Gruntle

Check the DPS warrior compendium pindle, this is not the right thread for these questions.

 

talliara

I have yet to see a slow off hand be better than a fast off hand weapon. I don't know why spreadsheets undervalue the dps numbers when a fast off hand is used. I've tested both slow badge fists against the slow MH and the fast OH badge weapons. I never found dps to change no matter which set was used; a little over 2000dps on anetheron with the fast OH or the slow OH. The AoE damage going through Hyjal also was not changed much with a fast or slow off hand, usually between top 3 and top 5 for overall damage on trash.

I just feel that the spreadsheets showing such a big boost in dps by going with a slow off hand, is inaccurate.

 

• landsoul

What benefit do you have of having a fast offhand? 3% more heroic strikes? 2% tighter rotations? What you feel is better? Rage generation is exactly the same amount. Its not more with a fast offhand. The difference comes from the WW damage. Its a calculator, so it really does not give care about what your personal preference is.

 

Gruntle

I agree completely talliara. It's known problem with the spreadsheets because they cannot model the effect of miss streaks on rage generation. It may be that slow offhands are slightly better, but the inherent variance of dps for a certain boss fight is probably significantly larger than this difference, making it impossible to find in practice.

That said, I don't think the boost the spreadsheet shows is that large, going from fast to slow offhand shows a dps drop by about 0.5% for me, that's not a lot. In fact I would say that all gear changes that result in dps changes that small should probably be taken with a grain of salt due to inaccuracies of the modeling.

 

talliara

I don't feel the rage generation is equal, I notice I can use more heroic strikes when I use a fast off hand over a slow one. I can't say I agree that rage generation is equal between the two, if I get two crits with a fast off hand, I generate more rage than 1 attack that doesn't crit with a slow off hand. While I realize spreadsheets calculate solid numbers, the WoW combat attack table is not so solid in making things happen to an accurate portrayal of stats.

Example: I have 32.97% crit unbuffed, on Kalecgos last week, there was an attempt where I had bloodthirst crit only 11% of the time, and my normal attacks were around 23-26%. As I understand it, spreadsheets calculate numbers in a static way; if that is true, then yes, a spreadsheet is going to show that a slow off hand is going to have the same, or very similar, rage generation.

As I originally stated though, the amount of times I can use heroic strike is considerably higher with a fast off hand and I can still maintain keeping up demo shout, battle shout and rampage, while hitting bloodthirst and whirlwind every cooldown.

Also, I can't say that I see lower numbers with whirlwind hits when I use a fast off hand, on an AoE mob my whirlwind with a fast off hand (all 8 crits) hits for 11-14k. With a slow off hand I've never seen over 14k with 8 crits on a whirlwind. Testing of course in hyjal with recklessness up for a full 8 crit whirlwind attack.

The simple answer is, I can use heroic strike much more often with a fast off hand than I can with a slow off hand. Perhaps not because a fast off hand is better, but because of the RNG.

And yes, I'm a broken record. lol

 

Gruntle

I'm parroting you again, but I have exactly the same experience when it comes to Heroic strikes (even though we're only in MH/BT so far).

The reason you're not seeing the effect on WW damage is that the effect is tiny. It's only the weapon damage that changes, the major part of your WW damage (in particular with the insane AP levels you reach in endgame gear and with full raid buffs) comes from normalized AP which is exactly the same for slow and fast offhands (unless you use a dagger, which is a bad idea).

 

Graul

This has been my argument since the whole "which is better" debate started. It's also why I refuse to dip below a certain amount of hit using slow/slow. You can get away with having less hit using slow/fast in comparison. Given both setups, rage generation will be "the same" but the rate at which you see it will play a larger role in your rotation. You can also get away with ignoring Rampage for a few more seconds using a faster off hand.

 

• landsoul

Fast offhands do favor shitty RNG or whatever the excuse is, yes I agree. But if you are very skilled with Slow/Slow, it is very very rare that you can't keep a perfect rotation or use as many heroic strikes. Having a slow offhand does not change the speed in which you use heroic strike. The main hand speed is the same between both combinations. If you cant use your abilities because you heroic strike at the wrong time that's player error.

Spreadsheets run off of absolutes. Slow/Slow is better by a little bit when dealing with absolutes.

 

talliara

Since WoW doesn't use absolute values the difference between the two will never be more than a couple dps at best.

As far as heroic strike usage with slow/slow, it's never been a problem, I just like that I am able to use it more liberally with a fast off hand.

I just find it weird that spreadsheets show a 20-40dps loss by going with a fast off hand, when the difference; if any, is going to be minimal. Again I realize that this is because the spreadsheets use absolute value, and there is probably not really a way to change how that is affected, but I am not positive. I just know that WWS has never shown a difference in dps between a slow and fast off hand with equal stats and dps.

 

• landsoul

I am making changes right now so that all avoided swings (dodge, miss) also consume flurry charges. Build is up. Also added inc to BT from stones.

 

Kaan

@talliara: landsoul is absolutely right, when you are skilled with slow/slow the difference is very very small. When you do less HS with slow/slow you are wasting in fact rage - and that's player fault/practise.

20-40 DPS difference between slow/fast with the absolutely same offhand is simply wrong information. For example, when you compare S4 slow offhand with S4 fast offhand on a very high gear/buff setup, the difference is at most 20 dps.

 

Gurlock

Nevermind.

I can't delete this post? Hmm.

 

Bronwyn

Landsoul, I have several questions on executes.

1) Does your spreadsheet factor in executes at all?
2) Somewhere you said that WW gives more damage/rage than execute. What is the AP value necessary to make a 25 rage ww more valuable than a 25 rage execute?

http://www.dkpsigs.com/sigs/Alleria/79591.jpg

 

Glory

Landsouls Sheet does not factor Execute (Points in imp Execute do not change anything)

Make your own Evaluation

Execute : 925+ Rage over * 21 (--> 1135 Without T6 ->> 1198 With T6 ->> 1240 With T6 + 1Point imp Execute)
WW : (Weapon DMG MH + Weapon DMG OH*OH factor)+(2.4*Ap/14*(MH factor+OH factor)) (Assuming you dont wear Daggers)

-> For 2x S3 Slow weapons and 2*T6+ 1Point in Imp Execute:

Solve(1240=435.5+(Ap/14*2.4*1.625),Ap)

-> Ap >= 2888.

-> For 90 DPS Weapons with 2.6 Speed Ap >= 3068

-> For Mounting Vengeance and Hand of the Deceiver -> Ap >=2754

For Mounting Vengeance and Hand of the Deceiver:

WW > BT > Execute for Ap < 4895
BT > WW > Execute for Ap >= 4895 per Rage Point.

 

Bronwyn

So the optimal cycle would be the regular BT-WW-BT-WW-BT, filling the blank CDs with Executes, right? I.e.: BT-WW-EX-EX-BT-EX-EX-WW, etc.

But what if, after the second Execute your rage allows an immediate execute but no BT? Wait or execute ?

http://www.dkpsigs.com/sigs/Alleria/79591.jpg

 

Glory

Assume 4000 Ap

Execute with 12 Rage (2*T6) -> 77 DMG per Rage

Execute with 20 Rage (2*T6) -> 54 DMG per Rage

WW has about 64 DMG per Rage
BT has 60 DMG per Rage

I think always Execute with 2 very Fast Weapons is Best, if you have Bloodrage ready use it and you can eventuals BT/WW and go on with Execute

 

Kaan

The goal is to make an instant every GCD, so don't wait for rage (for BT/WW), when you can execute.

 

Jothay

@landsoul:

First, I apologize if you have already covered this on a past posting.

I noticed when brining the sheet up that the stat lookups on the items are if/then statements regarding specific items with fall-through logic to end at 0 if nothing matches. As an alternative that would be a lot easier to program, I would recommend using a VLOOKUP against a table with all of the items in it. This would make it a lot easier to manipulate the item tables and cut the size of the formulas on the Gear Display sheet down to very minimal sizes.

The following is an example for looking up the AP of Breeching Comet in the Main Hand
Create a sheet named ItemList with the following colums:
IName, STR, AP, AGI, HIT, Crit, ArP, EXP, Haste, DPS/Spd
List out the items and their various pertinent stats (can even add more columns for sockets and socket bonus to make that easier too)

This will look up C70 (The Main Hand Weapon's name) in the ItemList (change the range to the actual table) pulling column 3, which is AP as seen above and state that the list does not have to be sorted in order to find the item (the "false")

This will pull back a 40 for the Breeching Comet and is much smaller and easier than trying to program individual item stats into every single cell.

 

• landsoul

I am doing this in the new sheet, however there are certain glaring problems with VLOOKUP. In the end it will make the sheet faster (I hope) and take up less space, being better. It is definitely though a more of a pain in the butt to write and make error free. I am having the gear selection area search through the tables on the item page and return a number in the statistics area. Same with gems, enchants.

I also plan to try an automatic cycle builder that will prompt you for your prioritized GCD<>rage rotation and construct a section of the cycle based on the prompt. And in the end, the calculation will return the DPS of a specific amount of cycle periods based on a customizable "length of a fight." This will also hopefully allow penalization of an unmet rage requirement based on your customized rotation setup.

 

Jothay

When I converted the sheet to Excel format myself, it took me a while to get all of the formulas moved over to VLOOKUP and it seemed to work faster. The biggest problem was figuring out how to make the sheet's circular references stop giving errors. I found in Excel that you could activate Iterative Calculations and it would allow it to cycle automatically, you would not need to hit Ctrl+Shift+Y multiple times as you do in Open Office.

Once this was completed, I was faced with a problem that perplexed me for several days. The sheet had a big fat #DIV/0 error listed in nearly every cell. What I determined was that a zero was in the sheet somewhere but due to the circular referencing, it was unable to work itself out. The resolution was to apply fall-through logic to create constants in the calculations so that if it read an error, it would apply a constant of 0. This may seem redundant to it's initial problem, but after inserting enough failsafes, the calculations suddenly started working. It wasn't erroring out and wasn't telling me my the DPS was zero.

However, the calculations were diverged from the original sheet's statements, at first drastically. After fuddling with the sheet for a few more hours I discovered I had accidentally modified a few of the original formulea. Once corrected, the new excel sheet now reads the primary output dps on a proportionate (roughly) 10% difference from the original sheet. I believe this to be a result of the iterative calculations providing a difference in the numbers. Unfortunately, my attempts to resolve this and coverge my edited sheet's numbers with your own have failed, I will continue to examine the difference in the sheets to attempt to correct the imbalance.

Frankly, I'm amazed at the incredible amount of work that you have placed into making this sheet possible and from one person with programming and calculative experience to another, bravo.

 

Graul

What would the formula be to determine your average special attack if you had say 35% crit and Impale? Also, what would the formula be to determine your average main hand or off hand auto after glancing blows at that same crit rate?

I'm trying to compare 2x 1h (Incite) using the lvl 80 Heroic every MH swing to 2x 2h using no Heroics at 8000 AP (before the AP from weapons gets added) and with WF and Flurry up 100% of the time and weapons that are equal item level, but I know the math is being screwed up somewhere because the 2h setup is coming out to 4359 DPS and the 1h setup 4469. This assumes 0% misses and no Heroics with the 2h setup, which wouldn't be happening (and no Deep Wounds), but then even so the results should not be like this either.

 

Gruntle

To Jothay: Hehe, this is, incidentally, exactly the behavior you would get with trying to use the old excel spreadsheet (drALLCom3's sheet) in OpenOffice, took quite a lot of tweaking to make it work. It seems OO and Excel handles circular references differently.

Graul,
Non-crit average damage (assume same MH and OH): D
Hit, Crit, glance are percentage values in decimals (assume expertise capped).
Assume +crit damage metagem.

White attack (MH): D*glance*0.76 + D * (1 - crit - glance -(0.28-hit)) + 2.06*D*crit
White attack (OH): 0.625 * White attack (MH)
Hit/exp capped heroic strike: (D+176)*(1 - crit) + (D+176)*2.27*crit

What's wrong with the results? You mean that the 2*1h should come out even more ahead? I think it looks reasonable. The fact that 2h weapons of the same itemlvl have about 30% more dps will be more or less equivalent to using HS on every white attack. Or do you mean that the dps values are too high? I can't really see why that is without looking at your calculations in detail.

 

• landsoul

After doing some experiments with Vulajin last night trying to pin down the armor level of the Dalaran Heroic Dummy, I made a painful realization that I made a mistake in this sheet. The (what I call Z value) 50% armor value for 70 is what is used for the attacker to determine amount applied through armor, not the mob level's value. I updated and uploaded the sheet with the correct value.

 

Gruntle

Hmm, I'm not quite sure I understood what the fix did. Was ArP overvalued or undervalued in the sheet before this fix?

 

belth

Maybe at some point a spreadsheet fails, such as for fury warriors.

Fast offhand = steady rage generation that allows for more heroic strike swings.
Heroic strike is ~35% of your total damage.
Heroic strike won't miss as long as hit rating is capped.

Slow offhand = more burst rage generation, if you miss the swing, you miss out on the heroic strike, and you might miss your next main hand swing.

Also, there are no slow off hands with expertise on them, so you're stuck with the crappy shard of contempt for the rest of your life.

And of course, Glaive set is always the best for a fury warrior, which is of course with a 1.4 offhand.

And now I can't remember why I even went to EJ in the first place, damn fury warrior threads.