Remove Advertisements

A Call to Arms - MoP Mechanics Testing

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Re: A Call to Arms - MoP Mechanics Testing

Postby Sabindeus » Mon Mar 26, 2012 8:42 am

tlitp wrote:For the moment we're interested in #2


So where do I find 3 dummies close enough for HotR?
Image
Turn In, an NPC interaction automator - http://wow.curse.com/downloads/wow-addo ... rn-in.aspx
User avatar
Sabindeus
Moderator
 
Posts: 10470
Joined: Mon May 14, 2007 9:24 am

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Mon Mar 26, 2012 9:33 am

theckhd wrote:Mirawen's log. Will look at it in detail tomorrow morning.


Avoidance matrix:

Code: Select all
   hit/tick block glance dodge miss parry total
cens  90.80  0.00  0.00  0.00  9.20  0.00  250
SotR  43.33 10.00  0.00 20.00 10.00 16.67   30
as    25.00  0.00  0.00  0.00 75.00  0.00   44
hw    31.73  0.00  0.00  0.00 68.27  0.00  104
cs    34.68 12.10  0.00 16.13 11.29 25.81  124
j     23.60  0.00  0.00  0.00 76.40  0.00   89
cons  26.61  0.00  0.00  0.00 73.39  0.00  575
sot  100.00  0.00  0.00  0.00  0.00  0.00  254
melee  0.00  0.00 39.48 23.61 12.23 24.68  466


Again, limited info since this is a 93 boss rather than level+3. More confirmation that J/AS/HW/Cons are all spells. SotR and CS are the only things that can be parried, but the sample size is again too small to say anything accurate about the rate. As we acquire more samples, we'll be able to say something more definitive.

Looking closer, all of the Censure misses seem to be missed applications of the debuff. The log doesn't seem to be giving me entries for Censure refreshes, but the initial ramp-up confirms that AS and Cons not stack Censure. It seems reasonable to assume that J and HW do not either based on SoT triggers (see below). If we assume that only damaging melee attacks (CS, SotR, melee) stack Censure, then the observed miss rate is 23/258=8.91%. Large error bars on that, of course, but it's likely that Censure's application is still a melee attack that cannot be dodged or parried, and that Censure damage automatically succeeds.

Mechanical stuff:
-Seal of Truth and Censure damage seems to lag the attack by up to 1 second. Not sure if this is a logging issue or an actual in-game delay.

-Seal of Truth does 22-27 damage regardless of Censure stacks, even on the first hit. Damage is as expected for 0.14.*([42 79]+1089/14.*1.5), or 14% of un-normalized weapon damage. Was able to confirm that Judgement, AS, HW, Cons do not trigger it (and thus likely don't stack Censure). SotR, CS, melee attacks are all confirmed triggers.

-Each Censure stack adds 76 damage to the tick, which is exactly 1.6x what the tooltip formula (0.018*sp+0.035*ap) suggests. I don't know of an effect which would give exactly that modifier, so it could just be an old tooltip. We'll have to check/confirm the formula later, once we're farther along.

-Grand Crusader still grants HP on cast, independently of whether AS succeeds or fails.

Let's hold off on doing any more repeats of Test #1 until we can get true level+3 data.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7710
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Mon Mar 26, 2012 9:57 am

Sabindeus wrote:
tlitp wrote:For the moment we're interested in #2


So where do I find 3 dummies close enough for HotR?


The boss-level dummy in SW is part of a set of 3 iirc, arranged like (85) (boss) (85)

If you attack the 85 with HotR while standing on the side closest to the boss dummy, I think it should cleave to hit all 3. That would be great, because the boss dummy has a much higher chance of spell miss. So if the HammerNova attack is independently determined (i.e. has its own attack roll), it should be apparent rather quickly.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7710
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby Jaannaa » Mon Mar 26, 2012 4:47 pm

Test #13: SoR

Build: 5.0.1 (15464)
AP: 563
SP: 260
Weapon: Dalaran Axe ( 1.5 ) http://www.wowhead.com/item=44642
Other gear: None
Glyphs: None
Toon level: 85
Dummy level: Boss (attacking from the front), cleaved to one lvl 70 dummy (test done in Exodar)

WoL: http://www.worldoflogs.com/reports/uu03jwbmpubz2fer/
Jaannaa
 
Posts: 7
Joined: Thu Mar 22, 2012 12:46 pm

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Tue Mar 27, 2012 6:49 am

Jaannaa wrote:Test #13: SoR

Build: 5.0.1 (15464)
AP: 563
SP: 260
Weapon: Dalaran Axe ( 1.5 ) http://www.wowhead.com/item=44642
Other gear: None
Glyphs: None
Toon level: 85
Dummy level: Boss (attacking from the front), cleaved to one lvl 70 dummy (test done in Exodar)

WoL: http://www.worldoflogs.com/reports/uu03jwbmpubz2fer/


1124 SoR procs (and no misses) from 561 glances - oddly, we have 2 more SoR procs than we should (561*2=1122). At any rate, it seems that SoR automatically hits the primary target if the triggering attack succeeds, otherwise we should see significantly fewer procs (even if the log didn't record SoR misses explicity). Unfortunately, you're cleaving to a level 70 dummy, which you have 0% chance to miss (spell or melee), so we can't conclusively say whether the cleaves make a separate attack roll. It's very unlikely given the primary target behavior, however. To confirm, try meleeing the 70 dummy and cleaving to the boss-level. If you don't see a miss within a minute or so of attacking, SoR doesn't make separate attack rolls for any targets.

Damage consistently falls between 7 and 9, consistent with 5% of normalized weapon damage (=0.05.*([45 84]+ap./14.*2.4)). Crits for double damage. All crits occurred on the 70 dummy, at a rate of 19/561=3.39%. Do you happen to know what your melee and spell crit rates were in that gear set?
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7710
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Tue Mar 27, 2012 9:39 am

Reposting here so I can find it later.
theckhd wrote:Combining Lumenactio's list with Klaud's list, it looks like base mana is exactly 20k at level 85 (it's 23,422 in Cata).
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7710
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby Jaannaa » Tue Mar 27, 2012 1:38 pm

theckhd wrote:
Jaannaa wrote:Test #13: SoR

Build: 5.0.1 (15464)
AP: 563
SP: 260
Weapon: Dalaran Axe ( 1.5 ) http://www.wowhead.com/item=44642
Other gear: None
Glyphs: None
Toon level: 85
Dummy level: Boss (attacking from the front), cleaved to one lvl 70 dummy (test done in Exodar)

WoL: http://www.worldoflogs.com/reports/uu03jwbmpubz2fer/


1124 SoR procs (and no misses) from 561 glances - oddly, we have 2 more SoR procs than we should (561*2=1122). At any rate, it seems that SoR automatically hits the primary target if the triggering attack succeeds, otherwise we should see significantly fewer procs (even if the log didn't record SoR misses explicity). Unfortunately, you're cleaving to a level 70 dummy, which you have 0% chance to miss (spell or melee), so we can't conclusively say whether the cleaves make a separate attack roll. It's very unlikely given the primary target behavior, however. To confirm, try meleeing the 70 dummy and cleaving to the boss-level. If you don't see a miss within a minute or so of attacking, SoR doesn't make separate attack rolls for any targets.

Damage consistently falls between 7 and 9, consistent with 5% of normalized weapon damage (=0.05.*([45 84]+ap./14.*2.4)). Crits for double damage. All crits occurred on the 70 dummy, at a rate of 19/561=3.39%. Do you happen to know what your melee and spell crit rates were in that gear set?


From the in-game stats page: melee crit = 0.01% and spell crit = 0.16%
This is wearing nothing but the axe linked.

I tested the cleave-to-boss-dummy bit, no misses in about 3 minutes of cleaving.
Jaannaa
 
Posts: 7
Joined: Thu Mar 22, 2012 12:46 pm

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Wed Mar 28, 2012 5:57 am

Jaannaa wrote:From the in-game stats page: melee crit = 0.01% and spell crit = 0.16%
This is wearing nothing but the axe linked.

I tested the cleave-to-boss-dummy bit, no misses in about 3 minutes of cleaving.

Did you happen to log it? Not that I don't believe you, but there have been occasions in the past where combat log entries weren't being generated for missed attacks. The proof would be to see that you had exactly 2x as many seal procs as you did damaging attacks on the 70 dummy.

Crit numbers aren't conclusive, I guess. I think the formula for increased crit chance against a lower-level mob is still 0.2%*level_difference, which should be 3%. I'm not sure if that's for both spell and melee crit, though, and the 95% confidence interval for a 3% crit rate with a sample size of 561 is +/- 1.41%. So it's too close to determine whether it's using spell or melee crit.

Easiest way to test this is to collect data in full Holy gear. High intellect will significantly increase spell crit without affecting melee crit. Meleeing any dummy from behind should tell us whether it uses melee or spell crit, because melee crit will still be around 3% while spell crit should be much higher. I'll add a parse request for this.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7710
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby Jaannaa » Wed Mar 28, 2012 1:08 pm

I logged the cleave-to-boss test....

There is a discrepancy of 6 extra SoR counts. 4 of those were accidental critter kills.
I'm not sure why there's an extra 2 SoR counts.

Build: 5.0.1 (15464)
AP: 563
SP: 260
Weapon: Dalaran Axe ( 1.5 ) http://www.wowhead.com/item=44642
Other gear: None
Glyphs: None
Toon level: 85
Dummy level: level 60 as primary target, cleaved to 1 boss dummy (test done in Darnassus

WoL: http://www.worldoflogs.com/reports/70pxryvmt1r77nry/
Jaannaa
 
Posts: 7
Joined: Thu Mar 22, 2012 12:46 pm

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Thu Mar 29, 2012 6:34 am

519 attacks and 520 Seal procs on "Training Dummy"
520 seal procs on "Raider's Training Dummy"
4 seal procs on "Red-Tailed Chipmunk"

In both of your logs, there are 2 seal procs that precede the first logged melee attack. My guess is that it's not logging the first melee attack on the dummy for some reason - either you started the log while you were already auto-attacking, or there's some sort of combat log bug happening. There seems to be a one-second delay between a melee attack and the next seal proc; from your first log:
Code: Select all
[19:05:36.170] Jaannaa hits Raider's Training Dummy Parry
[19:05:37.672] Jaannaa hits Raider's Training Dummy 13
[19:05:38.557] Jaannaa Seal of Righteousness Training Dummy 1 (O: 8)
[19:05:38.557] Jaannaa Seal of Righteousness Raider's Training Dummy 8


So, at this point, my best guess is that the logging started in the one-second gap between a melee attack and the seal procs? I usually start logging while I'm out of range of the dummy and cast some signal spell (like re-casting my seal) to indicate "this is where the log starts."

In any event, your test confirms that SoR does not make any combat rolls independently - everything is contingent on the attack roll of the ability that procs SoR.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7710
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby Tobrexa » Thu Mar 29, 2012 2:23 pm

test

I tested SoJ and SoT in Retspec, at level 85 with dalaran axe (paper doll dmg range: 108-151, 86,29 dps) AP: 563 SP: 260 Build: 5.0.1 (15508).

Conditions: naked no selfbuffs, no Glyphs, no Talents,,(as mentioned) Retspec -> 4% physical damage debuff, doing a full rotation (I presumed Exo>CS>J>TV under 5HP, TV>rest at 5,) without ever using Inq. I adhered as much as humanly possible to this rotation while drinking beer and watching a movie.

As I was not really familiar with how the combat log distinguishes between fights, I have to add myself that I had 4 fight segments:

I started with SoJ in IF on a training dummy (from the front) in IF. As they come only lvl 80 version in IF, I switched to lvl 85 in SW, where I attacked from behind (melee dps reflex) which I corrected after a short period of time. Thus follows a longer segment of attacking from the front with SoJ. My test is concluded by attacks from the front on the same dummy with SoT.

After I uploaded the log, I became aware that those fight segments as well as everyone around me doing damage were merged into a single log, so I hope that one can salvage some data from it.

I hoped to cover some of request #1, get some more into SoT mechanics from Ret point of view and provide data for SoJ, which was not requested, but in my opinion still interesting from a empiric point of view.
Tobrexa
 
Posts: 1
Joined: Thu Mar 29, 2012 10:09 am

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Fri Mar 30, 2012 6:18 am

Looking at a subset of the log to isolate SoJ:
105 SoJ misses, 1620 hits. 6.5% miss rate. No dodges/parries. 6.5% is within the 1% confidence interval of the expected 6% spell miss rate, and with the lack of dodges/parries it seems safe to say that SoJ makes its own attack roll on the spell table.

It looks like only melee attacks proc SoJ, because I see instances where you melee and Exo/J the dummy simultaneously while only seeing one proc. Which seems consistent with the fact that there are 1726 Melee+CS+TV in the log, and 1725 SoJ procs. It's likely I cropped in such a way that I cut out one proc.

Damage is strange though. Unless I'm missing something, the observed damage range is [8 14]. That's weird for a few reasons:
0.16*([48 90]+563/14*1.5)=[17 24], too high.
0.075*([48 90]+563/14*1.5)=[8 11], lower bound correct, upper bound too low
0.075*([48 90]+563/14*1.5)=[10 14], lower bound too low, upper bound correct

In fact, I suspect the ability is bugged, because if you assume that there's no weapon-speed scaling, you get:
0.16*([48 90])=[7.68 14.4], which is exactly the observed range. So the ability isn't using normalized or un-normalized weapon damage, it's using base weapon damage. That would make it vastly inferior to just about anything, because it wouldn't scale with AP. Hence why I think it's bugged.

Isolating this section with only SoT procs:
Again, melee attacks seem to be the only thing triggering SoT: melee, CS, TV. Oddly, while there are 2160 successful CS/TV/melee, there are only 2141 SoT procs. I can't seem to find an explanation for this without combing the log line by line, which I don't have time for right now.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7710
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Re:

Postby Jaannaa » Sat Mar 31, 2012 9:03 am

Test 17
Gear: As listed. 3048 int, 0 crit rating. ( There's also 931 haste and 469 mastery)

Attacks: Just autoattack.
Build:5.0.1 (15508)
AP: 563
SP: 3202
Spell crit: 4.70%
Seal: Truth
Toon level: 85
Dummy level: 60

WoL: http://www.worldoflogs.com/reports/7vnegr6j1qolwg6p/
Jaannaa
 
Posts: 7
Joined: Thu Mar 22, 2012 12:46 pm

Re: Re:

Postby theckhd » Sat Mar 31, 2012 9:19 am

Jaannaa wrote:Test 17
Gear: As listed. 3048 int, 0 crit rating. ( There's also 931 haste and 469 mastery)

Attacks: Just autoattack.
Build:5.0.1 (15508)
AP: 563
SP: 3202
Spell crit: 4.70%
Seal: Truth
Toon level: 85
Dummy level: 60

WoL: http://www.worldoflogs.com/reports/7vnegr6j1qolwg6p/

800 melee attacks (766 hit, 34 crit, 4.44% crit rate)
800 seal procs (753 hit, 47 crit, 6.24% crit rate)

What was your melee crit chance? The 95% confidence interval for 4.7% spell crit is [3.23% 6.17%], so the observed SoT crit rate doesn't seem to match your spell crit. The fact that you're getting 4.44% crit on melee attacks is a problem too - we won't be able to tell the difference between 4.4% (melee crit) and 4.7% (spell crit) with a sample size of 800. We need spell crit to be ~10% higher than melee crit to see the difference.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7710
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby Fenris » Sat Mar 31, 2012 9:21 am

The boss-level dummy in SW is part of a set of 3 iirc, arranged like (85) (boss) (85)

Not sure if they changed it,but seems the only close enough group of >lvl70 dummy in stormwind is Boss-85-Boss now

Unless i'm blind :?
ImageImage
ImageImage
User avatar
Fenris
 
Posts: 608
Joined: Wed Oct 24, 2007 3:51 am
Location: Italy

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: Bing [Bot] and 1 guest


Remove Advertisements

Who is online

In total there are 2 users online :: 1 registered, 0 hidden and 1 guest (based on users active over the past 5 minutes)
Most users ever online was 380 on Tue Oct 14, 2008 6:28 pm

Users browsing this forum: Bing [Bot] and 1 guest