Warrior Level 70 PvE DPS F.A.QFollow

Indeed that is true, very good point. Socketing hit over crit gems will indeed give more rage gain, be it a very small amount but still more none the less.

The respective dmg they add though is also a factor to consider.

8/15.8 = ~.5% Lets say your white dmg makes up 60% of your total dmg.

.60 * .005 = .3% increase in total dps.

8/22.8 = ~.35% crit.

Since crit increases all of your dps, not just white, you get a increase of .35% in total dps.

So given your white/yellow dmg distribution is 60/40 (fairly common for DW warriors) +8 hit rating gems will increase total dps by ~.3% while +8 crit rating gems will increase total dps by .35% (the more your dps is made up of white dmg the more +hit affects your total dps)

You have to factor in the extra rage from the +hit gems to get an exact percentage since more rage = more dps done. But that calculation is better left to someone who has a better attention span than me lol.

Once you factor in that crit helps keep flurry up and yellow crits are affected by impale it would seem (can't be absolutely sure without the math) that crit gems still comes out on top for total dmg done.

Edited for spelling and clarification of gems.

Edited, Sep 6th 2007 4:30pm by Jimpadan

Ok so now i'll apply impale math to your yellow crit damage (this will help bring in line the AP breakpoint for 8 Crit gems vs 8 Strength gems).

.35% ... 40/60 yellow/white.

40% * .35% = .14% increase to yellow dps

With 30% crit (assumed) * 20% increased damage = 6% yellow crit dps increase.

0.14% + 0.21% = total .35% increase.

6% * 0.14% = 0.0084% increase again.

total .3584% increase with impale and 30% crit rate, per 8 crit rating gem after.

So .003584 * X = 16 (ap)

X = 4464

A total of 107 AP difference with 30% crit rate and 2/2 Impale.

Still seems that 8 Strength gems out perform 8 Crit rating gems.

You have to convert 1.14 dps into a percentage of your total before you can compare the .35% increase of the crit gem. If your going to compare in dps terms you have to consider how much dps your doing before the .35% increase. Lets say 1000 (nice round number and not extremely hard to achieve)

1000 * .35% = 3.5 dps increase.

8 str increases dps by 1.14

8 str is always 1.14 dps increase, plus the very small amount it adds to BT dmg.

Since the 8 crit rating is a straight percentage increase it scales, while 8 str does not.

The higher the dps you put out the more favorable the crit rating gems are.

Breakeven point being:
X * .35% = 1.14
X = 325.71 dps

I just woke up and I'm nursing a hangover so I may have made a mistake but I've checked it a few times and can't find any.

So your agreeing that crit gems are better than str gems past 1500 AP?

I'm sorry I'm still kinda out of it. I definitely can't hang like I used to.

man i need to go to college all this math is killing me. i think when i get closer to 70 (at 32 atm) i'll just make a thread "tell the moron what he is missing" and then i'll just link gear. looks like nice work from what i get of it i'll have to take time to really read it at home instead of at work on breaks.

Aren't you forgetting that +str also scales with your current crit rate?

Say you put out 500 DPS without any crits. If you then have a 25% crit chance, your DPS will be 625. Now, in terms of gems, 1 str = 1 crit rating = 2 AP. You're comparing +8 str vs. +8 crit rating? 16 AP vs. 0.35% crit. Now, if I may calculate this in another way:

Forget about the 500 DPS example above (or anything calculating just your DPS). We'll look at attacks only.

First, which attack scales best with AP? Easy one: BT. Say you've got 2k AP selfbuffed, without Imp. Berserker Stance (since that would make +str scale). Your BT's then hit for 900. If you are to calculate crit chance: [900 * 1.25 = 1 125] (We'll be assuming at least +8.5% hit, so special abilities won't get hit reduction).

So, 1 125 is our starting point. Lets add 16 AP to that: [2 016 * 0.45 * 1.25 = 1 134]. 9 more damage.

Lets do the crit instead: [2 000 * 0.45 * 1.2535 = 1 128.15]. A little over 3 more damage.

Now, what scales badly with AP, as a DW warrior? WW. You've been using merciless weapons so far, so I'm going to do the same. A WW will hit for 597, adding crit to that will make it 746.

Add 16 AP: [((2 016 * 2.4 / 14) + 254) * 1.25 = 749.5]. 3.5 more damage.

Use +0.35% crit instead: [((2 000 * 2.4 / 14) + 254) * 1.2535 = 748.16]. Slightly more than 2 damage.

Let's do it with white attacks. Now, forgive my noobishness, but I was under the impression that the offhand gets full AP bonus, but only does 50% of its tooltip DPS. I'm gonna calculate it your way though, since you're probably much more experienced DW warriors than I.

MH (Merciless Gladiator's Slicer):

[((2 000 * 2.6 / 14) + 254) * 1.25 = 782]
+8 str: [((2 016 * 2.6 / 14) + 254) * 1.25 = 785.5]. 3.5 more damage.
+8 crit rating: [((2 000 * 2.6 / 14) + 254) * 1.2535 = 784]. Slightly less than 2 more damage.

OH (Merciless Gladiator's Quickblade):

[((2 000 * 1.5 / 14 * 0.625) + 146.5) * 1.25 = 350.53571428571 (~350.5)]
+8 str: [((2 016 * 1.5 / 14 * 0.625) + 146.5) * 1.25 = 351.875 (~352)]. ~1.5 more damage.
+8 crit rating: [((2 000 * 1.5 / 14 * 0.625) + 146.5) * 1.2535 = 351,51721428571 (~351.5)]. 1 more damage.

Just realized I forgot hit penalty. Aaw @#%^ it!



Now, if every one of these abilities benefit more from +AP than +crit, I don't see how crit could up your DPS more than AP. If you then consider that +str scales with BoK, and Imp. Berser. Stance, where as +crit doesn't... You get the point.

Are we supposed to find a break point in this? I must confess I'm not adept as such things. First, lets see if the we get the same increase, or if the AP mysteriously scales with the rest of your AP:


Lets assume twise as much AP:

[4 000 * 0.45 * 1.25 = 2 250]
[4 016 * 0.45 * 1.25 = 2 259]. 9 More damage (and so it's true, the increase is static in this case).
[4 000 * 0.45 * 1.2535 = 2 256.3]. 6.3 more damage.

So, the break point for BT should be [X * 0.0035 = 9] X = 2 571,42857142857

[2 571 / 1.25 / 0.45 = ~4 572]

Yeah... 4.6k AP should do it. However, that's only true with this particular crit rate. Lets see if it's correct in the other cases:

First, the scale control.


[((4 000 * 2.4 / 14) + 254) * 1.25 = 1 174,64285714286 (~1 175)
[((4 016 * 2.4 / 14) + 254) * 1.25 = 1 178,07142857143 (~1 178). ~3.5 more damage (again: it fits).
[((4 000 * 2.4 / 14) + 254) * 1.2535 = 1 177,93185714286 (~1 178) ~3 damage increase (3.289)

Lets try with the BT break point:

[((4 600 * 2.4 / 14) + 254) * 1.25 = 1 303,21428571429 (~1 303)
[((4 616 * 2.4 / 14) + 254) * 1.25 = 1 306,64285714286 (~1 307). ~3.5 more damage (3.43).
[((4 600 * 2.4 / 14) + 254) * 1.2535 = 1 306,86328571429 (~1 307). 3.649 more damage.

Yup. Adds up, if you've got 25% crit, you need 4.6k AP for crit to be more effective than crit AP. (Edit)



Now I've gotten myself curious... lets make a quick check to see if this is somewhat accurate with other crit rates. Say 33%, since that's said to be optimal for Flurry.



BT: (Edit: More precise numbers)

[2 000 * 0.45 * 1.33 = 1 197]
[2 016 * 0.45 * 1.33 = ~1 207]. 10 (9.576) more damage.
[2 000 * 0.45 * 1.3335 = ~1 200]. 3 more damage.

[4 000 * 0.45 * 1.33 = 2 394]
[4 016 * 0.45 * 1.33 = ~2 403]. 9 more damage (slight difference, but nevermind) (9.576).
[4 000 * 0.45 * 1.3335 = ~2 400]. 6 more damage.

[4 600 * 0.45 * 1.33 = ~2 753] (2 753.1)
[4 616 * 0.45 * 1.33 = ~2 763]. 10 (9.576) more damage.
[4 600 * 0.45 * 1.3335 = ~2 760]. 7 more damage.

Not quite, but I can't be ***** to calculate a new breakpoint... or can I?

[X * 0.0035 = 9.576] X = 2 736.

[2 736 / 1.33 / 0.45 = ~4 571]. So, the new break point is still ~4.6k? Can't be. We'll try with 4.8k:

[4 800 * 0.45 * 1.33 = ~2 873]
[4 816 * 0.45 * 1.33 = ~2 882]. ~9 more damage.
[4 800 * 0.45 * 1.3335 = ~2 880]. ~7 more damage.

No? Close enough... maybe it's at 5k. If so, you can theorize about the break point being +50 AP per 1% crit. I can't be ***** doing more maths now though.

Edit: Did some more maths, turns out you have to have a bit over 6k AP at 33% crit chance for crit to be more effective than AP. (The AP case was still 0.126 damage higher than the crit case at exactly 6k AP and 33% crit).

Edited, Sep 9th 2007 3:20pm by Xordon

...

Well, since you seem to think it needs simplification:

devious said that 25% crit will be sufficient for flurry use (although I've always heard 33, or at least 30%). So, basicly:

If your crit chance is lower than 25%, go for crit.

If your crit chance is 25% and your AP is lower than 4.6k (fully raid buffed, with shaman, pally, self and pot buffs) go for crit.

If your crit chance is 33% and your AP is lower than 6k (again, with any buffs imaginable) go for crit.


Now, I've only heard about RP getting the odd 4k AP (was is 4.2k?) so getting 4.6k would indeed be an acievement. Don't beat yourself up about it though, since all you have to do to compensate is go for +str gems instead of +crit gems.


Heh... I tend to get a little carried away :P



Oh, and I just realized I'm quoted at (at least) 2 places in the OP. Cheers devious ;)

Ah sorry, I'll edit that to make more sence:



You've kind of highlighted the wrong parts though... If I only listed when to go for crit, you only need a simple sence of elimination to see when to go for AP. If/Then/Else scenarios, very simple.

Actually it doesn't. "real" dps is affected by debuffs, buffs, miss rate, etc...

1.14 dps from 16 ap, w/ imp berserker stance or BoK is more dps.

Edited, Sep 9th 2007 1:37pm by devioususer

Changed and updated a few things.

::EDIT::
Damnit I can't spell...

Edited, Sep 9th 2007 11:26am by devioususer

I think I gave up on the;



part.

Lets just go ape ****.

I specced d/w tonight for the fun of letting another tank tank. No, it does not get full AP.

With 2500 ap and a 91.1 dps offhand I got something like 210 DPS on tooltip (hard to tell sometimes when your paying more attention to the bottle and threat meter than exact numbers).

The hit rating requirements are off. I know it's based off the WoWWiki numbers, but if you go in game with 127/128 hit rating you'll be at the cap as 2H/1H + Shield, not 138. WoWWiki truncates the % needed for 1% hit slightly, resulting in incorrect numbers as you get more and more.



It gets the full AP bonus, but the tooltip reflects the DW penalty. If you had a 100 DPS weapon and 200 DPS worth of AP (2800), you'd see the offhand at 150 DPS. With max DW Spec, you'd see it at 187.5 DPS.



Are you forgetting to factor in Battle Shout or other buffs? The number is too high for that AP number.



"Perma Flurry" is a bit of a misnomer. The better phrase would be Flurry Uptime, and that's not terribly easy to calculate for theorycrafting and is far more suited to simulated testings (much like Shaman Windfury). Too bad nobody's done a Warrior Simulator. =/

Using just autoattacks for a moment, and since I'm nursing my caffeine now I may ***** this up royally...

25% Crit Rate
Going from the first autoattack crit, which can be at any point, and then calculating the probability/percent that Flurry will be active over the next 100 attacks.

.75 * .75 * .75 = 42.2% chance that Flurry will not refresh itself before it expires (75% chance of a combat event that is != a crit).

Over 100 attacks, one would expect 25 of them to be crits. Assuming perfect distribution, this would have a Flurry uptime of 75% (Crit Hit Hit Hit Crit Hit Hit Hit - Flurry expires between the last hit and the next Crit). The chance of that occurring is very, very, very small.

Come to think of it, I have absolutely no idea how to calculate this properly. I'll leave the above up as a testament to _why_ we need to either adapt or write our own simulator, but a whole bunch of factors come into play. Even assuming that you have only autoattacks and a 50% crit rate, the chance of Flurry expiring without refreshing itself is 12.5%.

The mitigating factors are Bloodthirst/Whirlwind/Hamstring (additional attacks which do not consume charges but do increase the number of chances to refresh Flurry), Windfury Totem complicates matters (contrary to popular belief, Windfury _does_ consume a Flurry charge if it procs, but it also provides the additional chance to crit; the benefits far outweigh the cost, but it does slightly counteract the addition of instant attacks when calculating uptime) and when you add in instant attacks you have to make assumptions about the speed of the weapon versus the cooldowns, which means that the dream of theorycrafting any of this with a formula becomes farther and farther away. I don't know about you guys, but if I see a dy/dx in theorycraft my head probably will explode.