Archived

This topic is now archived and is closed to further replies.

Lactose

Hunter damage cycles breakdown (Nerubian Slavemaker update)

289 posts in this topic

Okay, but just realize that the corrections are only a few percent, and probably dwarfed by other errors in practice. For the sake of good theory, however:

For an 11s cycle (say, an 8/3=2.66s Auto)

<snip>

et cetera. If you start by firing Multi right before an Aimed (rather than after, as I've shown above), you'll get off a Multi every 10s for a little while, but eventually, you'll run into your Aimed Shot. You won't be able to sustain a long-term average of more than 1 Multi every 11 seconds.

First, just looking at the speed, etc, the 9 second cycle would actually be better :P

2.66 Auto Shot fires

5.32 Auto Shot fires

6.00 Start Aimed Shot

9.00 Aimed Shot fires

9.00 Auto Shot fires

Took Rhok'delar stats, changed weapon speed to 3.06 --> 2.66 after quiver

So, commenting on the 9 sec cycle first.

Let's try adding Multi-Shot here. Testing several options, so bear with me :P

Removing Auto Shots for less clutter.

0.00 Multi-Shot fires

6.00 Start Aimed Shot

9.00 Aimed Shot fires

10.00 Multi-Shot fires

...

51.00 Start Aimed Shot

54.00 Aimed Shot fires

60.00 Dilemma!

Looking at our cycle, a closer look at the last part would be:

54.00 Aimed Shot fires

54.00 Auto Shot fires

56.66 Auto Shot fires

59.32 Auto Shot fires

60.00 Multi-Shot ready

60.00 Aimed Shot ready

61.98 Auto Shot ready

3 options, delay Multi-Shot until after Aimed Shot. Delay Aimed Shot until after Multi-Shot (this turns it into the 10 sec cycle, so I won't write it down again). Delay Aimed Shot until after Multi-Shot + Auto Shot.

60.00 Start Aimed Shot

63.00 Aimed Shot fires

63.00 Auto Shot fires

63.00 Multi-Shot fires

65.66 Auto Shot fires

66.00 Start Aimed Shot

Multi-Shot delayed by 3 seconds. Same will happen every 7 cycles,

So instead of 0.9*Multi-Shot, it should be... 0.86*Multi-Shot? 6/7 * Multi-Shot would seem correct based on the above...

60.00 Multi-Shot fires

61.98 Auto Shot fires

61.98 Start Aimed Shot

64.98 Aimed Shot fires

64.98 Auto Shot fires

67.64 Auto Shot fires

70.00 Multi-Shot

70.30 Auto Shot fires

70.98 Start Aimed Shot

Aimed Shot delayed by 1.98 seconds. Conflicting cooldowns ever 2 cycles or so.

Seems 0.86*Multi-Shot is the best bet at the moment, based on this weapon speed at least. Which seems to agree with what you're saying, no? Well, so much for the 9 second cycle, let's check with a cycle exceeding 10 seconds.

Testing, the longest (optimal) cycle was with the speed of Rhok'delar, making the cycle 10.57 seconds. So, using that.

Actually, scratch that. It just dawned on me now. Basically, if the cycle is longer than 10 seconds, you only get 1 Multi-Shot per cycle due to Aimed Shot casting.

For a cycle shorter than 10 seconds... Hmm...

In the example above, the cycle is 1 second faster than Multi-Shot cooldown. Thus it takes 6 fully run cycles for Multi-Shot to catch up with Aimed Shot? If it takes x cycles, the coefficient for Multi-Shot would be x/(x+1) -- as it is now, 6/7 fits that at least :P

Would that mean that you could do something similar to... [messy math]

IF CycleLength > 10; 1*MultiShot_PerCycle

IF CycleLength < 10; ((6/(10-CycleLength)) / (1+6/(10-CycleLength)))*MultiShot_PerCycle

[/messy math]

Question is, if it's right, should I round it up or down? :P

Can't think of anything else right now... although I think I'll be able to reformat the comparisons a lot. I'll save it for after the formulae are commented on though ;)

Share this post


Link to post
Share on other sites

I'm currently 0/21/30 I'm wondering what is more effective now or how it would affect your calcs as all the AP, Auto, Aimed and Multi has different values

Share this post


Link to post
Share on other sites

This reworked shot calculation is done manually, so it takes some time using other variables. It's basically done so I can test the automatic version of it.

Once I'm sure it's working - and it seems to be up until now, with pretty varying weapon speeds - I'll only be a bit unsure of some hit%. After that, I'm all set. I've been working on a spreadsheet for comparing gear and builds in terms of damage (mana not considered due to lots of reasons). Various versions of it have been seen on the EU hunter forums, and I'll post a link to it here if there's any interest.

It won't give the timestamp version for cycles, but will show DPS with crits included, and factor in every related talent, as well as choosing between which cycle you should use (wait for Auto Shot or not).

Basically, when it's done, you'll be able to check yourself ;) Up until that point, I really can't do each calculation with different builds, gear, etc; it would take too much of my time, and I would probably fall asleep while doing it.

Share this post


Link to post
Share on other sites

Lactose, it's ironic this was post with such close timing to a post I put up at the official boards...after someone reading my post, they linked me to your's. I encourage you to take a look -- http://forums.worldofwarcraft.com/thread.aspx?ForumName=wow-hunter&ThreadID=1183364

Basically looks at a lot of the same things you do, from the perspective of "which weapon is best, assuming any shot cycle?" I don't explain the math or logic that goes into it in as much detail, but I have an entire spreadsheet I could send you if you want to mail me at f1lman@yahoo.com

Share this post


Link to post
Share on other sites

Yup, I've seen the thread ;)

Actually, the list (similar to the image you link to in your first post) was something that made my make my spreadsheet (which is basically what all this is for) in the first place; I was certain Ashjre'thul would be much better than other weapons (like for example Huhuran's Stinger).

Which it has been since I got my thread stickied I think... around march 14th ;)

Although, I made it way too good then, due to lots of things, amongst others how to calculate the best cycle times for weapons not fitting a perfect 9.0 second cycle (Huhuran's Stinger specifically). Now, I don't list ranged weapons. I let you compare 1 set of build / gear with a new one (so basically, you can only compare 1 weapon at a time), allowing you to min/max your gear. For example, what gives the most increase out of a +1% crit ring and a +28 RAP one, stuff like that.

Anyway, cycles not acting correctly lead to a lot of testing and thinking, which lead to some rude formulae. They in turn lead to more testing, etc, until I posted this here. Feeling pretty good about it now, and as soon as someone can check my cycle formula* and comment on it, I'll update and upload my newest version. Will link it here and stuff, if other people want it.

Trust me though, I lurk a lot on the US Hunter forums ;)

I don't think I need the spreadsheet just yet (takes a long time to look through thoroughly, which I don't have time for right now), but I'm curious... are you using ammo as normalized as well as RAP? I seem to remember tests showing both were affected, contrary to what you might think when reading patch notes...

*[messy math]

IF CycleLength > 10; 1*MultiShot_PerCycle

IF CycleLength < 10; ((6/(10-CycleLength)) / (1+6/(10-CycleLength)))*MultiShot_PerCycle

[/messy math]

Share this post


Link to post
Share on other sites

Updated every comparison with conflicting timers taken into consideration. Also corrected a couple of minor errors.

Share this post


Link to post
Share on other sites

Hello again. Hope you don't mind me necroposting a bit, but new weapon, new breakdown :P

For reference, using exact same stats as with the previous calculations, just changing the stats you get from ranged weapon (so, going from Rhok'Delar to Nerubian Slavemaker, calculations are made with +7 RAP).

Stats when using Nerubian Slavemaker:

1400 RAP

+7 scope

5/5 Ranged Weapon Specialization

3/3 Barrage

3/3 Slaying vs target

15% quiver

Jagged Arrows

NERUBIAN SLAVEMAKER

15% quiver. Speed = 2.61

Set AutoShot to 579

Set AimedShot to 1201

Set MultiShot to 810

Average damage against a target, before armor is applied.

Aimed + Multi (9 sec default)

A					B

2.61 Auto Shot fires			2.61 Auto Shot fires

5.22 Auto Shot fires			5.22 Auto Shot fires

6.00 Start Aimed Shot			7.83 Auto Shot fires

9.0 Aimed Shot fires			7.83 Start Aimed Shot

9.0 Auto Shot fires			10.83 Aimed Shot fires

					10.83 Auto Shot Shot fires

+ Multi-Shot				+ Multi-Shot

CycleDPS A				CycleDPS B

= (3*579 + 1201 + 0.86*810) / 9.0	= (4*579 + 1201 + 810) / 10.83

= 3634.6 / 9.0				= 4327 / 10.83

= 403.84 (optimal cycle)		= 399.54
Multi + Aimed cycle (10 sec default)
A					B

2.61 Auto Shot fires			2.61 Auto Shot fires

5.22 Auto Shot fires			5.22 Auto Shot fires

5.50 Multi-Shot fires			6.33 Multi-Shot fires

7.0 Start Aimed Shot			7.83 Auto Shot fires

10.0 Aimed Shot fires			7.83 Start Aimed Shot

10.0 Auto Shot fires			10.83 Aimed Shot fires

					10.83 Auto Shot fires

CycleDPS A				CycleDPS B

= (3*579 + 1201 + 810) / 10		= (4*579 + 1201 + 810) / 10.83

= 3748 / 10.0				= 4327 / 11.70

= 374.80				= 399.54 (optimal cycle)

Optimal cycle: Aimed + Multi (9 second default) A - 403.84

Compared to Ashjre'thul's optimal cycle:

Optimal cycle: Aimed + Multi (9 second default) A - 412.54

Nerubian Slavemaker has 1% more crit than Ashjre'thul. This does not compensate for the lower DPS; Ashjre'thul still comes out on top in a PvE setting.

Before you mention anything regarding cycles being impossible etc:

They're not. Of course, timing them perfectly (especially the ones with optimal cycle lengths of for example 9.55) is practically impossible. However, getting as close to it as possible will be the most damaging cycle. Thus, the theoretical calculations I give you will show you the most damaging cycle.

This is even more true for the "pure" cycles (9 sec Aimed Shot + Multi, or 10 sec Multi + Aimed). You don't have to timer, you don't need to pay attention to your Auto Shots, just spam the button, the cooldowns make sure you can't fire too fast.

If Nerubian Slavemaker had been 3.3 or 3.4 speed, I'd be happier.

For those interested, optimal cycle for 3.3 = 429.11; 3.4 = 433.44

Share this post


Link to post
Share on other sites

So even though the slavemaker has higher weapon damage and higher dps than the Smiting xbow, because of the loss of autoshots due to aimed shot still casting in each cycle (you lose .39 seconds each aimed shot) it still ends up less? Or is it just a trick of the math that a weapon with a higher damage range and higher dps post-normalization still is inferior dps? /headscratch

Share this post


Link to post
Share on other sites

Edit: Nevermind, I'm dumb.

Autoshot needs to be normalized too, clearly.

So even though the slavemaker has higher weapon damage and higher dps than the Smiting xbow, because of the loss of autoshots due to aimed shot still casting in each cycle (you lose .39 seconds each aimed shot) it still ends up less? Or is it just a trick of the math that a weapon with a higher damage range and higher dps post-normalization still is inferior dps?

Just look at the numbers above the math -- that's all that matters. SO much of hunters' damage comes from their RAP that a difference of 0.3 or 0.4 in speed will lower autoshot damage enough that it will outweigh any gains from base weapon damage.

Share this post


Link to post
Share on other sites

Dang, I hadn't done the calculations myself, but had hoped the extra damage would outweigh the speed difference. And I allready got noticably more ap then that even with the same specc, so once you got the naxx gear, as well as the naxx xbow, it should be very much higher, which will favour the bwl xbow even more since it'll add equally to the dps, except nerubian loses some due to speed, and bwl one loses nothing. (oh well, at least we should get a new best ranged wep at some point in expansion, patch 1.6 -> expansion isn't too long to wait I guess. Can always hope they change the way Aimed/auto interact, to make hunter weapon about more then speed).

edit: And while as mentioned, a perfect cycle is rather hard to attain, the more ap you get, the more significant the dpsloss for a non-optimal speed weapon like this would be, making even a flawed cycle, running around at times, etc, not change the outcome.

Share this post


Link to post
Share on other sites

Sophomoric question but: at what point does a weapon swap (in your damage cycles) begin to make sense?

EDIT: for clarification

Share this post


Link to post
Share on other sites

A separate question, if I understand correctly -- is it possible that, because of normalization, you could get better cycle DPS if the weapon were faster?

The problem is that if you can only fit in 2 autoshots in 6 seconds, you want them to hit as hard as possible, so you want a really slow weapon. But what if a weapon were 2.3 speed, such that you could fit in a third autoshot? Yes, you'd obviously lose some base weapon damage on your specials, but would the third autoshot make up for it?

Since I'm lazy, what would the numbers be if Slavemaker were 2.3 speed instead of 3.0, with the same DPS?

Share this post


Link to post
Share on other sites

KalelScilla -- What Praetorian said

Moridin -- What do you mean by noticably? Talking pure self-buffed RAP, you can't possibly that much higher than me, you don't have tier 3 yet :P

Good point regarding flawed cycles btw.

probiscus -- Not sure what you're asking..?

Praetorian -- Will check and post in a sec, have to run to the store to get some groceries first :P

Share this post


Link to post
Share on other sites

can you do the math with 1900 attack power? that's much closer to realistic raid buffed AP with naxx gear - i already have over 1700 already :-P

Share this post


Link to post
Share on other sites

Not to discount your math, as I know absolutely nothing about hunter mechanics, but according to your work the Xbow of Smiting is the best, but several other cycle DPS breakdowns I've seen claim that Huhuran's Stinger is the best out of current available weapons (though only slightly higher than the Xbow). Is it because they are allowing auto-shots to finish as opposed to adhering to the strict 9 or 10s cycles?

Share this post


Link to post
Share on other sites

Elendril -- Will do, but will only post optimal cycle DPS, can't be bothered doing timestamps etc :P

chalon -- The error is that they've used a 10 sec rotation as standard, based on the Multi-Shot cooldown. This favors faster weapons, able to get an extra Auto Shot in their cycle compared to slow weapons like Ashjre'thul. What I do is I take the 9 sec default cycle, then compare it to a 9 sec cycle with an added Auto Shot (obviously making the cycle longer than 9 sec), then the same with the 10 sec cycles.

Edit: Like your screenshot, repeatable cycle for Ashjre'thul is set to 10 seconds, while it performs better under a 9 second cycle. Read my first post for more info =)

Share this post


Link to post
Share on other sites

math hurts my head.

if you don't mind, i'm going to forward this thread along to some people i know who know people who are in positions to look at something like this in a meaningful way :-P

Share this post


Link to post
Share on other sites
Just look at the numbers above the math -- that's all that matters. SO much of hunters' damage comes from their RAP that a difference of 0.3 or 0.4 in speed will lower autoshot damage enough that it will outweigh any gains from base weapon damage.

Ultimately though autoshot dps is weapon-speed irrelevant unless you measure over a very short period of time, unless I'm completely retarded. With the way we're doing the cycle math, it depends very heavily on the speed of the weapon whether you get 3 or 4 autoshots in the time allotted - it's a very short window.

However the slavemaker should have more autoshot damage than the smiting xbow over time. And since the weapon damage is higher, the multi-shots and aimed shots should be hitting harder too. I'm still scratching my head over this.

Edit: math that doesn't matter to my point is removed

The trick being, if you do in fact lose .31 seconds per 10 seconds on auto-shots due to the faster speed waiting for aimed shot to cast, vs. losing .04 seconds per 10 seconds with smiting bow, that means 3.1% of your autoshot damage is gone from the slavemaker vs. only .4% with the smiting xbow. And I think thats the real loss.

How does blizzard fix this? Adding more 3.4 speed weapons? Or change the aimed shot mechanic so that the slow weapons aren't universally better until 20dps improvements happen?

Share this post


Link to post
Share on other sites

The strict theory (and shot cycles) analysis of ranged weapons always makes a bad assumption - an unlimited source of mana (probably not a bad assumption for alliance :( ).

Once out of mana (if not potting or FD-drinking) then the Slavemaker would obviously do more significantly more damage whenever the hunter is just autoshooting. And for each raid setup (gear/spec/buffs) what is the breaking point between the loss of DPS from FD-drinking to the gain of having a full mana bar to return to full DPS cycle... as straight autoshot DPS increases FD-drinking becomes less likely to provide a positive DPS return over simply continuing to attack.

Share this post


Link to post
Share on other sites
The trick being, if you do in fact lose .31 seconds per 10 seconds on auto-shots due to the faster speed waiting for aimed shot to cast, vs. losing .04 seconds per 10 seconds with smiting bow, that means 3.1% of your autoshot damage is gone from the slavemaker vs. only .4% with the smiting xbow. And I think thats the real loss.

How does blizzard fix this? Adding more 3.4 speed weapons? Or change the aimed shot mechanic so that the slow weapons aren't universally better until 20dps improvements happen?

That's another way of looking at it, yes. The problem is, indeed, that going from 3.4 to 3.0 is that, in theory, your individual autoshots hit a bit less hard (due to a lower AP multiplier), but you get more of them so you end up doing the same DPS. But when you're fitting them in between specials, which you almost always are, you aren't actually getting any more autoshots.

One random idea completely off the top of my head would be to add a new hunter passive skill (might be a good example of a mid-60s skill for the expansion) called Quick Shot, that, when you do a special, simultaneously fires off a weaker autoshot, which is regular autoshot multiplied by [(time elapsed since last autoshot) / (autoshot cooldown)], above some minimum floor like 0.2.

So if your autoshot time is 2.5 sec modified, and you shoot at 2.5sec, then again at 5.0 sec, then when you do Aimed/Multi at 6.0 sec, you'd also immediately fire off a Quick Shot that does (1.0/2.5) = 40% of your normal autoshot damage. This way you could recapture that lost damage and autoshot would be weapon-speed-independent as always intended.

Share this post


Link to post
Share on other sites
But if multi-shot has a 10s cooldown, how's a 9-sec cycle possible?

it's a 9 second average cycle, with the multishot damage per second calculated in, rather than a 10 second cycle including each skill.

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.