Forum Settings
       
Reply To Thread

Az MathFollow

#77 Aug 25 2006 at 2:55 AM Rating: Decent
*
217 posts
Woo! You finally stickied Az Math! Would like to ask though, which are the correct figures? The one in this post or the other Item Comparison List you posted? You seem to have quite varying weights in the two posts. Thanks :)
#78 Aug 25 2006 at 8:20 AM Rating: Decent
I had a DPM (Damage per mana) meter running on the whole raid for the last 5 raids I've been on, and the top 5 positions were always hunters (no surprises), excluding rogues and warriors ofc. While this may seem pretty obvious, all the Rhok'delar using hunters were on about 7 points of damage per point of mana, while the lone hunter with Xbow of smiting was pumping out 10 or 11 points of damage per point of mana, an enourmous increase from Rhok'delar, especially considering the very small difference in their actual weapon dps.

This is much more than the overall dps increase shown on a standard cycle, being a 50% increase in mana efficiency wheras it is only a 10-15% (iirc) increase in total damage. This is probably down to the extra damage dealt by Xbow auto-shot hits, as this wasnt normalised to 2.8 in the nerf patch - it remained at the weapon speed damage.

Clearly using a fast weapon has even less use than previously thought, one hunter was using heartstriker, and only getting 5 or 6 DPM.

Yet more proof that hunter itemisation sucks.
#79 Aug 25 2006 at 9:18 AM Rating: Good
****
6,678 posts
I would argue that, zysac. When a mage casts a Fireball, 100% of that damage is from the spell. When a hunter fires an Aimed Shot, only the bonus damage is from the ability and the rest is from what is essentially an autoshot. Rank 1 multi costs 100 mana, for 0 damage bonus. However, if I have a normal autoshot of about 600 (averaged with crits,) and I have Barrage and full Giantstalker, that multi is gonna land for 800 -- which is 8 DPM. It's mathematically impossible for a mage to match that, but one could argue that it's only 2 DPM coming from the fact that it's a Multi Shot. Making arguments about hunters and their DPM is fairly meaningless unless it's to compare different hunter skills. (This is also why many hunters don't use Serpent Sting on raids.)

Quote:
This is much more than the overall dps increase shown on a standard cycle, being a 50% increase in mana efficiency wheras it is only a 10-15% (iirc) increase in total damage. This is probably down to the extra damage dealt by Xbow auto-shot hits, as this wasnt normalised to 2.8 in the nerf patch - it remained at the weapon speed damage.

Clearly using a fast weapon has even less use than previously thought, one hunter was using heartstriker, and only getting 5 or 6 DPM.


DPM should be slightly higher on slow weapons because of the higher damage range, but your Multi and Aimed should be similar. Less than 100 difference, tops. Granted, 200 difference in 10 seconds is still 20 DPS, but that's what that last post about if slow weapons are actually dead was about. 100 DPM at a cost of 100 mana is 1 DPM, however. If you're getting radically different values beyond that, it's not an issue of the weapon -- did you ask the Heartstriker hunter what their RAP was, or if they were using max ranks, or if they were dilluting their DPM with serpent sting?

And no, autoshot is not normalized, but autoshot is also time-based. In the same time window of autoshooting, a Huhuran's Stinger will fire about 5 shots compared to Ashjre'thul's 4. AXoS's autoshots will be about 25% harder, but they will also be 25% more infrequent. With the exception of how to interrupt autoshooting to use Aimed, the mechanics of autoshot boil down to hard, cold basic DPS.
____________________________
Only the exceptions can be exceptional.
#80 Aug 25 2006 at 3:15 PM Rating: Decent
Quote:
DPM should be slightly higher on slow weapons because of the higher damage range, but your Multi and Aimed should be similar. Less than 100 difference, tops. Granted, 200 difference in 10 seconds is still 20 DPS, but that's what that last post about if slow weapons are actually dead was about. 100 DPM at a cost of 100 mana is 1 DPM, however. If you're getting radically different values beyond that, it's not an issue of the weapon -- did you ask the Heartstriker hunter what their RAP was, or if they were using max ranks, or if they were dilluting their DPM with serpent sting?


Ok, a few questions to answer here: all the hunters in my guild have very even gear, all of us come very close to one another damage wise (the asjrethul hunter has been pulling away a little recently), and I know for sure all of them use serpent sting and max ranks of everything (one thing I can be thankful for is that paladins in my guild arent lazy with JoW). I dont think I conveyed the maths of my post very well, nor explained my point properly.

The main fact is that with a slow weapon, you can do alot more damage for the same amount of mana. Unless you're using serpent sting every second, its pretty tough to use more or less amounts of mana when all of you are using the same cycle. Even if you were using different cycles; if anything the hunter with xbow of smiting should be using a faster, more mana consuming cycle anyway. Even if the hunter with the xbow was slacking, and relying on auto shot more than a proper cycle to attain damage, the fact remains that he could beat myself and other hunters while we were using a cycle.

Anyway, I was only using using DPM meters as a small experiment to ponder what DPM was actually like, not a way of finding hard proof for something or another.

Quote:
With the exception of how to interrupt autoshooting to use Aimed, the mechanics of autoshot boil down to hard, cold basic DPS.


Surely this isnt true; with crits included, it was weapon speed which made slow weapons so superior prior to normalisation, why should it be any different when the moves which had been normalised (aimed and multi shot) are removed from the dps equation?
#81 Aug 26 2006 at 11:38 PM Rating: Good
****
6,678 posts
Quote:
The main fact is that with a slow weapon, you can do alot more damage for the same amount of mana.


Printed weapon damage + AP/14 * 2.8 + ammo * delay.
A slower weapon may have up to 100 higher on printed damage, and maybe another 25 damage on the ammo. There shouldn't be any wider margin that this, beyond multipliers like RWS, Barrage, and crits. If you're seeing a margin of over 200 on a non-crit, Blizzard's normalization isn't working.

Quote:
Anyway, I was only using using DPM meters as a small experiment to ponder what DPM was actually like, not a way of finding hard proof for something or another.


I understand. Mana efficiency is not something I've done a great deal of personal research on. I really should look into what a point of MP5 is worth, but it's very hard depending on what people do in their cycle. A poster earlier on this page analyzed it moderately well, but I'm all for hearing comments about DPM because I've studied DPS, not DPM.

Quote:
Surely this isnt true; with crits included, it was weapon speed which made slow weapons so superior prior to normalisation, why should it be any different when the moves which had been normalised (aimed and multi shot) are removed from the dps equation?


The shot cycles are purely based around the way the use of Aimed Shot interrupts your autoshooting.

Let's say I'm out soloing, and firing my Ashjre'thul, which takes 2.96 seconds per autoshot (with quiver.) And let's say I do nothing but autoshot. Every 2.96 seconds, I'm going to land a shot that is worth 2.96 seconds worth of autoshot DPS. However, if the pet kills the target when I'm about to fire my next autoshot, I don't get 2.8 seconds of my DPS in damage. The shot simply never fires.

If I were firing a steady laser beam of damage, I would actually get that 2.8 seconds. It seems like I *should* get it. If my DPS is 100, and I was autoshooting for 20 seconds, I should do 2000 damage, darnit! Instead, I only do 1776. (100 damage/sec * 2.96 sec/shot * 6 shots)

That all should make sense to you. Now, imagine you're shooting at a target where you can shoot at it for 6 or 7 seconds, and then it despawns and pops up again 3 seconds later. (Kinda like the purple ghosts in West DM, but more annoying.) This is what firing the cycle is like. So obviously the best situations are going to be ones where I can stretch out my DPS across the entire period I have to fire. I want to be shooting from the instant I can to the last moment I could possibly get a shot off.


The part that hasn't already been made clear is that the autoshot timer starts after your first shot, which happens pretty much right as you push the button. So if I have a 6 second window, and I'm firing in almost exactly 3 seconds, I can potentially get 3 shots in. When you're just autoshooting for a long long time, that isn't very noticeable, but when you're only firing for 6 seconds and then pausing, you ought to notice that 3 shots * 2.96 seconds means I should be getting almost 9 seconds worth of DPS...in 6 seconds!

Aimed Shot interrupts your actual shots, but not the cooldown on them. So when you start an Aimed Shot, you are postponing your auto until it finishes; however, Aimed Shot takes long enough to fire that with a 15% quiver, no weapon in the game will not finish the cooldown while you're aiming. So you are basically repeatedly getting 7-9 seconds worth of shot time every 6 seconds...because it's actually 9 seconds elapsed counting the AS, and you're losing whatever time your weapon is ready to go during the charge-up.

And this is why Ashjre'thul is so powerful. It's a 3.4 delay, which works out to almost exactly 3 seconds with a quiver. So you fire at t=0 seconds, t=3 seconds, and t=6 seconds, and then it's just ready to fire again as your Aimed Shot finishes. It fits not only the firing window but the charge length of Aimed perfectly.


Unfortunately, this is all in a perfect world. If you have an Ashjre'thul, you know that it doesn't work out quite so cleanly. Although it's listed as a 9 full cycle, in practice you end up using a 9 clip or a 10 full, because the first auto shot does not occur instantly, Multi Shot will sometimes partially interrupt your shooting, and internet lag and your inability as a human being to push the button at exactly the right moment keeps this from working out quite the way it should in theory.

My later renditions of the weapons spreadsheet have incorporated some arbitrary amount of delay to try to better imitate the "real" world rather than pure math theory. Ironically, this is why Huhuran's Stinger has taken a dive from my initial version to present. It fits a 10 second cycle perfectly, without the lag and delay and all that junk.


Overly long explanation to basically come back and say: Yes, it's all because of Aimed Shot. Multi doesn't even really factor in, just Aimed.
____________________________
Only the exceptions can be exceptional.
#82 Aug 27 2006 at 2:15 AM Rating: Decent
I see ^^

Thanks.
#83 Aug 27 2006 at 9:13 AM Rating: Good
****
6,678 posts
Does that make sense, or was that more of an "umm...yeah. Ok, sure..."?
____________________________
Only the exceptions can be exceptional.
#84 Aug 28 2006 at 4:38 PM Rating: Decent
It was 90% 'yeah I understand' and 10% 'wtf is he talking about' :D
#85 Aug 28 2006 at 8:50 PM Rating: Good
****
6,678 posts
If you don't understand, other people reading don't either, so anything I could clarify?
____________________________
Only the exceptions can be exceptional.
#86 Sep 08 2006 at 10:50 AM Rating: Decent
So, I'm pretty confused by one thing.

A lot of this math is based upon the stated assertion that 14 AP = 1 DPS. In fact, it is the fundamental basis for a lot of the calculations.

But, when I add or remove gear that has a straight AP or RAP bonus from my character, my tooltip ranged DPS increases or decreases by MORE than this formula.

For example, if I remove my Drake Fang Talisman (+56 AP), I'd expect my Tooltip DPS to decrease by EXACTLY 4. Instead, it actually decreases by 4.6.

Or if I remove either my Master Dragonslayer's Ring (+48 AP) or my Royal Seal (+48 RAP), I'd expect my Tooltip DPS to decline by 3.4. Instead it actually decreases by 3.9.

Based on this impact it seems that it takes less RAP than 14 to get 1 additional DPS. In fact, based on these numbers 12.2 AP = 1 DPS.

But can this be right?! I think I must be making a mistake somehow here.

Can anyone explain this to me?

Thanks.

(BTW, I realize that if you take crit into account, it takes less than 14 AP to get an additional EFFECTIVE 1 DPS. But the problem I'm having is that my TOOLTIP DPS is showing this effect, and the TOOLTIP DPS is before the crit calcuation.)
#87 Sep 08 2006 at 7:46 PM Rating: Good
****
6,678 posts
Modifiers of some kind? Ranged Weapon Spec would make 56 AP worth 4.2, although I realize that's still not enough...does your RAP actually change by 56 when you remove the trinket?
____________________________
Only the exceptions can be exceptional.
#88 Sep 09 2006 at 12:27 AM Rating: Decent
Quote:
Modifiers of some kind? Ranged Weapon Spec would make 56 AP worth 4.2, although I realize that's still not enough...does your RAP actually change by 56 when you remove the trinket?


Yes, as you would expect, my RAP changes by exactly 56 when I remove the DFT. And by exactly 48 when I remove Master Dragonslayer's or the Royal Seal. But the tooltip DPS declines by more than 4 and 3.4 respectively.

I'm Survival specced, so I don't have Ranged Weapon Specialization.

Also, this issue does not exist on my melee damage tooltip. When I remove the DFT, my Melee Power goes down by exactly 56 and my melee tooltip DPS goes down by exactly 4. This issue only seems to affect Ranged DPS.

I asked one of my Hunter friends to try it also and he noticed the exact same thing.

Can you please try this yourself and see what happens? When you remove a trinket or ring like this, does your tooltip ranged DPS goes down by exactly the item's AP divided by 14?

Thanks
#89 Sep 09 2006 at 9:50 AM Rating: Good
****
6,678 posts
I figured it out. Quiver haste. My Royal Seal went up by 3.9 instead of 3.4 as would be expected. However, 3.4 * 1.15 = 3.9.
____________________________
Only the exceptions can be exceptional.
#90 Sep 09 2006 at 11:46 AM Rating: Decent
Yep. I noticed that it was a 15% difference too. And I tried equipping a 10% speed quiver just to confirm. And low and behold then the expected difference was 10% instead of the 15%.

OK, so what does that mean for our formulas though?

In comparing two items, we've assumed adding gear with 14 additional AP yields 1 additional DPS (before crit).

Now, we see that it only takes 12.17 (i.e., 14/1.15) AP to yield 1 additional DPS (before crit).

Doesn't this mean we have to adjust our comparison formula accordingly? And doesn't it mean that +AP items and +AGI are a little more valuable relative to +crit items than we thought they were?

(I'm using a spreadsheet to calculate relative values of different items, and have been using 14 AP = 1 DPS as a basis.) Using 12.17 makes the calculation materially different.

My math skills are decent, but not spectacular. So let me know if there is an error in my thinking.

Thanks.

#91 Sep 09 2006 at 11:56 AM Rating: Decent
To clarify the preceding post a bit, I've been using the following formula from the post above to calculate the relative impact of new gear.

Quote:
Considering new gear? Only care about damage output? While this analysis is probably pretty meaningless if you aren't a level 60 PvE raider, here's the direct conversions of numbers...

Find your tooltip DPS (T) and your crit rate as a percent (R). Take out all the "*1.3" factors if you don't have Mortal Shots.

1 DPS =...
14 / (1+R*1.3/100) Attack Power
100 / (T * 1.3) Crit %
1 / ( (1+R*1.3/100)/7 + 1.3*T/5300 ) Agi

And divide the agi line by 1.15 if you have Lightning Reflexes.


Doesn't the first line of the formula need to be adjusted? It captures crit rate and Mortal Shots, but it doesn't capturing quiver speed bonus, which we now see has an effect on tooltip DPS.

And, as I think about it, I realize it also doesn't capture Ranged Weapon Specialization -- (which I never remember, being Survival specced).

So, I think instead of: 1 DPS = 14 / (1+R*1.3/100) Attack Power

The formula should be: 1 DPS = 14/(1+(Q + RWS)) / (1+R*1.3/100) Attack Power

Where Q = Quiver Speed Bonus (likely 15%), and RWS = Ranged Weapon Spec (likely 5%, if you have it at all).

Let me know if I'm on to something, or if I've just gone completely down a blind alley here. Thanks.

Edited, Sep 9th 2006 at 5:26pm EDT by GioHunt
#92 Sep 11 2006 at 1:21 AM Rating: Good
****
6,678 posts
All 3 equations would be scaled by an equivalent amount. If you want to divide them all by 1/1.15, be my guest, but it wouldn't change the proportion.


edit: all this math and I can't keep multiply from divide straight. *holds up hands to see which one makes the L*

Edited, Sep 11th 2006 at 2:21am EDT by Azuarc
____________________________
Only the exceptions can be exceptional.
#93 Sep 11 2006 at 6:14 AM Rating: Decent
Quote:
All 3 equations would be scaled by an equivalent amount. If you want to divide them all by 1/1.15, be my guest, but it wouldn't change the proportion.


I think this might not be right. The first equation would be adjusted, but the second equation would not.

The first equation does not take into account Quiver Speed (or Ranged Weapon Specialization). And, so, adding in modifiers for these would make the translation of AP to DPS more accurate.

The second equation, by contrast, already takes Quiver Speed and RWS into account because these factors are already reflected in the tooltip DPS. So, there is no need to adjust this equation. I think it is accurate as is.

If this is the case, it means the proportion DOES change. (And it means that AGI and AP are slightly more valuable relative to +crit than they were before.)

Now, I'm willing to believe there might be some flaw in my reasoning, but I can't see it yet. Please feel free to critique.

Also, I don't know about the third equation, the math in that one is beyond me! :-)

Edited, Sep 11th 2006 at 7:22am EDT by GioHunt
#94 Sep 11 2006 at 8:24 AM Rating: Good
****
6,678 posts
Hmm, you might have something there. Fortunately, the later item comparison lists are all done from scratch in a spreadsheet, so they don't face that discrepancy, but I may have to go back and edit when I'm not at work and can actually concentrate. I'll comment if I find some reason why they're actually right, but chances are I will just be editing the equations.

And the third one isn't really beyond you. Look at it carefully. It's line one x2 plus line two / 53. (1 agi = 2 RAP, and 53 agi = 1% crit.)

Edited, Sep 11th 2006 at 9:32am EDT by Azuarc
____________________________
Only the exceptions can be exceptional.
#95 Sep 11 2006 at 1:31 PM Rating: Decent
Quote:
And the third one isn't really beyond you. Look at it carefully. It's line one x2 plus line two / 53. (1 agi = 2 RAP, and 53 agi = 1% crit.)


Ah, yes, now I see it.

What confused me were the fixed numbers 7 and the 5300.

But I should have realized that 7 = 14/2. Duh!

And I've been using 52.91% for my crit calcuations, so I had no idea where the 5300 came from. But now I see it. And it's easy to adjust this formula to take into account Quiver and RWS. Then the math works like a charm.

Thanks!
#96 Sep 11 2006 at 10:50 PM Rating: Good
****
6,678 posts
Dang, I just realized I think this actually does affect the spreadsheet. Doesn't reorder much, but it does throw the numbers off a little bit.

Weights should now be...
AP = 1
crit = 21.49958983
agi = 2.405652638
agi(LR) = 2.766500534
hit = 2.07

Ick. Well, I'm not redoing the list. They don't change enough to worry about it.
____________________________
Only the exceptions can be exceptional.
#97 Sep 15 2006 at 9:11 AM Rating: Decent
Terrorfiend
*****
12,905 posts
Coulda used you in guild chat last night, Az.

I was trying to explain how great Bloodseeker is and was told I didnt know what I was talking about, and that Aimed Shot was useless in raids.

:-\
#98 Sep 15 2006 at 3:43 PM Rating: Decent
***
3,043 posts
KTurner wrote:
Coulda used you in guild chat last night, Az.

I was trying to explain how great Bloodseeker is and was told I didnt know what I was talking about, and that Aimed Shot was useless in raids.

:-\


Should've just linked them here. Tell them "zam.com, hunter board, Az Math sticky"

Keep it really simple ;)
#99 Sep 15 2006 at 10:08 PM Rating: Good
****
6,678 posts
Or you coulda just said "Aimed shot fills dead time with extra damage, and does a better job of it than Arcane." That is, more or less, what happens. Unless you cost yourself more than 2.8 seconds of autoshot time, you're coming out ahead without counting the bonus damage, which makes up for the Arcane Shot.
____________________________
Only the exceptions can be exceptional.
#100 Sep 17 2006 at 7:44 PM Rating: Decent
hi i have a question related to the shooting cycle

I;ve been reading a lot of Forums about hunter stuff and i cant say i'm an expert (i have a twinked 22) but talking about the Shooting cycle i seem to notice that there hasn't been much talk about the fact that tyour enemy Comes to you .When it reaches the dead zone you inevitably have to use melee (wing clip usually and run to concussion shot him) but you where talking about like 30 seconds of shooting ..wel to be honest in reallyti you dont really have that much time (unless ur instancing) but when you are about to do an attack this is what i do.

I put hunters mark, send pet , , put a trap (untill all that the pet has reached the Enemy ) then i start with :
1 Aimed shot, concussion , Sting, then arcane or multi ....mostly after that The enemy ahs reached me so i have to melee him so i dont rly see the use of calculation of like 30 sec shots u you dont have that time!
After u wing clip him ..u run ...do a 180 degrees jump shot (concussion ) then u use arcane or whateva ...and so on ...anyways sorry for bothering you guyz and i apologize if i seem like a nub but i'm here to learn and i'm waiting for critics
#101 Sep 18 2006 at 12:30 AM Rating: Good
***
1,441 posts
The "clean" shot cycle as discussed here applies mainly to grouping / raids where a tank is there to hold better aggro than your pet does. Also note that for a long time, most of your PvE fights will be over before starting a second cycle.

In my own experience, what I'd adapt from your fighting method is the use of concussive shot: don't keep it in your cycle, use it if you grab aggro and the mob starts heading to you.
As your pet levels and gains higher versions of aggro-generating abilities, you'll soon be able to open with aimed, immediately followed by multi, then wind up another aimed when the cooldown has reset - provided your target's remaining health warrants it.

Last but not least, if you bother to trap at all (depends on the situation), you will want to heel your pet while your enemy is frozen, wind up another aimed and re-sic your pet on it. As soon as the aimed shot lands, pop another concussive shot, your pet should have regained aggro at this point.
The hardest bit is to heel your pet before it breaks the trap, though.
Reply To Thread

Colors Smileys Quote OriginalQuote Checked Help

 

Recent Visitors: 1 All times are in CST
Anonymous Guests (1)