dr_AllCOM3

Shouldn't MH Executioner only be more like 25%+? It confuses me.

 

• songster

Autoattack only, so you'd expect to be exactly half the number of procs compared to dual executioner. However, these stats are definitely skewed by the small number effect. If you look, you'll see that he only had 10 executioner procs on the executioner/Mongoose test. Crit rate is also drastically variable between tests, probably again from the small number effect (though the 120 Agi from mongoose will contribute).

In any case, with autoattacks only, your proc rate is drastically lowered, so you won't really uncover the full effect of dual procs. So this test is certainly nonoptimal in that regard, which is strange since as I understand it he deliberately set out to discover the effects of dual procs.

We need hours of data, not 15 minutes, and we need it to not be excluding specials.

 

Osse

The only viable way I see to test this is to get someone with 2190 or 3690 armor to stand in Nagrand / Blade's Edge / Gurubashi arena with two or more healers while he/she takes the beating from someone.

 

Gokey

I don't understand why he didn't use yellow attacks... am I missing something here?

 

Obould

Why should i use something like this, raw white dmg with same weapons (thus same weapon speed on attacks) will give a more accurate number regarding raw procs on white attacks.

Of course, the PPM be less, but not drastically less, i mean check double_mongoose and double_executioner:

dm: 1171 hits
de: 1127 hits

thats almost 50 more attacks, gained through speed increase on mongoose (double procs etc), yet the rate of procs is "the" same.

If i would use every possible skill during this test, someone will come along check the WWS and then argue that i Bloodthirsed 4-5 times more in this sample then the other and say its compeltly off.

I can tell some things:

- regarding the Crit% rate while using double mongoose, its not off nor was i lucky, i could have done this for an hour too and it would +/- 1%.

- harder hits while under the effect of Executioner give more dmg, so it would be interresting to test it on a 2H and compare hits.

 

Obould

Okay i just did 14mins of

- MH (Rising Tide) - Executioner
- OH (Swiftsteel Bludgeon) - Mongoose

i was using Bloodthirst / WW more or less on Cooldown, Heroic Striked when possible and kept BattleShout + Rampage up.

Wow Web Stats



First of all ignore the crit rating in the ProcMeter, its way off no idea why tbh, but if you check from UptimeMeter its around 42-43% which WWS supports.

Also just with BloodThirst / WW (not counting HS since it comes instead of a white hit) i was able to raise uptime of Executioner quite nicly (so it is actually dramatically less without doing those instant attacks).

==========================================

- MH (Swiftsteel Bludgeon) - Executioner
- OH (Swiftsteel Bludgeon) - Mongoose

no WWS

 

Obould

And here again, this time over 100mins, ONLY WHITE attacks

- MH (Rising Tide) - Executioner
- OH (Swiftsteel Bludgeon) - Mongoose




=================================0

- MH (Swiftsteel Bludgeon) - Executioner
- OH (Swiftsteel Bludgeon) - Mongoose




Conclusion:

You can drop this PPM thing, and change it into Procs / Hit!!!

 

◊ Vulajin

Not likely. Based on your data, if you figure out on average how many of your attacks should have been performed by the weapon with Executioner on it, your Rising Tide had a proc rate of 3.41%, while the Swiftsteel Bludgeon had a proc rate of 2.48%.

If you'd like to help establish the proc rate of Executioner, you should do a test of this sort: take one main hand of one speed with Executioner on it, and a second main hand of different speed with Executioner on it, and swing a crapload of times with those weapons. You should also use a buff-canceling macro of the sort indicated in this post: [Rogue] Mechanics Testing Thread

Submitting that data would greatly help us establish a proc rate for this enchant.

 

Obould

Thanks for the post, using this scripts it would have been possible to tell a more accurate way on the % of the proc-rate.

But then again, check the UPTIME, which in the END is the thing what counts, a procrate of 50% is useless when the buff just lasts 1second.

Rising Tide / Swiftsteel: 28,97%
Swiftsteel / Swiftsteel: 37,98%

Why do you think it differs that much, if it has the same PPM regardless of weapon speed this two numbers should be the same.
Throwing in instant attacks you can get upto 50% with 2 x Bludgeons and 40% with Rising Tide / Bludgeon (only 14-15mins sample).

 

◊ Vulajin

Yes, the uptime counts in the end, but as theorycrafters, we can't accurately value the enchant in an arbitrary situation unless we know the mechanics that result in that uptime. Right now we know roughly how good the enchant is if you're auto-attacking with Rising Tide and Swiftsteel Bludgeon, or if you're auto-attacking with dual Swiftsteel Bludgeons. From this data we can't easily generalize to, say, a combat sword rogue using a full cycle with Talon of Azshara and Merciless Gladiator's Quickblade.

I'm not asking you to do any more work than you already have done. I wouldn't be making such a big deal over this if I had access to the enchant on my PTR character, I would surely go test it myself. Unfortunately, I instead have to sit here badgering people like you to do it. My apologies if I'm coming across as demanding or ungrateful, as that's not my intent.

 

paper

The dual mongoose test uptime is wrong, due to UptimeMeter incorrectly tracking proc uptime if done by two different sources. For example MH procs and 5 seconds before ending, OH procs it, giving you effective double proc for the next 5sec and another 10sec of pure offhand proc.

The thing is tho, that UptimeMeter will stop measuring it as "uptime" when MH proc fades, thus giving incorrect uptime values.

 

Carud

I too have done testing on executioner on PTR.

I dual wielded Wicked Edge of the Planes and Talon of Azshara (both 2.7 speed). Both weapons had executioner enchant. I had 17/44 build, so got flurry. 99% of hits were autoattacks. I used procwatch for tracking the procs.

I have this far just tested with executioner/executioner combo. But here's results this far:

10133 hits
892 procs
4.4 ppm

Due to flurry the ppm is higher then it would be with 2 weapons with flat 2.7 speed all the time.

Weapons were both the same speed and there's over 10k hits, so the amount flurry gave to each weapon should be quite even (those 10k hits took about 3h 20mins). So you could roughly count autoattacks give about 2.2ppm per enchant, with flurry beign up most of the time. With wf and instants the ppm would rise. Considering you cannot have 2 proccs on at the same time i'm atleast leaning to think that having double executioner enchants would not be wise, executioner/mongoose or executioner/potency would most likely be better. This for warriors ofc.

Next i will test ppm using a mainhand weapon and shield, but im guessing it will be around that 2.2ppm.

 

• Aldriana

4.4 PPM is significantly higher than I would have guessed based on the other data in this thread... are you using ProcWatch to record procs and, if so, is it possible that you're catching both gain and fade messages?

Also, do you by any chance have the combatlog for that test available?

 

Tamonten

So it is confirmed that Executioner does not stack like mongoose? That would be a pretty big consideration, as Armor Penetration seems to scale quite well in large amounts. That would be a pretty decisive factor for the executioner/potency debate which I have been having with myself...

 

Carud

Went to ptr and checked and sadly that's just what had happened. Procwatch was registering both gain and fade. So the 2.2ppm per enchant is roughly double of what it actualy is, making it 1.1ppm, which seems to be quite inline with other peoples tests. Also i wasent running any addon that saves combatlogs, so sadly dont have that.

Also to clear things for Tamonten:
Two executioner enchants dont stack in anyway in it's current form, the enchant simply renews it's duration timer if it procs during when it already it is procced.

 

paper

You can just type /combatlog next time you do this and WoW will log your combat messages by itself in /Logs directory.

 

borbes

Most raid bosses have 7685 base armor. You will want to build your calculations around that number.

 

Carud

Nice, thx for the info. I'll do that the next time i test. Every day you can learn something new.

 

Carud

I countinued my tests on ptr. This time learned from my mistakes, so i used two addons to track the amount of proccs (procmeter and procwatch) and i saved combatlog. So now the info was more inline with other peoples tests.

Testing was done using mainhand weapon (2.7 speed) with executioner and shield in offhand, so only mainhand weapon was doing the hitting. Specc was 17/44 so flurry was present. About 99% of the hits were autoattacks. So here's what i found out:

10542 hits
447 proccs (same on both addons and verified from combatlog)
23.5 hits/procc
1.1 ppm

Ok here's the bad news (which in a way is good news actualy). I noticed a thing with executioner enchant. Due to it works a bit diff then other enchants like crusader and mongoose for example that can stack, diffirently from those enchants executioner renews it's timer instead of stacking on proccs. So if you have executioner already procced and it procs again, then combatlog does not show the "You gain executioner." line, but the timer of the procc is renewed anyways (100% positive about this, witnessed it several times). Addons didn't ofc register that second proc either. This is due to a bug or then it's intended. It looks like finding out the real or close ppm will be almost impossible or really hard. On the positive side it means that the ppm is actualy higher then tests show!!!!

So getting 1.1ppm from autoattacks with just 1 executioner and having warrior flurry is quite good, as the ppm is actualy higher.

A way to checking for more accurate ppm would be to check the combatlogs for the time between the "gained" and "faded" messages and compare that amount to the duration of one proc, this way you could find out more accurate number of procs. It's just that the combatlog from those 10542hits i did is 810 pages long and contains thousands lines, so it atleast is out from my ability to parse. Ofc if someone has the means do to it, then pls do tell and ill happily upload the combatlog.

-Carud

 

• Aldriana

If you would post the combatlog somewhere so we can run some analysis on it, that would be very helpful. Thanks.

 

paper

Carud you can use the buffcancel macro, which cancels the buff as soon as you get it.

more: here

 

Carud

Ok uploaded the combatlog to filefront. Hope you can get more out of it.
WoWCombatLog.txt - FileFront.com

Heh ty again for good advice Paper, gotto try that one out.

 

Jairek

Eviscerate is not the best example, as stated, so it's best to include variations for Sinister Strike and Backstab. There are also other variables that will affect the value of Mongoose vs. Executioner, like the Relentless Earthstorm diamond.

All PvE combat builds from here on will likely include Aggression, so it should be included as a multiplier as well.

For Sinister Strike (20/41):
SS = wpn + 98 + ((AP / 14) * 2.4)
damage = SS * (1-armor_mitigation) * 1.06 * (2.3 * crit + (1-crit))

For Backstab (15/41/5):
BS = 1.5 * (wpn + ((AP/14) * 1.7)) + 255
damage = BS * (1-armor_mitigation) * 1.26 * (2.24 * (crit + .35) + (1 - (crit + .35)))


In the case of Mongoose + ToA / FoV, 1720 base AP, 24% crit rate:
armor_mitigation = .17179 or (2190 / (2190 + 10557.5))
SS = 260.5 + 98 + (((1720 + 120) / 14) * 2.4) = 673.93
ss_damage = 673.93 * .82821 * 1.06 * (2.3 * .27 + (1-.27)) = 799.31

BS = 1.5 * (252.5 + (((1720 + 120)/14) * 1.7)) + 255 = 968.89
bs_damage = 968.89 * .82821 * 1.26 * (2.24 * (.27 + .35) + (1 - (.27 + .35))) = 1788.40


In the case of Executioner + ToA / FoV, 1720 base AP, 24% crit rate:
armor_mitigation = .11337 or ( (2190-840) / (2190-840 + 10557.5) )
SS = 260.5 + 98 + 1720/14*2.4 = 653.36
ss_damage = 653.36 * .88663 * 1.06 * (2.3 * .24 + (1-.24)) = 805.63

BS = 1.5 * (252.5 + 1720/14*1.7) * + 255 = 947.04
bs_damage = 947.04 * .88663 * 1.26 * (2.24 * (.24 + .35) + (1 - (.24 + .35))) = 1832.22


So, by comparing:
SS w/ mongoose = 799.31
SS w/ executioner = 805.63
BS w/ mongoose = 1788.40
BS w/ executioner = 1541.22

The damage difference on Sinister Strike is fairly negligible, but is a bit more substantial for backstab.

When it comes to white damage, executioner should come out ahead. The rogue specials have far more multipliers when it comes to critical strike damage, which do not translate to white damage.

This also only applies to the damage of the two specials that occur within the duration of the buff itself. It also only applies to a boss like Vashj, and not a more heavily armored one like Void Reaver. It seems to me that Mongoose holds more "general" utility for a rogue, while Executioner will grant fair dps gains for encounters where a boss's armor is not incredibly large.

 

Asur

What's the bolded modifier, and why's it not the same in both?

 

Jairek

That would be a typo.

It looks like the "break even" point on armor, where executioner and Mongoose have the same effect on specials damage, is somewhere in the 10-10.5k range.