I think the armor penetration procs are not modelled correctly in the dps sheet.

Warp Spring Coil, for example, modeled as -350 armor (which is probably equal to [Armor Penetration Value]*[Uptime of the proc]). However, 1000 armor penetration for half of the time is not equal to 500 armor penetration all the time, the first one yielding more dps than the last.

Also, in my experience, Warp Spring Coil has a 10% chance to apply the buff for each special attack. It approaches to 1/2 uptime for a hemorrhage using rogue (15 secs / 35 secs). However, the sheet assumes around 1/3 uptime.

I've done calculations on this very topic, and it turns out that the overall DPS effect provided by X armor penetration with U uptime is actually pretty close to having an X*U armor penetration effect all the time. The difference is small enough that the approximation works unless you have a ton of separate really small armor penetration effects with less than 100% uptime.

Warp-Spring Coil has a 30-second cooldown and a 25% chance to proc on any special attack. An uptime figure of 35% is roughly correct.

This is mostly true, except in low-armor situations. In particular, against a Hyjal boss (6200 armor) with Sunders, Faerie Fire, and Curse of Recklessness, if one has both WSC and Executioner, one can observe a reasonably significant difference between the average case and the proc model. We're still not talking world-shaking differences, but, depending on the details of your implementation, it can be off by up to 5-10 DPS - even with the best possible implementation of the approximation, it'll be off by 2-3 DPS. So I do think there is value in doing the actual proc uptime computation + model for WSC and Executioner, due to them being large amounts of armor pen while active. Madness of the Betrayer is small enough to probably not be worth messing with, though.

True enough. Here's the formula I worked out for the average armor penetration effect, X, provided by an armor penetration debuff P with uptime U, assuming the target has A armor to start with. F is a constant, 10557.5 if you're level 70.

X = A + F - (A + F)(A + F - P) / (A + F - P + UP)

By this formula, on a fully-debuffed 6200 armor boss (Sunder + FF + CoR = 4010 constant ArP), if Warp-Spring Coil has 35% uptime, then its static armor penetration value is actually ~369.

I have a question about how the Hemo debuff works. I often use the spreadsheet for PVP and am therefore looking for personal DPS only.

I understand that the "Include Hemo debuff DPS estimate" check box treats the debuff as if a whole raid will be using all the charges and makes an estimate. If I have that box unchecked is my unbuffed DPS treated as if I am using charges on my attacks?

I noticed in the Unbuffed and Buffed DPS tabs my attacks do not gain any damage when I am Hemo, so I am assuming that I am not gaining any DPS from the debuff. If that is the case then it would be safe to assume I am actually easily doing 5% more personal DPS in a setting where I am the only one using charges then what the sheet is telling me.

Thanks in advance for any help. I've been a long time user of this great tool.

Dagger of Bad Mojo seems to be missing as a mainhand dagger in this last release. It's listed as an offhand, but not as a mainhand.

Also, Ancient Sin'dorei Longbow has been renamed Ancient Amani Longbow in 2.3.2, same stats.

Thank you. Correction and suggested addition noted.

It's on my "to do" list to correct in some fashion. I don't mind that the sheet uses averages for AP, agility, crit and the like as those are linear functions by themselves (i.e. adding 200 AP for 50% of the time is equivalent to 50 AP 100% of the time; add 1% increased to hit 1/2 the time and by itself, thats equivalent to 0.5% to hit all the time - at least if you aren't butting against the hit cap). This doesn't account for synergistic effects, but it's a purer modeling than with armor penetration.

My initial thought was that I could take the percentage possibilities on and off, arrive at the change in % damage done, then back calculate to the amount of armor penetration that is equivalent to said damage done to be more accurate. It appears Vulajin has essentially done that with his formula. The only thing that wouldn't be so accurate is when multiple armor pen proc effects stack as the initial armor would essentially change. Still, it does better reflect reality than the current modeling.

It gets a little more complicated than that to implement as we still have buffed vs unbuffed and there are inherent problems with models that are based on procs when those procs are partially determined by the cycle you are using. I've seen that the initial developers went to some lengths to avoid circular references (and there are some solid reasons why including compatibility issues). Still, I plan to do something to at least give a bit more boost to armor penetration as it's currently consistancy undervalued when it's not a 100% effect.

I've noticed that when adding or removing serrated blades from the talents page, DPS is affected accordingly, however, the -Armor tallies don't change at all, and there's no entry placed in the -Armor field for talents on the Gear_Buffs page. Is this being accounted for somewhere else, and is it possible to make it more "standardized", and add it to the other totals if so?

@ Dontmindme

you've done an amazing job over the last weeks improving the sheet. it is so much more comfortable and looks awesome.

as a rogue who uses the spreadsheet to optimize his dps since the first version, I just wanted to say thank you!


btw. totally in love with the armory-import. saves me so much time when i compare my gear setup to other rogues on my server <3

I found a problem with the 200ap to demons Twinblades set bonus.

Right off I notice that there is no equation to confirm that you have both Warglaives. This should add 200ap to the MH when you're fighting a demon, but for whatever reason my sheet isn't registering 1 = TRUE (at least for this cell), resulting in the 200ap bonus not being calculated into my dps, even with only one of the main hand.

Changing Gear_Buffs Z36 to the following fixes both problem for me, but you're welcome to sleek it up a bit Dmm.

Yes this is precisely what I saw that made me look further. Whats happening is you're getting the dps boost from each sword, as well as the haste proc. It isn't, however including the 200ap, and when you select demon you lose the 2% damage from murder, so your dps actually goes down.

Has anyone provided any like substantial tests over say a long amount of attacks of whether mongoose or executioner would be better? And if not or if so, what's the conclusion? I looked back a few posts and earlier ones but no avail.

[Raid] Executioner vs. Mongoose, preliminary numbers
They are no longer "preliminary" numbers.

Actually already fixed for next version, there was a post about it a few pages back. I noticed the inconsistent set coding when it was brought up that the Warglaives weren't giving a bonus to Demons. I'm actually phasing out "TRUE" and "FALSE" where possible (primarily because using such conflicts with non-english language versions of Excel at times). I'm hoping to post a new version by the end of the weekend. Thanks especially to Pulynerain, there will be some very nice added features. You will be able to designate Gear and Talent sets that you can load on demand. I'll probably include some default sets.

Actually I haven't noticed any problems with occurrences of TRUE or FALSE yet, and I'm using the German version. Afaik Excel is 'intelligent' enough to replace those expressions with the correct localized ones, just like IF, etc.

Corrected in the upcoming version. I actually forgot I even added that to the page. I was playing around with what I might add to the boxes up there, plugged in some quick formulas to see how it looked. I've now linked the -Armor boxes to the totals on the (Un)Buffed DPS pages which should account for all additions.

I have started a new thread as this one has become rather long and a new thread will let me update the first few posts to keep the project up to date. If a moderator could, pretty please, close this thread it would be appreciated. For all others, please post in the new thread.