Fury DPS Spreadsheet for 3.0 - Page 20

Halle · 05/09/08, 7:03 AM

Yes, all re-appears on scrolling or selecting all so it looks like a simple screen redraw problem.

However this doesnt explain Ctrl+Shift+Y not doing anything

edit: Disregard, latest Open Office patch sorted it

landsoul · 05/10/08, 11:31 PM

I found the [Netherbane] error. It wasn't using the correct normalized speed because I misspelled it there.

Kaan · 05/11/08, 8:51 AM

[Muramasa] has its stats counted 2 times (34 hit and 70 Haste).

Edit: wielding in offhand [Muramasa] has counted zero stats.

landsoul · 05/12/08, 4:41 PM

Damn, i'll get it fixed sometime tonite, sorry. I am starting to suck at these updates. I need to check and re-check this stuff before I put up the file.

Kaan · 05/13/08, 12:10 AM

Originally Posted by landsoul

Damn, i'll get it fixed sometime tonite, sorry. I am starting to suck at these updates. I need to check and re-check this stuff before I put up the file.

No prob for me, was just interested how Muramasa would perform compared to S3/Warglaive/S4 weapons.

Keep up the good work.

shed · 05/13/08, 12:34 AM

Just eyeballing it, I think S4 will probably be better because it has crit and some arp. The haste on Muramasa is nice but the hit is sort of wasted.

Kaan · 05/13/08, 12:42 AM

Yeah, I subtracted the excess stats from Muramasa and entered it ... atm Muramasa would be +3 DPS to a S3 mainhand, and -11 DPS to S4 with my equip (full raidbuffs w/o BF/LOTP, demonslaying, 7700 armor, demon boss), so nowhere near that good as it should be.

landsoul · 05/13/08, 2:08 AM

Yeah Muramasa is just a little bit over s3 for me also. Would be a sick rogue MH though. Also, the fix is up with the MH/OH Muramasa bug.

Beefcakes · 05/13/08, 6:07 AM

Can you add [Carapace of Sun and Shadow] on the sheet? many have taken lw for brut so would like to see how better this is from t6 and on.

Also is bloodfury implemented for ze orcs in the sheet? couldn't find it or I'm blind :p. would be cool to be selectable along with the 2min consumables :p.
Cheers

Gellor · 05/13/08, 5:00 PM

Just an FYI, but DW dragonstrikes, the haste proc now refreshes rather than stacks.

brutalbovine · 05/14/08, 2:44 AM

Originally Posted by Gellor

Just an FYI, but DW dragonstrikes, the haste proc now refreshes rather than stacks.

This was already pointed out in the thread, post #468.

shed · 05/14/08, 11:59 AM

Originally Posted by brutalbovine

This was already pointed out in the thread, post #468.

However in post #468, 2.4.2 hadn't hit live servers, thanks for the info brutalbovine. Saves me some money in testing it out myself.

zinv · 05/14/08, 1:02 PM

Hi, I have a question regarding the forthcoming of the Execute model: The spreadsheet right now models with T(fightdur)=infinite but in an infinite fight you cannot reach 20%. So the whole spreadsheet has to be placed on Boss HP and according to those and the own dps a new T hast to be modeled in.

Because you and other classes deal more dmg during 20%
(or even from 35% on, i think thats hemo-rogues and fire mages from 30% on)
you will spend less time in execute phase and thus the dps gain from execute will result in an overall smaller dps than it can be modeled.

basically the spreadsheet would have to be changed to rather calculate the dmg done in cycles than single abilites, that would also allow the modeling of improved whirlwind and improved execute.

all you need then for an Execute model would be Rage. you have your RPS, you have your Execute rage cost
first you check whether bloodthirst / whirlwind deal higher damage than a single normalized execute, if not it will not be used in the cycle

so simplest execute model : spam
take rage per second * 1.5 = rage available for an execute
if your RP1.5s is less than the rage needed for an execute, simply (rage needed for execute, sorry for the acronym = RN4E) RN4E/RPS=Te (time between executes)
So your cycle dps for execute would be (spam cycle consists of ONE use) Executedmg/Te + white dps
of course one has to take in calculation the increased frequency of instant attacks --> more procs

now for the bloodthirst / execute cycle, we start with 30 rage, Te seconds minimum is 1.5 secs , assuming Te=1.75
0.00 - bloodthirst
1.50 - no rage for execute
1.75 - execute
3.25 - no rage for execute
3.50 - execute
5.00 - no rage for execute, bloodthirst not ready
5.25 - now is the question for this execute cycle, would it be better to use another execute and then use bloodthirst at 6.75s or wait to 6.00 seconds and use bloodthirst ?

so you take the two cycles, compare the dmg done to the cycle length and chose the better one

I think that it could be easier to use execute at the beginning of each cycle and then bloodthirst so you dont have to calculate leftover-rage, but then its not me who is developing the sheet ^^

I dont really see the problem in the Execute, the problem lies within the rage!
Because you cannot use a 08/15 model but rather have to use hits within 1.5 seconds and normalized rage gain of those hits.

landsoul · 05/14/08, 4:21 PM

Originally Posted by shed

However in post #468, 2.4.2 hadn't hit live servers, thanks for the info brutalbovine. Saves me some money in testing it out myself.

This would not change how it is in the sheet now. I'd have to look at it again to make sure, but I did a lot of thinking on it when I put it in. Simplified down, its %uptimeMH + %uptimeOH = %uptimeTOTAL. You can take a look at it in the sheet as well.

Edit coming for thought on execute modeling.

Crimsonstorm · 05/14/08, 4:58 PM

Originally Posted by landsoul

This would not change how it is in the sheet now. I'd have to look at it again to make sure, but I did a lot of thinking on it when I put it in. Simplified down, its %uptimeMH + %uptimeOH = %uptimeTOTAL. You can take a look at it in the sheet as well.

If the sheet was already accounting for the fact that a proc happening when the buff was already up is just a refresh, not a full extra buff, then yeah, it wouldnt effect things. Having MH + OH dragonstrikes just increases the PPM rate.

Opta · 05/14/08, 6:00 PM

S4 are more dps then Muramasa but the 2050 rating for s4 weapon even with lot of skill is going to take some (or a lot) time to get.

shed · 05/14/08, 11:42 PM

Originally Posted by Opta

S4 are more dps then Muramasa but the 2050 rating for s4 weapon even with lot of skill is going to take some (or a lot) time to get.

Ya, I will likely go in on Muramasa since I have no interest in grinding up to 2050 rating (especially early on when S4 hits).

Gruntle · 05/15/08, 11:20 AM

Originally Posted by landsoul

This would not change how it is in the sheet now. I'd have to look at it again to make sure, but I did a lot of thinking on it when I put it in. Simplified down, its %uptimeMH + %uptimeOH = %uptimeTOTAL. You can take a look at it in the sheet as well.

Edit coming for thought on execute modeling.

Hmm, will that be correct modeling of proc uptimes? If you get a proc while the haste is up the old one will be overwritten, this means that the total uptime will be smaller than the sum of the MH and OH uptimes. I made a quick python script to check the permutations with only white attacks (and 0% miss/dodges) and 1.4 ppm, the separate uptimes were then ~20% while the combined uptime was ~35%. DS combo will be overvalued in the spreadsheet by a couple of percents from this.

Crimsonstorm · 05/15/08, 6:19 PM

S4 Gloves are showing up as having 0 stats for me in the latest spreadsheet.

landsoul · 05/18/08, 2:16 AM

Originally Posted by zinv

[about execute model]

This isn't something that I haven't already thought of. The problem is way more complicated than this.

here are some issues I have with the sheet

1) It needs to be believably accurate
2) It has to consider many different preferences of DPS optimization
3) It has to be small enough to where its time to edit or tweak remains reasonable

While current know-how can fulfill the first problem, it doesn't fulfill the 2nd or 3rd problems.

-Rage generation frequency and amplitude in relation to Global Cooldown
Imagine a graph where you plotted expected rage gain with respect to time, and then plotted another line with a different gear and weapon setup. How would they be the same? How would they differ? How would different buff and raid orientation makeups effect these plots? How would using different talents affect this makeup? How would using massive cooldowns during this phase affect it? How would different thresholds of AP affect our cycle? When does Bloodthirst or even Whirlwind begin to become more effective than execute and at what rage levels?

I think at this point we might be ready to do some research to determine what gear-stats are best for executing, but are we no way near being able to wright a usable tool that helps us make very important gear decisions. I think we have gotten the normal DPS cycle to a very usable, reliable, and accurate state. Executing may be a little bit too advanced for us right now. I do think that people should definitely consider execute when gemming or selecting gear, but I'm afraid we don't have solid answers yet.

Jevade · 05/22/08, 6:17 AM

Not only gemming but switching weopons out during the execute phase and knowing if you should save recklessness and or bloodlust for executing. This spreadsheet kicks ass and is highly acurate but we can only use it for cycle dps. Not only hard but basically different for every fight since rage is highly influenced by incoming damage too so knowing if you should use other attacks or just the executes would not only have to consider current rage but incoming rage per second for that phase of the fight ect..

I throw in a combination of fast/fast for the executing phase - which seems do about 200 more dps than my main weopons i use for cycling- funny thing is if i were to put those execute weopons in the spreadsheet my dps goes down significantly-- but not for executing.. in fact if you use slow/slow you might even be executing for less damage than your overal cycle is doing. To give you some estimated numbers i was doing 1400 DPS in a fight.. the executing didnt change that DPS by much and often times went down....where as i put in daggers that would gime 1100 DPS for cycle and 1620 DPS for executes. Using different weopons for executes is definately the way to optimize your overal damage.. just because we cant graph out a good estimated number that goes up and down with stats doesnt mean shouldnt find yourself at least one fast weop for the execute stages imo.

zeux · 05/22/08, 11:28 AM

Originally Posted by Jevade

Not only gemming but switching weopons out during the execute phase and knowing if you should save recklessness and or bloodlust for executing. This spreadsheet kicks ass and is highly acurate but we can only use it for cycle dps. Not only hard but basically different for every fight since rage is highly influenced by incoming damage too so knowing if you should use other attacks or just the executes would not only have to consider current rage but incoming rage per second for that phase of the fight ect..

I throw in a combination of fast/fast for the executing phase - which seems do about 200 more dps than my main weopons i use for cycling- funny thing is if i were to put those execute weopons in the spreadsheet my dps goes down significantly-- but not for executing.. in fact if you use slow/slow you might even be executing for less damage than your overal cycle is doing. To give you some estimated numbers i was doing 1400 DPS in a fight.. the executing didnt change that DPS by much and often times went down....where as i put in daggers that would gime 1100 DPS for cycle and 1620 DPS for executes. Using different weopons for executes is definately the way to optimize your overal damage.. just because we cant graph out a good estimated number that goes up and down with stats doesnt mean shouldnt find yourself at least one fast weop for the execute stages imo.

I'm not 100% sure about this but I think Bloodthirst yields more dps than execute does IF you got more than 3500AP buffed or close to that.

edit: Now I saw you're slam specced, but what I said should apply to fury specc atleast. For slam specc execute probably would be better, anyone can confirm any of this that I'm saying, 'cause I just remember hearing it somewhere.

landsoul · 05/22/08, 12:24 PM

That is 100% correct. BT > execute when around that Ap level. Not sure exactly the number off my head.

Im releasing the Kil'Jaeden Loot Table update later today.

zeux · 05/22/08, 12:30 PM

Originally Posted by landsoul

That is 100% correct. BT > execute when around that Ap level. Not sure exactly the number off my head.

Im releasing the Kil'Jaeden Loot Table update later today.

Could you implement the Kel'thuzad trinket aswell? 150AP constantly (on UD/demon) should outpreform many TBC trinkets.

edit: Mark of the Champion that would be.

Wump · 05/22/08, 12:54 PM

Just curious as this is the second thread I've seen in the past 10 minutes mention Kil'Jaeden's loot table, but where did it come from? Was it unearthed in the Alpha datamining? Or has someone downed him and I've totally missed the boat?