Graul

I doubt Loderunner was being sarcastic when he said it gives a headache to whoever is updating it.

 

Tiburon11

I wouldn't mind doing this. It would be a good way to "refresh" my warrior knowledge. Unless someone else has started it, I'm going to give it a try.

Reclaimed is currently looking for 2 Mages, 2 Warlocks, and a Holy Priest who are Elitist Jerks just like the rest of our guild.

 

Colan

I'm an idiot, ignore this post. Preferably somebody delete it...

 

Sinnermighty

I have a question about Warrior DPS.

I heard that 17/44/0 is the best Fury spec to go with and I'm wondering if someone can link me a Wowhead build or something around that. I'm also wondering how I should work my rotation to best output my DPS as DW Fury 17/44/0.

Thank you,
Sinnermighty

 

nolena

Actually there are several topics around this, also a nice topic on the same page with an excel which you can alter to see what gear/build changes provide to you in terms of dps.

 

Paa

Have been using this build for quite some time now and find it to be a very good build for fury dps.

Talent Calculator - World of Warcraft

Some talents were taken for tanking/Support, Something you often do as a dps warrior and no other talents for example in the arms tree are worthwhile taking for fury dps.

 

p o j

I applaud this. I would have done so myself, but the math and statistics from college days aren't as keen as they would need to be. I will contribute as best I can, as I'm sure more knowledgeable and experienced than I.

Topics I've always wanted to see proved/ disproved/ gone into detail:
- More hit doesn't mean more chances to crit
- x% crit is needed to keep flurry active for y% of combat (Try as I might to come up with a formula for this, I fail. Seeking ex-professor's help)
- What is 'crit cap' and how does it apply?
- All things considered, how much AP is equal to 1% crit (with regards to dps gain assuming law of averages)?
- Specific interactions with WF and the enhancement shaman (our most potent "buff")

Hopefully, should such a thread get started, the typical 'Here is my armory- tell me what I need to change' posts would be discouraged.

You can't stop the love train, baby.

 

p o j

So I've given up on trying to use ODE's, but instead made 2 different (and very simple) calculations concerning flurry.

z_3, z_2 = chance to not crit (%)
x = crit chance (%)
y = hit chance (assuming 100 - x = y) (%)

Autoattack (3 before flurry expires after activation):
x_3 ~> y^3 = z_3

Instants (assuming 2 hits inside the 3 auto-attack window- 2 chances to reapply flurry):
x_2 ~> y^2 = z_2

Now, how to relate the following senarios:
x_3 series crits, x_2 series crits
x_3 series hits, x_2 series hits
x_3 series crits, x_2 series crits
x_3 series hits, x_2 series crits

You can't stop the love train, baby.

 

Tiburon11

I had some time to work on it today, so here's what I've written. There are still a lot of things left to do, such as add meta gems and armor penetration. Most of this stuff is off the top of my head, so a lot isn't backed by math. I'd also like to know if this looks like it has potential, or if it's more annoying spam. Any suggestions would be nice. And definitely point out any errors I may have made.

The Hit Table and Itemization

To the best of our knowledge, warrior attacks are based upon a roll table. This may not be the exact way Blizzard’s system does it, but it has been confirmed by Blizzard to give the nearly the same results as their system. It works very simply. Every attack you deal and receive is based off a roll table similar to:

0-28: Miss
28-33.6: Dodge
33.6-58.6: Glancing
58.6-78.6: Crit
78.6-100: Hit

This table is assuming you have no +hit gear and 20% crit, and you are fighting a normal level 73 boss. Say you roll a 10, that would mean your next attack would be a miss.

It is important to note the change when you add, say, 1% hit. The table would now look like this:

0-27: Miss
27-32.6: Dodge
32.6-57.6: Glancing
57.6-77.6: Crit
77.6-100: Hit

While adding 1% hit obviously increased the hit on the table by 1%, it also reduced your miss rate by 1%. The same thing happens when you add crit; you gain 1% crit and lose 1% hit.

Note that this table is different for yellow attacks. Since yellow attacks only have a 9% miss rate and cannot glance, they have a correspondingly higher hit rate.

Crit Cap

As you might have noticed from reading above, there is a point on the table where you will have removed all hit from the table. At this point, adding more crit would be useless, since it would be ignored. This occurs at 42.4% crit for white attacks against a standard level 73 mob. You can, however, raise this cap by adding hit. Every 1% hit you add increases the crit cap by the same amount. Since there are no ways to reduce the dodge or glancing rates (except possibly weapon skill, who’s effects are still relatively unclear), there is a maximum amount of crit you can stack. This occurs for white attacks when you have capped your hit, which means the maximum useful amount of crit occurs at 69.4%. Note that yellow attacks have a much higher cap since they only have a 9% miss rate and they cannot glance.

“Perma Flurry”

This is a myth. Unless you have 100% crit, there will always be a chance that flurry can drop off. 33%, 30% or any similar numbers hold no special value.

Diminishing Returns on Crit

Deep Wounds- This debuff ticks every 3 seconds, but this counter is refreshed every time you crit. Therefore, as you gain more crit, this debuff will refresh more often, and tick less often.

Flurry- Imagine you have 1% crit. Every time you crit, you will most likely consume all three flurry charges afterwards. Now imagine you have 98% crit. If you added 1% more crit, the returns from flurry would be negligible. The chance of flurry falling off would be so low, that the 1% crit you just added would have almost no effect on your flurry uptime.

Hit vs. Crit

Hit and Crit both cause around the same increase in damage. 1% hit adds one attack in every hundred hits, and 1% crit doubles the damage of one hit in every hundred hits. Hit has the added value of triggering trinkets and procs, such as Dragonspine Trophy or Mongoose. It also causes more consistent rage generation, which is an underrated feature. Crit has the added value of triggering Deep Wounds and Flurry, as well as being modified by impale. Crit has a slight advantage, though they are pretty even overall. Thus, you can say that 1% hit is roughly equal to 1% crit.

However, hit caps complicate this. While white attacks cap out at 28% hit, yellow attacks cap out at only 9%. As you can probably realize, stacking hit past 9% will only effect your white damage. An average fury warrior’s white attacks usually make up around 50% of his overall dps. Thus, you can very roughly say that 1% crit is equal to 2% hit once you have over 9% hit.

Note that stacking hit past 9% is worthless for any build which uses a two-handed weapon.

“But you can’t crit if you don’t hit.”

Yes, you can. Adding hit does not increase your chance to crit (unless you have some sort of proc associated with hiting, such as a Dragonspine Trophy). The only exception to this is if you are at the crit cap, which has been explained above.

Crit vs. AP

This is a very tough comparison to make. Crit and AP scale inversely. As you gain more crit, AP becomes more valuable. As you gain more AP, crit becomes more valuable. Because of this, your personal gear, buffs, and group layout can have significant impact on the values of both stats.

The best way to determine which is better for you, is through actual dps testing during a raid. If you don’t wish to do that, you can also try using the spreadsheet.

In general, however:

AP is significantly better when self buffed.
AP is slightly better when raid buffed (no group buffs).
AP is slightly better if you are grouped with a feral druid.
Crit is slightly better if grouped with enhancement shaman and no feral druid.

Crit is superior in a Mortal Strike build, due to normalization.

Haste

At low levels, haste is very comparable to hit. Adding 1% haste will cause 1 extra attack in your next 100 attacks, similar to hit. For example:

How to Itemize or “But don’t I need 200 Hit Rating?”

There are no magical numbers when it comes to warrior damage. Achieving say, 200 hit rating does not suddenly increase your dps. When deciding on an item to equip, just figure out which item will increase your dps by the largest amount. Don’t equip an item just to hit a certain value. In the end, your stats should even out due to blizzard’s itemization. Due not trust generic values such as “AP is worth 1 point, Crit 2 points, etc.” These values can change and vary significantly between players.

Specific Gear

Trinkets

[Madness of the Betrayer] Rumored to have a 2.4ppm proc. This would give it a static value of -120 armor.
[(-300armor*10seconds*2.4ppm)/60seconds]

[Tsunami Talisman] Rumored to have a .9 ppm proc. This would give it a static 51 attack power.
[(340ap*10seconds*.9ppm)/60seconds]

[Dragonspine Trophy] Rumored to have a 1.5 ppm proc. This would give it a static 81 haste rating.
[(325haste*10seconds*1.5ppm)/60seconds]

[Bloodlust Brooch] The “use” function will add 46 attack power if triggered on every cool down.
[(278ap*20seconds)/120seconds]

[Hourglass of the Unraveller] Rumored to have a .9 ppm proc. This would give it a static 45 attack power.
[(300ap*10seconds*.9ppm)/60seconds]

[Abacus of Violent Odds] The “use” function will add 21 haste rating if triggered on every cool down.
[(260haste*10seconds)/120seconds]

Slow MH vs. Fast MH

Slow MH- Larger Whirlwinds, More rage efficient Cleaves

Fast MH- More Heroics Strikes and Cleaves

The slower the MH, the better. However, DPS is still king. If there is a significant difference in DPS, a fast MH may be used. In this situation, you may want to prioritize heroic strike over whirlwind.

Slow OH vs. Fast OH

There is very little difference between the two. The weapon with the highest DPS/stats will win. However, a fast OH is usually preferred since it provides more consistent rage generation.


Talents, Specs, and Abilities

The cookie cutter build is 17/44, laid out like: Talent Calculator - World of Warcraft

Improved Whirlwind

Improved Whirlwind will provide the smallest DPS improvement among talents, which is why there is usually only 1 point put into it. Some people choose to take piercing howl instead of this talent.

There is another advantage to putting one point in improved whirlwind. It causes your whirlwind and bloodthirst cool downs to line up exactly. For example:

0- Bloodthirst
1.5-Whirlwind
6-Bloodthirst
10.5-Whirlwind
12-Bloodthirst
18-Bloodthirst
19.5-Whirlwind
Etc.

Heroic Strike

Keep in mind that this ability costs extra rage. It costs the 15 rage to cast, plus the rage that would have been generated if it were a white hit. Also keep in mind that this is a yellow attack, which means it only has a miss rate of 9% and can't glance.

Slam

Slam's main draw is the fact the it is not normalized. It's damage is calculated using this formula:
Damage=Base Weapon Damage + (Weapon Speed * Attack Power / 14)

Normalized abilities, such as whirlwind and mortal strike, are calculated using this formula:
Damage=Base Weapon Damage + (X* Attack Power / 14)

X is a constant, it does not vary upon your weapon's speed. When you are using a two-hander, this value will be 3.3. When using a one-hander, it is 2.4. For a dagger, it is 1.7.

The extra damage really becomes obvious with very slow weapons. Take for example, a [Stormherald] on a player with 3500 AP.

Slam=579 + (3.8*3500/14)=1524 + 140=1664

Mortal Strike=579 + (3.3*3500/14)=1404 + 210=1614

As you can see, Slam can actually hit harder than MS. It also costs half the rage. The only downfall is that it resets the swing timer. Improved Slam has a casting time of .5 seconds, so if it is used immediately after a white hit, you should only lose .5 seconds + your latency off your swing.

Note that a slower weapon will not greatly increase damage over time. As long as your weapon + buffs stay at less than 2.5 speed (so you can complete the rotation), the extra instant attack damage will be mostly made up by a faster swing time.

Threat

Battle Stance and Berzerker Stance have a passive -20% threat generation mechanic. This combined with salvation (-30%) will usually let you go all out. If you are still having trouble, you can do a number of things:

- Use [Prism of Inner Calm]. This reduces your threat by around 10%.
- Get soulstoned, and die to wipe your agro.
- Limit your heroic strikes.

“Couldn’t I replace Heroic Strikes with Cleaves to do less threat?”

Yes, this is an option. Cleave does actually do more threat per bonus damage. However, the total threat added is lower than heroic strike. Because of this, you can do it more frequently and take advantage of the benefits from impale and and yellow damage.

Rotations

Dual Wield Fury

100%-20%

In this range, always ensure that you have rage (and that you abilities aren’t on the global cool down) to use Bloodthirst every time it comes off cool down. Whirlwind is your next priority, keep it on cool down, but Bloodthirst has priority if they both come off cool down at the same time. Keep rampage up when Bloodthirst and Whirlwind are on global cool down. If you have extra rage after that, spend it on heroic strikes. Additionally, if you extra rage after using heroic strike on every auto attack, you can throw in some hamstrings between Bloodthirsts and Whirlwinds.

20%-0%

Execute. Don’t save a full rage bar for this phase. If you do go into it with a full rage bar, Bloodthirst than start executing. The key to maximizing dps in this phase is to ensure that you can execute on every global cool down. Equipping a fast MH usually helps to do this, by smoothing out your rage generation. Something like [King's Defender] works well. Also, if you have more than 2989AP, Bloodthirst is more efficient than execute.

(30 Rage Execute=1245 Damage; 2989AP Bloodthirst=1245 Damage Thanks Zoroaster)

Mortal Strike/Improved Slam Build

100%-20%

In order to get the most out of your slams, you will want to time them to hit immediately after a swing. To do this efficiently, you will want a swing timer. I recommend using either Swinger or Quartz.

You will than have a rotation like this:

White Attack->Slam->Mortal Strike->White Attack->Slam->Whirlwind->White Attack->Slam->Mortal Strike->White Attack->Slam->Hamstring/Shout

Make sure that the global cool down is free after after a white attack, so you can cast slam freely. "Spam Slam" is not an efficient form of dps. The 1.5 second global cool down means you can only get around 2-3 slams in place of white attacks. Heroic Strike won't be used much here because this rotation is very rage intensive.

Keep in mind that slam basically has a built in /stopcasting macro. You can watch your latency, and start you slam early. For example, if you have 200ms, you can start your slam .2 seconds before your auto attack lands.

If your hasted with something like heroism, or very high end gear, or you have a relatively fast two-hander, you may wish to drop Slam from your rotation. There is a point where cutting .5 seconds off your auto attack is a bigger loss in dps than you gain from the slam. This rotation is also impossible if you swing faster than 2.5 seconds. Again, you would want to cut slam out of your rotation if this occurs.

20%-0%

Continue the same rotation above. Execute spam is not very efficient with a two-hander. If you have heroism, or drink a haste pot, or something similar, than start spamming. Those buffs will make your weapon fast enough to not waste rage. You will still want to keep MS on cooldown if you have the rage. Alternatively, you can equip a couple fast one-handers. If you have around 9% hit, than this can be effective. However, if you have low hit like many MS warriors, you may miss too much to get off consistent executes.

Useful References

Slam Guide Thanks Turpin
Catalog of Theorycraft
DPS Spreadsheet

Reclaimed is currently looking for 2 Mages, 2 Warlocks, and a Holy Priest who are Elitist Jerks just like the rest of our guild.

 

gemmanite

excellent post

the use of Bloodlust Brooch gives 278 AP, giving a static 46.33... AP

 

Rapid

Thankyou Tiburon.

We have had many discussions amongst our DPS Warriors and other Melee classes on the relative value of +hit and +crit. Your post has just alleviated me several hours of headaches.

 

Tiburon11

Woops :P

Reclaimed is currently looking for 2 Mages, 2 Warlocks, and a Holy Priest who are Elitist Jerks just like the rest of our guild.

 

Mundis

Great post Tib.

and yeah, the brooch is 278 AP on use, not 260, 46.33 would be the correct amount.

 

Rezarel

Haste doesn't work like that. 1% haste means you get 1% more attacks in a given time, not that the swing time is reduced by 1%. You can check this by watching your weapon speeds when flurried: they'll drop to Weapon Speed / 1.25, not to Weapon Speed * 0.75.

Going from 0% haste to 1% haste takes a 4.00 weapon to 4.00 / 1.01 = 3.96.
Going from 98% to 99% takes you from 4.00 / 1.98 = 2.02 to 4.00 / 1.99 = 2.01.

Nice post, otherwise.

 

Natural

Very nice post, however, you should read some of the earlier discussion about cleave and heroic strike.

Both heroic strike and cleave add bonuses BEYOND their +damage because they make a white attack a special attack. This adds the following bonuses:

1) Cannot glance
2) Increased chance to hit
3) +10% damage on crits

For example, let's say that "removing glancing blows, reducing chance to miss, and +crit" increases the expected damage of an attack by 200.

This changes your calculation as follows:

Heroic strike adds 176 damage (+200 for hit/glance) for 196 threat. 1.92 damage/threat
Cleave adds 70 damage (+200 for hit/glance) for 130 bonus threat. 2.07 damage/threat

Here, cleave is the victor. If you can calculate the value of these other bonuses (no glance, +hit, +10% crit damage), THEN you can calculate what is better damage per threat.

 

Turpin

Nice Writeup Tib

I believe the hit cap for yellow hits and 2 handed whites is 8.64%

Regarding slam and normalizaion, Ceasar has a nice write up arguing that weapon speed is not a factor for slam on page 3 here:
http://www.honorbound.se/Slamdps2.pdf

 

◊ Zoroaster

Overall very good post. Definitely would help cut down on some of the more commonly asked questions. Couple things that I noticed while reading through your post that I've quoted and commented on below. Not trying to bash on the hard work you put into this though, very nicely done.



"Since yellow attacks only have a 9% miss rate, they have a correspondingly higher hit rate."

Yellow attacks also cannot glance, increasing the crit cap by 25%


"Thus, you can roughly say that 1% hit is equal to 2% crit once you have over 9%."

Shouldn't it be reversed? 1% crit is equal to 2% hit once your yellow attacks are capped? Of course it isn't quite this easy with rage generation and all, but I know the point you are trying to make.


"X is a constant, it does not vary upon your weapon's speed. When you are using a two-hander, this value will be 3.3. When using a one-hander, it is 2.4."

Daggers have a 1.7 speed for normalized attacks, really is noticeable with WW/Overpower.


"Execute. Don’t save a full rage bar for this phase. If you do go into it with a full rage bar, Bloodthirst than start executing. The key to maximizing dps in this phase is to ensure that you can execute on every global cool down. Equipping a fast MH usually helps to do this, by smoothing out your rage generation."

With Hyjal/BT gear I believe Bloodthirst is actually more rage efficent then execute if you have 30 rage. So during execute phase if you get a couple crits while global CD is up and you have 30 rage, don't use it on excute, BT instead. (Can someone else verify this? I have seen a number of comments about it in the past and the napkin math I did seemed to confirm.)

" White Attack->Slam->Mortal Strike->White Attack->Slam->Whirlwind->White Attack->Slam->Hamstring/Shout"

The best rotation in my experience is actually, white - slam - MS - white - slam - WW - white - slam - MS - white - slam - shout/hamstring. Basically a 4 slam rotation, since WW has such a long cooldown.

 

◊ Zoroaster

I think you are misreading his post, or you mean something different. He states that you do not want a 2h faster then 2.5 seconds post flurry/haste, and a bit slower would be preferred since 2.5 speed requires very precise timing of your slams/instants.

 

Tiburon11

I've heard 8.64% and 9%. I've never seen a clear definition on which is right, though, so I decided to go with the rounded version.

Oops, forgot to add that.


Yep, my mistake


I suppose I could add that. I wasn't expecting anyone to MH a dagger though.

I've actually been wondering this too. I would love to see some math behind it.

My mistake again. Thanks for the corrections, it's hard to spot that stuff immediately after writing it.

Reclaimed is currently looking for 2 Mages, 2 Warlocks, and a Holy Priest who are Elitist Jerks just like the rest of our guild.

 

Turpin

top of pg 4 he says "as we can see slow weapon=more damage is a myth". his argument for it is before this.

 

◊ Zoroaster

When you are in the 3.5-3.8 speed range it's not a big difference. However faster then 2.5 (post flurry/haste of course) will hurt your DPS. Granted, there aren't very many 2h weapons that are faster then 3.5, also with the somewhat limited choices in end-game 2h's most people take whatever drops first.

 

◊ Zoroaster

Ok, here is the math for execute vs BT when you have 30 rage. This assumes you have 2/2 imp execute.

30 rage execute does 1345 damage
Bloodthirst does 1345 damage with 2,989AP.


So if you have 2,989AP or higher it's better to BT then execute with 30 rage, the higher your AP the bigger the difference. When I was 17/44 what I just did was hit BT then execute, if you don't have rage for BT then no wasted global CD and you execute.

Commonly accepted fact is that 15 rage executes are your best damage per rage. However once you hit 4573 AP, BT is more rage efficient. Now this requires very good gear, but I'm sure people with mostly Hyjal/BT gear are either at this point already, or nearing it. I know I was often around 4000 AP as 17/44, this was with 2 pieces of tier 6 loot, the haste ring off trash and a Swiftsteel mace.

 

Gruntle

You can't calculate the values of damage per threat like that. You have to include the bonus damage in the threat done to compare Cleave and HS (they both do more threat than just the bonus threat). It seems you are right however, including the white damage part of HS/Cleave seems to indicate that if your average swing damage is greater than around 140 Cleave becomes better than HS in terms of damage/threat (and Imp. Cleave is better already with 0 swing damage).

If you include 130 swing damage you get damage/threat values like this:
HS: damage/threat = 306/(306+196) = 0.610
Cleave: 200/(200+130) = 0.606

[top]> i.e. HS is better than Cleave
If you assume a 150 swing damage:
HS: damage/threat

326/(326+196) = 0.624
Cleave: 220/(220+130) = 0.629
=> Cleave is better.

Not sure I went wrong somewhere, but this is actually quite interesting, I never thought that including the swing damage could make such a difference.

edit: this is just wrong, read the later posts by Natural. It turns out that the turnover point in efficiency for HS/Cleave comes at the same value of extra damage (due to 9 %miss and no glances) regardless of if you model the efficiency as extra damage/bonus threat or as extra damage /(extra damage*mod + bonus therat)

 

Gruntle

A very nice and good guide, thanks for writing it Tiburon.

I agree with the 200 hit rating statement of course but using static values for the worth of different stats is very well supported by theorycrafting (in the spreadsheet). Once you have reached 9% +hit the relative value of the different stats stay pretty constant (at least they don't switch places) for a specific build. For DW fury the order (after 9% hit): 1 haste ~> 1 str/0.5 AP > 1 crit rating > 1 hit rating, is supported by the spreadsheets for all gear setups I've tried. Maybe the spreadsheets do not mirror "reality", but if you trust the theorycrafting then this seems to be how it works.

About the 9% miss chance. A lot of testing in the weapon skill thread showed that the base miss chance is 9% (with a very small margin of error) vs 73 mobs, without any weapon skill. It might be interesting to include a section on weapon skill in this guide. Check that thread for the results (one result is that the first 5 weapon skill gives you 3% hit).

 

D4vE

If this is true [Belt of One-Hundred Deaths] is the best fury dps belt.