A Call to Arms - Cataclysm Mechanics testing

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Re: A call to arms - Beta/PTR parses on WoL

Postby Klaudandus » Thu Nov 18, 2010 3:03 pm

Are the reports of CS being 115% Weapon Damage true?
The Element of Forum Hyperbole
Image
---
Flüttershy - Draenei Protection Paladin, Aerie Peak
Klaudandus - BE Protection Paladin, Feathermoon (Semi-retired)
User avatar
Klaudandus
 
Posts: 11146
Joined: Thu Apr 02, 2009 7:08 am
Location: Texas' Armpit

Re: A call to arms - Beta/PTR parses on WoL

Postby Flex » Thu Nov 18, 2010 3:07 pm

Klaudandus wrote:Are the reports of CS being 115% Weapon Damage true?


It is a known change from several patches ago.

they reduced the baseline CS damage to 115% weapon damage and increased Prot's CS from a 60% increase to a 100% increase.
We live in a society where people born on third base constantly try to steal second, yet we expect people born with two strikes against them to hit a homerun on the first pitch.
User avatar
Flex
 
Posts: 7500
Joined: Tue Jul 10, 2007 7:29 am

Re: A call to arms - Beta/PTR parses on WoL

Postby theckhd » Thu Nov 18, 2010 4:23 pm

Arianne wrote:Sorry? Did I upload the logs wrong? Is it still broken? Did I do a bad test? /confused :x I tried uploading the naked +/- glyph again to see but my WoL isn't loading after I upload to it (the ad servers are timing out for me for some reason), so I can't really check.


Engaging tlitp translation matrix:

Code: Select all
            Base   Base+Glyph   T10     T10+Glyph
Physical    90.9      97.4      119.0     130.4
Nova       864.7     856.8      966.4     950.2


The glyph is definitely not affecting the Nova portion, as you can see. The physical portion is seeing increases of 1.072 and 1.096 respectively in the two tests. For the first we'd expect to see 1.4/1.3=1.077 (provided the glyph is still additive with Crusade), which is pretty close. If the T10 bonus is working on the melee portion additively as well, we'd expect to see 1.6/1.5=1.066 in the second test, so something's funny there.

This test is incapable of telling whether T10 is affecting the nova component, since your AP/SP changed after putting on the T10 gear.
"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: 7956
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A call to arms - Beta/PTR parses on WoL

Postby theckhd » Thu Nov 18, 2010 5:57 pm

SoT / Avalanche testing, mostly for Redcape.

Stats:
STR: 2787
AGI: 99
STA: 5236
INT: 109
SPI: 112
AP: 5809
SP: 1771
Haste: 0
Hit: 8.15%
Expertise: 29
Weapon: Reclaimed Ashkandi, 3.8 speed, 1894-2842 damage, enchanted with Avalanche.
Paper doll damage: 3471-4419, 3.8 swing speed
Attacking a boss (88) dummy from behind, auto-attack only with SoT active, no self-buffs or raid-buffs. I'm doing this while AFK though, so there will be many outside boss debuffs I can't control for.

http://www.worldoflogs.com/reports/rt-b ... details/0/

I was planning on collecting more data, but apparently the server went down. That's about 2 hours though, which should be enough.

tlitp: Do you have some code already written to parse this and filter out procs due to melee attacks from procs due to Censure? Or do you use the expression editor of WoL?
"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: 7956
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A call to arms - Beta/PTR parses on WoL

Postby tlitp » Fri Nov 19, 2010 2:42 am

Theck : I'm reasonably sure that there are ~265 Avalanche procs that can be attributed to Censure's ticks. Maybe a few less or a few more (Censure overlapping autoattacks basically implies educated guessing), but that should be the range. The observed rate was ~9.5% (no ICD).
Either a flat 10%, either 1.5 PPM. We need a test with a much faster weapon, below 2.0 swing time.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A call to arms - Beta/PTR parses on WoL

Postby theckhd » Fri Nov 19, 2010 5:58 am

Interesting. Redcape and I had guessed it was 15% based on the wowhead proc details. I won't have time to collect another data set tonight, but I should be able to get a fast weapon parse this weekend.

<edit> I don't have time to browse the log file line by line (hence why I asked if you already had code), but a simple estimate seems to agree with your interpretation:

2207 melee+glances
2792 Censure ticks
982 Avalanche procs

at 5ppm for melee attacks, we'd expect 5*3.8/60*2207=698.9 procs from melee attacks, leaving 283 attributable to Censure. 283/2792=10.14%, in good agreement with your value.
"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: 7956
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A call to arms - Beta/PTR parses on WoL

Postby tlitp » Fri Nov 19, 2010 10:07 am

Yeah, I did check it against the known 5 PPM from autoattacks. Also worthy of notice : min / max. The minimum shows that there were no self auras/modifiers. The maximum, compared to the DB value, is a clear indicator that the proc does benefit from external auras/modifiers (~1.078, matching CoE well).
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A call to arms - Beta/PTR parses on WoL

Postby theckhd » Sat Nov 20, 2010 6:14 pm

SoT / Avalanche testing, round 2

Stats:
STR: 2646
AGI: 99
STA: 5228
INT: 255
SPI: 209
AP: 5527
SP: 3786
Haste: 0
Hit: 8.31%
Expertise: 29
Weapon: Andoros, Fist of the Dragon King, 2-242 damage, 2.3 speed
Paper doll damage: 910-1150, 2.3 swing speed
Attacking a boss (88) dummy from behind, auto-attack only with SoT active, no self-buffs or raid-buffs. I'm doing this while AFK though, so there will be many outside boss debuffs I can't control for.

Parse: http://www.worldoflogs.com/reports/rt-7 ... 41&e=10735

Note that you have to look at a portion of the time range, after a certain point my weapon apparently broke and Avalanche stopped proccing.

4653 melees
3542 Censure ticks
1203 Avalanche procs

Expected procs from melee: 4653*5*2.3/60 = 892

Remaining 311 procs due to 3542 Censure ticks would be an 8.78% proc rate.
"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: 7956
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A call to arms - Beta/PTR parses on WoL

Postby theckhd » Sat Nov 20, 2010 10:47 pm

More Avalanche testing:

http://www.worldoflogs.com/reports/rt-a ... details/4/

Didn't write down detailed stats, but it's most of the holy gear that comes on a premade. Just cast Exorcism over and over to get proc rate estimate from spells.

Weapon: Andoros, Fist of the Dragon King, 2-242 damage, 2.3 speed, Avalanche
SP: 5621
Spec: 21 points in prot.
556 Haste rating (4.34%)

521 Exorcisms
51 Avalanche procs

9.79% proc rate.
"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: 7956
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A call to arms - Beta/PTR parses on WoL

Postby tlitp » Sun Nov 21, 2010 1:01 am

In the case of spells/DoT, Avalanche clearly isn't PPM-based. The DB value is proven incorrect, as it's actually a flat 10% chance. I think we need one more test before we call it done - check for AoE scenarios. Grab a fast weapon (minimizes the number of procs from the physical component), then cast ~200 HotR on a group of well-stacked low-lvl dummies. These ten minutes of testing should be enough to determine the per-target behaviour.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A call to arms - Beta/PTR parses on WoL

Postby theckhd » Sun Nov 21, 2010 1:15 am

Yeah, I should be able to find some time tomorrow to run a Dalaran Sword test.
"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: 7956
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A call to arms - Beta/PTR parses on WoL

Postby Rasmfrackn » Mon Nov 22, 2010 4:28 pm

Hello, diligent number crunchers.

Something just occurred to me, and it strikes me as something of reasonable consideration for gearing. Have you, or will you at some point once coefficients are nailed down, show threat stat weights at different levels of vengeance?

Obviously I think we all dislike vengeance as it is, and it strikes me that it makes Str from gear sort of tertiary in concern. Against a target dummy, Str is probably #1 after getting to the hit/exp caps, but when I'm rocking 300% normal AP, I have a feeling it may take a back seat to say crit.

This is mostly for the occasional agi piece. Obviously agi is not as strong as dodge anymore, especially without the armor boost, but for example if I'm working with a mastery trinket like say unsolvable riddle vs impatience of youth, the ~1200 dodge rating worth of agi is much more appealing than the ~400 parry rating of str for survival, and the crit of the agi may outstrip the AP of the str if I'm assuming a lot of vengeance already.

Or more generally, I may want to go for agi jewelry that's got mastery + hit/exp as its stats to finish capping and go for pure tanky pieces in my plate slots.

Now, this may be totally wrong and we may scale in such a way that more AP is always better than crit, but hopefully you can tell us. Good luck representing all this information in a graspable way. :) 3d contours anyone?
Rasmfrackn
Dwarf Paladin
Icecrown Server
Eng/Scribe/Masochist
User avatar
Rasmfrackn
 
Posts: 1014
Joined: Fri Aug 17, 2007 12:11 pm

Re: A call to arms - Beta/PTR parses on WoL

Postby theckhd » Mon Nov 22, 2010 5:55 pm

Rasmfrackn wrote:Something just occurred to me, and it strikes me as something of reasonable consideration for gearing. Have you, or will you at some point once coefficients are nailed down, show threat stat weights at different levels of vengeance?

I've been calculating the stat weights with nearly a full stack of Vengeance (95% iirc) already. I plan to continue doing so, though I could certainly include some commentary on what happens when you drop to 50% or 0% Vengeance.
"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: 7956
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A call to arms - Beta/PTR parses on WoL

Postby tlitp » Mon Nov 22, 2010 11:01 pm

theckhd wrote:
Rasmfrackn wrote:Have you, or will you at some point once coefficients are nailed down, show threat stat weights at different levels of vengeance?

I've been calculating the stat weights with nearly a full stack of Vengeance (95% iirc) already. I plan to continue doing so, though I could certainly include some commentary on what happens when you drop to 50% or 0% Vengeance.

Vengeance is an input for [EM] since r174. Them Rogues be pesky, and now with 100% more Anticipation. :P

Grr, beta is gone. I'll post here the latest DBC paladin_spell_scale, as we'll need it for checking the @85 values against @80 tests : 1.115548762542.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A call to arms - Beta/PTR parses on WoL

Postby tlitp » Wed Nov 24, 2010 10:03 am

Right, time to move on to live testing.
I. I'd like to see 55283 in action. Grab the level 80 meta, activate it, and then proceed to smack down just about anything that hits for 1000+ (PvP against a 2H-based spec, dungeons/raids).
II. Scaling tests (I'll emphasize what Theck has already pointed out : be as verbose as possible, there are a lot of variables - we need to fully characterize the testing environment).

In order to avoid confusion : in terms of priorities, II >> I. However, I is considerably less time-consuming.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: No registered users and 1 guest

Who is online

In total there is 1 user online :: 0 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: No registered users and 1 guest