Machinator

I find it interesting that while our ratio of DW to MH procs was the same, you got a hell of a lot more procs than I did(+50% more!). In that SS you have 210 hit rating and I used 171, but 2.5% hit and your passive haste doesnt seem like enough to cover it. Perhaps its something with uptime monitor, which I got today and I noticed displayed an incorrect crit value as well, like your screenshot. Hopefully I will get some ptr testing done tonight.

"Information is ammunition."

 

koaschten

@Machinator you did read he was using styles, didn't you?

 

Machinator

I assume you mean special attacks. (correct me if I'm wrong I see Europeans call it styles, translation or something?) Yes, and I did too, which is why I dont understand how he got so much more than me. This is my WWS for DW Executioner: Berserkr - WWS
24 procs in 10m to his 77 in 23m. Perhaps I should just try again.

Anyway I loaded up the PTR and ran autoattack on a servant for a while. Besides parry I think its a good indication of proc rate, no flurry, no haste. Only 2.7 speed weapon with executioner and shield so no DW miss penalty.
136 procs in 135 minutes seems pretty close to 1 PPM to me.
WWS: Berserkr - WWS

"Information is ammunition."

 

suLonized

Mhh i dont know, but did you use a addon which cancels the buff? If not i know the mistake because WWS dont note the overwrites from the buff. So if the Buff is up and proccs again there will be the new buff which overwrites the old, but in WWS you will only see one buff proc.

 

Machinator

I double checked this in the logs earlier. If you already have the buff and you get another proc it cancels the first buff and applies the new one. It does not simply reset the buff to 15s.

12/5 02:17:21.623 Executioner fades from you.
12/5 02:17:21.623 You gain Executioner.

And since I forgot to bring procwatch to the PTR with me I also manually counted the procs in the log. There are 136 gains and 136 fades of executioner. :P

"Information is ammunition."

 

suLonized

Ah. I didn´t know that, mhh learnd something today but the problem is not solved.
Mhh youre just unlucky over 2 hours, problem solved lets have a picnic :/

 

mogun

The difference between us is the number of hits I think. I did 1576 hits (all attacks) in 23 mins, you had only 513 in 10 mins. Try to get a bigger sample (When I've done the test for exec in MH only, in the first 2 mins I had an uptime ~80%, it procced insanely much, but then stabilized around 50%). I'll try myself to do some more tests.

 

Vanadi

RapidShare: 1-Click Webhosting

4 hours of autoattacking a mob in diremaul with only my mainhand weapon equipped. Tooltip attack speed was 2.56 (1.71% passive haste) and no haste procs where used at all. From what I saw ingame the few times I was watching the combat log does not show if its refreshed and that happened quite frequently.

And this is the WWS report. Wow Web Stats

 

Machinator

I think the difference in hits is because you appear to OH Swiftsteel, 1.5, and I OH Fury, 2.6. The proc/min should have been closer. I havent had time to go out to blasted lands again but in a 7m VR I did 460 hits with ~68% uptime(24 procs), thats with windfury, bloodlust, drums, ect thrown in. So I still think 75% uptime by yourself is probably high, and a lot higher than what you should get on average.

@Vanadi: Unless your combat log works different from mine somehow there is a loss of executioner at the same time as a gain when its refreshed. Glancing at your log I dont see anywhere where it proced within 15s of the previous proc. Can you show an example where it refreshed and didnt show up in the log?

"Information is ammunition."

 

Vanadi

12/5 12:00:55.661 You gain Executioner.
12/5 12:00:58.292 You hit Gordok Spirit for 570.
12/5 12:01:00.665 You hit Gordok Spirit for 531.
12/5 12:01:03.314 You hit Gordok Spirit for 537.
12/5 12:01:06.224 You hit Gordok Spirit for 494.
12/5 12:01:08.202 You hit Gordok Spirit for 574.
12/5 12:01:10.811 You crit Gordok Spirit for 1193.
12/5 12:01:13.319 You hit Gordok Spirit for 494.
12/5 12:01:16.095 You hit Gordok Spirit for 534.
12/5 12:01:18.636 You hit Gordok Spirit for 512.
12/5 12:01:21.043 You hit Gordok Spirit for 517.
12/5 12:01:23.535 You hit Gordok Spirit for 513.
12/5 12:01:26.277 You crit Gordok Spirit for 1176.
12/5 12:01:28.685 You crit Gordok Spirit for 1049.
12/5 12:01:30.909 Executioner fades from you.

Thats a 35 second uptime of executioner.


12/5 12:06:18.130 You gain Executioner.
12/5 12:06:20.500 You crit Gordok Spirit for 1106.
12/5 12:06:23.137 You hit Gordok Spirit for 473.
12/5 12:06:25.800 You hit Gordok Spirit for 474.
12/5 12:06:28.155 You crit Gordok Spirit for 927.
12/5 12:06:30.533 You hit Gordok Spirit for 551.
12/5 12:06:33.164 You crit Gordok Spirit for 1074.
12/5 12:06:35.815 You crit Gordok Spirit for 1028.
12/5 12:06:38.518 You hit Gordok Spirit for 504.
12/5 12:06:40.901 You hit Gordok Spirit for 508.
12/5 12:06:40.901 Executioner fades from you.

22 second executioner uptime before it fades.

12/5 13:22:54.563 You gain Executioner.
12/5 13:22:56.961 You hit Gordok Spirit for 512.
12/5 13:22:59.611 You crit Gordok Spirit for 1010.
12/5 13:23:01.993 You hit Gordok Spirit for 573.
12/5 13:23:04.384 You hit Gordok Spirit for 467.
12/5 13:23:07.302 You hit Gordok Spirit for 502.
12/5 13:23:09.701 You crit Gordok Spirit for 972.
12/5 13:23:12.358 You hit Gordok Spirit for 475.
12/5 13:23:14.746 You hit Gordok Spirit for 500.
12/5 13:23:17.380 You hit Gordok Spirit for 544.
12/5 13:23:20.021 You crit Gordok Spirit for 1188.
12/5 13:23:22.424 Executioner fades from you.

28 seconds uptime before it fades.

These are just 3 examples I quickly found glancing through the log. My tests where done on live server btw. Maybe they changed the way the combat log records on the PTR?

 

Machinator

OK, went out and did a 40m test with double executioner on live server with my raid gear. I realize now that the reason mogun had a higher uptime than I expected is probably because servants are not +3 lvl Bosses but lvl 57. More hits, less dodge means more uptime.

WWS: Loading...
Combat Log: RapidShare: 1-Click Webhosting

40 minutes, 2193 hits, 99 procs, 62% uptime
Now it seems procwatch is only counting normal attacks, wws shows 3035 attacks with normal, BT, and WW so thats what we will use.

For simplicity assume both hands had same proc chance of .045(1 PPM).
Expected uptime = 1-(1-.045)^(15/~.79sec per hit) = 58% uptime

This was on live and I can confirm that my combat log shows refreshing the proc as a loss and gain at the same time.

12/6 06:14:48.655 You gain Executioner
...
12/6 06:14:51.123 Executioner fades from you.
12/6 06:14:51.123 You gain Executioner.

My guess is that for some reason the EU version does not do this as thats the only difference I can think of as the combat log should not be modified by addons, ect. Can anyone confirm this? Who knows what kinda stuff that has thrown off of true. Earlier there were some tests that got odd results and it would not surprise if this was the cause.

Conclusions I think we can make so far:
1. Executioner is 1 PPM
2. 1-(1-P)^(D/S) is a fairly accurate formula for determining uptime - It better be because thats what all the spreadsheets use for procs anyway
3. Something screwy is going on with wowcombatlog.txt - This is probably something worth reporting as a bug, but since they are revamping the whole thing anyway I don't see much of a point now. But it is something to look out for.

If those are true its time for the fun part, determining if all that uptime is useful. Im not sure how much of this is already started in other class specific threads though.

"Information is ammunition."

 

keshian

I'd suppose this comes down to if 120 agi + small haste buff for the uptime of OH mongoose does more dps than the armor reduction of (2xExecutioner -1xExecutioner)

 

mogun

I'm playing on EU and I see both loss and gain at the same time when Executioner refreshes itself. I'm using SimpleCombatLog, it shows something like:

--Executioner
++Executioner

And yea, I agree with the fact that my higher uptime could be because of lower lvl mobs.

 

Vanadi

Very odd, I tried simplecombatlog, even deleted my WTF folder and started with a clean ui and all addons except simplecombatlog disabled and I still didn't see the refresh. Might this be that it only shows the refresh if another weapon procs then the original proc comes from? E.g mh procs executioner and 5 seconds later offhand procs it and refreshes it.

 

Machinator

This could also be a possibility, but I havent gone through every proc checking if its exactly 15s yet. Anyone have any tools for doing that easily? I also was using simplecombatlog, but addons shouldnt be able the output of wowcombatlog.txt, unless Im wrong.

"Information is ammunition."

 

◊ Shaker

Yes - what Vanadi is saying is almost certainly correct - Crusader worked something like this - if the same hand proc'd twice, it would refresh the buff, if a different hand proc'd crusader, you'd get 'stacking' buffs (well, two of them).

In this case, it seems that they both refresh the buff, one just 'overwrites' the other.

From Machinator's Logs we have:

12/6 06:14:48.655 You gain Executioner. <-- one hand procs
12/6 06:14:51.123 Executioner fades from you. <-- other hand procs, 3s duration
12/6 06:14:51.123 You gain Executioner.
12/6 06:15:04.889 Executioner fades from you. <-- first hand procs again, 13.7s duration
12/6 06:15:04.889 You gain Executioner.
12/6 06:15:30.233 Executioner fades from you. <-- first hand proc'd again refreshing the buff somewhere in there

The last one is the interesting one - 25.334 second uptime. Several other incidents of it dropping before 15s are up and immediately coming up again (other hand overwriting).

 

◊ Shaker

Did analysis on Vanadi's log.

For # of procs, this is a minimum number - since the buff refreshes we don't know exactly how many procs were fired, but I counted every incidence of "You gain Executioner" then added 1 incident for every duration > 15s, another for every duration > 30s, and another for the sole incident of > 45s. This is a MINIMUM, though it is probably pretty damn close.

XLS file attached if someone wants to review the data

Attached Files

ExecAnalysis.xls (131.5 KB, 37 views)

 

Machinator

Nice catch. That is an annoyingly inconsistent way to log. Have to watch out for that but I'm guessing we have to wait for 2.4 for something to be done about it.

Edit: Do you have something to put that stuff in excel or was that by hand?

Something I have been thinking about.

If, and I need to confirm this, PPM is unaffected by haste/flurry, that is doubling the hits/min will not yield double the procs in one minute there would be a cap on average uptime. White hits are capped by PPM and there are only so many instants you can do in a set period of time. I base this on the PPM and Haste thread here.

Using 2x 2.7 speed(4.5% proc), BT(6s) and WW(9s) on cooldown. Lets say you have all the +hit and expertise needed to land all attacks. So in one minute you have 44.4 white attacks, 10 BT, 13.3 WW, 4.4 windfury for 72.1 attacks/min or an average speed of .83s. This would give you an uptime of ~56%.
The max MH uptime would be ~39% with the same setup.
This is roughly close to what I got, but mogun had more, not sure if I'm way off target or he is just lucky.

Using the assumption that white hits proc % is calculated off current, not base, weapon speed, adding haste would not increase uptime. Hit and expertise are already capped, so I dont see any other way to increase procs other than getting slower weapons.

"Information is ammunition."

 

◊ Shaker

That excel sheet was done by hand. Good ol' unix command line tools, then tossed into excel.

 

soujeh

This is a very interesting discussion. I am however a simple man, the only thing I care about is how well this enchant performs in PvP. I couldn't seem to find a mod that actually measured the exact amount of hits you did WITH EXECUTIONER UP. I found plenty that measured your % uptime, of course in time not hits, and ProcWatch that measures the amount of procs. Either way, I wrote my own mod and I've been gathering data the last week or so from 2v2 with S3 sword + Executioner to see what the average uptime is around in my usual 2v2-fights. The mod is really, really easy. All it does is on the event that you hit in way, it checks if you got Executioner up or not. So you get two variables, total hits and hits with Executioner.

Over the 3675 hits I got recorded so far, 1812 was with Executioner up. You don't have to be rocket scientist to see that it's around 50% uptime. ALL this data is recorded within 2v2, which I play with a resto druid in the top 10-tier in my battlegroup, so it should be a good estimate of the kind of uptime you could expect as a warrior in a real arena setting.

So the question becomes, how much better is 400-ish armor penetration (840 with 50% uptime) compared to the alternatives; 70 AP (Savagery) or 60-ish agility (Mongoose, from what I understood the uptime should be about the same). This is naturally with normal warrior PvP gear, AP around 1700, crit around 32, hit at 80-ish and some penetration. Let's say the average armor you are hitting at is around uhm 4k or whatever. How's that math on this?

 

Machinator

PvP is very hard to model because your target changes and burst can mean more than average damage. For reference my warlock has something around 2500 armor with mark which is around 20% reduction, probably the least armor of anything in arena. If you have executioner up it drops to ~14%. However against say a warrior with 11k armor(51%) you only get it down to ~49%. How the other enchants affect your personal dps in pvp is really only something you can say, Executioner is probably overhyped for Arena but its really only a call you can make for your gear and playstyle.

That said running around the 1500s in PvE gear with 1200 armor penetration up is fun against clothies. :P

"Information is ammunition."

 

Lithraviel

What am I going to do with this thread is over time establish the effectiveness of both the mongoose and executioner enchants for rogues. First will be the individual comparisons that will eventually include my own studied PPM, then when combining them.

As a rogue, it takes ~14 AP in order to increase our DPS by 1. That being said at AP of (see below) you get an increase to DPS of (see second number
AP: DPS:
1400 100
1500 107
1600 114
1700 121
1800 129
1900 136
2000 143
You can kinda see a pattern here, 100 AP = ~7 DPS increase, now with that being said we all know that mongoose proccing gives a 120 agility increase which is 3% to crit and ~9 DPS increase.
Now let us calculate the % of damage increased when mongoose procs
AP: DPS:
1400 9%
1500 8.4%
1600 7.8%
1700 7.4%
1800 7%
1900 6.6%
2000 6.3%
As you can see, mongoose gives diminishing returns the higher your AP gets.

Let us now look at executioner. What this gives you is 840 armor penetration. A fully debuffed boss will give you -4010 armor reduction (2600 for 5 sunders, 610 for faerie fire, 800 for Curse of Recklessness) I do not believe that we use CoR so our fully debuffed mob will be at -3210 armor. Here is the armor value for most high end bosses.

Serpentshrine Cavern:
Hydross the Unstable: 7700
The Lurker Below: 7700
Leotheras the Blind: 7700
Fathom-Lord Karathress 6200
Morogrim Tidewalker: 7700
Lady Vashj: 6200

Tempest Keep:
Void Reaver: 8800
High Astromancer Solarian: 6200
Al'ar: 7700
Kael'thas Sunstrider: 6200

Hyjal Summit:
Rage Winterchill: 6200
Anetheron: 6200
Kaz'rogal: 6200
Azgalor: 6200
Archimonde: 6200

Black Temple:
High Warlord Naj'entus: 7700
Supremus: 7700
Shade of Akama: 7700
Teron Gorefiend: 6200
Gurtogg Bloodboil: 7700
Reliquary of Souls:
- Essence of Suffering: 0
- Essence of Desire: 7700
- Essence of Anger: 7700
Mother Shahraz: 6200
Illidari Council:
- Gathios the Shatterer: 6200
Illidan Stormrage: 7700

As we can see, on average most bosses have either 6200/7700 armor, minus VR who is at 8800. What this means is that the mitigation done by a lvl 70 attacker is going to be R = (Ar)/(Ar + B)*100
where
R=damage reduction
Ar=armor
B=(467.5*Enemy Level - 22167.5) = 11960 for a 73 mob or 10557.5 for a 70 mob
So the damage reduction for a mob with x armor is equal to y reduction
X: Y:
6200 36.34%
7700 39.16%
8800 42.38%
So lets take out the -3210 armor that we should have from sunders and FF
X: Y:
2900 19.51%
4490 27.29%
5590 31.85%
NOW for the part that we've all been waiting for! the damage increase when executioner procs on a fully debuffed boss mob.
X: Y:
2060 14.69%
3650 23.38%
4750 28.42%
Thus, the damage increase for a mob with x starting armor will be y when fully debuffed and executioner proccing
X: Y:
6200 4.82%
7700 3.91%
8800 3.43%

As you can see, mongoose (at least with these preliminary numbers) does do more for a rogue when looking at damage % increase.

Over the next couple of weeks, I will be tracking the PPM, uptime and downtime of both my mongoose and executioner so expect more data as it develops.

I tested Executioner's procs using Procwatch mod. During the Fathom-Lord Karathress fight this recorded how often it procced per hit and every second of combat.
Recall:
803 Hits
27 Procs
11:11 of Combat time
which equates to ~29.7 Hits/Proc
total of 2.4 PPM which translates this:
Executioner on a boss fight, when consistently using SnD and attempting to hitting the target at all times, this enchant is up around 36 total seconds out of 60 seconds in a minute. These hits btw are counting my TOTAL hits, MH and OH
Trash, this proc is MUCH lower, around 1.3 or so.
My OH speed is at 1.15 with SnD up

Mongoose is currently only on my MH, with SnD up, i run 2.00 speed
On trash clearing to Fathom-Lord, here are the numbers
1451 Hits
58 Procs
21:21 Mins of Combat Time
which equates to ~25 Hits/Proc
total of 2.7 PPM which translates (on trash) to:
Mongoose on trash, when consistenly used with SnD, will run about 40.5 seconds out of a total of 60 seconds per minute.
On Fathom-Lord Karathress, here are the stats (remember that there is some running + wipes, etc etc STFU already!!!!)
692 Hits
27 Procs
10:13 Mins of Combat
which is equal to ~ 25.6 Hits/Proc
total of 2.6 PPM which translates (onto the boss) to this:
Mongoose on Fathom-Lord Karathress (remember that we did have to run around A LOT), when consistently used with SnD will run about 39 seconds out of a total of 60 seconds per minute.
One thing to remember about the proccing of mongoose is it is on my MH only, OH had executioner at the time I was testing.

Posts I used were:
Mitigation formulas:
Mitigation
http://forums.worldofwarcraft.com/th...74319573&sid=1
Armor Values:
[RAID] Boss armor values

 

• Latito

Small note.. that 14 AP => 1 DPS thing is "per attack". Your mainhand will get a 1 DPS increase from 14 AP, as will your OH and your instant attacks. This is then further modified by your crit%, armor-based damage reduction, and OH penalty for your OH attacks. Then theres rupture, which scales at a different rate.. and deadly poison which sees no increase from AP.

In general, a low-end rogue with entry Kara gear will be gaining ~27 DPS per 100 AP, a mid-geared rogue (SSC/TK) will be gaining about 30, and a high-end rogue (MH/BT) will be gaining ~33 DPS per 100 AP. The higher boost from AP is generally due to how AP scales with Hit, Crit, Armor Pen, Haste, etc. Basically AP scales with everything except itself.

 

Maurice2u

Latito beat me to the added complexity of AP. I'd also just briefly mention the potential added complexity of Executioner in that many are now using the Combat-Hemo hybrid spec, so Serrated Blades has an impact in armor reduction, as well as the fact there are about 5 other viable armor reduction items now in the game for rogues, and there is even the Night Blade dagger (though it and dagger spec were written off, in general, awhile ago now) that can further push the armor down. Basically, with all the armor pen out there now, a Rogue could seemingly get a boss quite close to a sub-500 level with respect to their attacks.

I am going to toy with it for a few minutes, but I'd love to see one of our better theorycrafters here explore all the items available (shadowpanther.net) for armor penetration being utilized and it's effect on DPS, given the trade off most likely of other stats.


EDIT: Just in quick glance, I noted about 1300 armor pen from decent (top 10 per item types) epics, and of course about another 1300 armor pen from the Night Blade dagger.

 

Lithraviel

When I calculated the DPS i was just looking straight at my AP and dividing it by DPS. I wasn't factoring in hit/crit nor do I want to attempt it, lol. Simple numbers man just looking to contribute a little, =D