a true occurance (at least until the time we fully understood the 3:2 rotation......or Blizz changes it again).

Oh, sorry, I forgot to add. This is 1:1 rotation auto-steady rotation. Nothing to do with the bizzaro 3:2 crap.

Sarutobi · 03/13/08, 11:55 AM

Originally Posted by Praxx

I believe your macro would force three steady shots between each auto because of the use of the /castsequence. Further, I believe it would hang in the /castsequence until kill command could be fired. Note: I am at work so I cannot test this in game to be certain.

It would also, due to the conditional statement in the castsequence, if his pet were to ever die, do nothing but fire auto shots.

KraxisSingular · 03/13/08, 11:58 AM

Originally Posted by Zurgat

It could also be worth investigating whether caster pets : Petopia: "Caster" Pets
Take any difference in the spell coefficient compared to normal pets. There aren't a lot of caster pets, but as most pets are used for melee purposes, their melee DPS will be lower, while their spells "might" get a boost.
I think it's doubtful however, and suspect that caster pets are simply a flaw in the pet system which blizzard hasn't fixed yet. But one can never be sure until it's tested.

It is very much worth the effort. Recent testing on caster pets has revealed some rather considerable, and interesting, differences between them and 'normal' pets. Now these change don't seem to have much positive effect outside Wind Serpents. But I must direct you to this thread regarding 'caster' pets. From page 2 it begins to get interesting, and it is overall a fairly joined effort to find out the effects.

Basically what we have noticed (many thanks to Skätha for the taming of a high level caster) is that caster pets scale much better with agi in relation to crit (25 agi per 1% crit compared to 33 agi to 1% crit). They also crit 5% more with LB inherently, and their Int can be buffed for even more LB crit (something normal pets are completely incapable of). Inherently they also appear to have more hit, which is really odd.

So far it appears that caster pets can very much be worth the effort in large raids with full buffs. Especially if some Shammy is so kind to put down Grace of Air totem. But they suffer from a destinct lack of HP so pet unfriendly fights aren't their strength at all.

But really, what I would like is for some assistance in testing this out more. Currently there is not much data to go on, but it is highly encouraging. And again, many props to Skätha for daring to bring a caster pet to a raid.

Zurgat · 03/13/08, 12:09 PM

Originally Posted by KraxisSingular

So far it appears that caster pets can very much be worth the effort in large raids with full buffs. Especially if some Shammy is so kind to put down Grace of Air totem. But they suffer from a destinct lack of HP so pet unfriendly fights aren't their strength at all.

Very interesting, i wonder if this would this apply to monkeys, bats, serpents and scorpids as well?
Dots can't crit, but it might increase the spell damage and hit chance. (thus uptime)

Definitely worth researching.
Also a shame that there's no lightning bolt above level 60 yet. A serious oversight from blizzard, or intended...