Theck's MATLAB thread - Cataclysm/4.x

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby theckhd » Mon Sep 20, 2010 12:26 pm

I should have a "calc_abilitydmg" module finished later today or early tomorrow. I'd like to get someone on the PTR and/or beta to do some testing in order to see whether the values seem accurate. If there are no volunteers, I'll take this request to the Cata/PTR threads and ask around.

I'd need the gear, talent spec, and buffs you used. It would be nice if you gave me your AP and SP as well so that I could check the code for consistency there as well. Uploading a log of a dummy test run should be fine, I just need the average damage of a "normal" hit for each ability. In fact, a dummy would be preferable given that you won't have Vengeance effects to worry about.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6205
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Mon Sep 20, 2010 1:19 pm

I should be able to do that this evening.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby theckhd » Mon Sep 20, 2010 1:27 pm

Teaser numbers:


Code: Select all
Spell     Raw    Net    Thr
ShoR    14386  13773  35452
CS       2743   2867   7380
JoT      4785   5082  13081
AS       3105   3297   8486
HW       2824   3262   8396
HoW      3329   4509  11606
Exor     2875   2909   7488
SoT       376    409   1053
SoR       338    338    870
SoJ       154    154    396
Cens     7957   8310  21390
Cons     2898   3065   7889
HotR      797    763   1964
HaNova    944    955   2458
Melee    1319   1198   3084


The "raw" should be for a non-crit, "net" is after avoidance and crit is taken into effect, and "threat" is just "net" times our threat modifier. Note that all of these values will vary slightly from gear set to gear set. The talent spec used is this. The simulation is assuming full raid buffs and full boss debuffs.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6205
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Mon Sep 20, 2010 7:24 pm

Attacks performed on a Grandmaster's (level 80) training dummy. I was watching for other debuffs to appear on my target and would ignore a value if something damage-affecting occurred even close to it. My talent build gave me Hammer and Shield but contained no talents for increasing the damage of anything. None of my glyphs were damage-affecting. This test is from the PTR as it was on Sep 20, 2010 (talking to posterity for a second).

The first data set is naked except for normal Neverending Winter (needed it to do AS/SotR) and a grey Hatchet (7.4dps, 2.50 speed, no stats).
Middle data set is my T9/early ICC/Frost gear, weapon was Shiver (178.9 DPS, 1.60 speed).
Third data set is the same as the middle except I put Blessing of Might on myself.

Format is xxx (yyy-zzz), x is the average, y is lowest hit observed, z is highest hit observed. Crits weren't counted but I did fast-head-math-check them for consistency. Consecration ticks can crit, I don't know if that's news. I didn't see any correlation between Hammer's direct and splash damage so I think the splash is a separate random amount.

Sample sizes were generally six. Some were three, for things with little variability like Judgement. For Crusader Strike I went to 11 samples; for Avenger's Shield I did ten. Middle gear set melee was based on 10 samples, third set melee based on 7 because I crashed at that point. More data pending.

Code: Select all
                              682AP/226SP             2890AP/889SP           3179AP/889SP
Melee, no seal                NOT TESTED YET          361 (333-392)          404 (358-447)
Avenger's Shield              3167 (2961-3334)        4188 (3999-4407)       4384 (4169-4543)
Judgement of Truth, 0 stk     146 (146-147)           601 (601-602)          642 (641-642)
Judgement of Truth, 5 stk     219 (219-220)           902 (902-902)          963 (963-963)
Censure, 1 stk                20 (20-20)              84 (84-84)             91 (91-91)
Censure, 5 stk                100 (100-100)           419 (419-420)          455 (454-455)
Hammer, splash                716 (651-845)           1010 (956-1075)        1011 (923-1076)
Hammer, direct                26 (25-27)              120 (110-131)          119 (112-125)
Crusader Strike               83 (81-86)              482 (412-522)          505 (463-544)
Shield, 3 holy power          824 (824-824)           3474 (3474-3474)       3821 (3821-3821)
Holy Wrath                    808 (808-808)           2688 (2687-2688)       2688 (2688-2688)
Consecration                  97 (97-97)              175 (175-175)          183 (183-183)


I can reproduce this gear and I've still got the testing spec in case anything else comes up. I still have the raw data from this if you want to check it.

(Does Seal of Truthgeance still 'burst' when it procs on a 5-stack? If so, I forgot that number until just now.)
Last edited by Arcand on Fri Sep 24, 2010 6:25 am, edited 4 times in total.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby trellian » Tue Sep 21, 2010 7:16 am

Just out of curiosity, shouldn't you consider melee as an ability aswell?
User avatar
trellian
 
Posts: 343
Joined: Wed Jul 16, 2008 3:02 am
Location: Silvermoon, EU

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Tue Sep 21, 2010 7:53 am

trellian wrote:Just out of curiosity, shouldn't you consider melee as an ability aswell?


I assumed its mechanics were unchanged and well understood, but I can get some numbers for it easily enough I guess.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby theckhd » Tue Sep 21, 2010 8:15 am

Thanks Arcand, I'll try and do some consistency checking with those today.

If you get a chance, I'm curious to know what damage ranges the in-game tooltips give for the abilities that aren't weapon-damage based.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6205
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Flex » Tue Sep 21, 2010 8:31 am

Arcand wrote:
trellian wrote:Just out of curiosity, shouldn't you consider melee as an ability aswell?


I assumed its mechanics were unchanged and well understood, but I can get some numbers for it easily enough I guess.


It'll confirm the 100% weapon damage CS and weapon damage portion of SoT at minimum.
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: 4677
Joined: Tue Jul 10, 2007 7:29 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Tue Sep 21, 2010 8:40 am

Flex - fair enough, I'll get some numbers.

Just tried to log into the PTR and crashed to desktop; going to reboot now and see if that was luck or skill.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Chicken » Tue Sep 21, 2010 8:49 am

Arcand wrote:Flex - fair enough, I'll get some numbers.

Just tried to log into the PTR and crashed to desktop; going to reboot now and see if that was luck or skill.
If it keeps happening, go to your PTR installation and remove the contents of the Cache and Data\Cache folders.
Image
User avatar
Chicken
 
Posts: 425
Joined: Fri Jun 26, 2009 2:19 pm

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Tue Sep 21, 2010 8:55 am

Thanks, I may have to try that. Just got in, was in long enough to get 34 swings at a dummy and then CTD again.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby tlitp » Tue Sep 21, 2010 8:59 am

Interested in seeing several data sets on the topic of avoidance DR, conveying both 80ptr and 85beta.
- pure dodge (i.e. no buffs, no gear with agility)
- pure agility (no buffs, no gear with dodge)
- parry
Only character panel info is needed at this time.

EDIT : also include a set of data conveying base values. No buffs, naked, untalented. Fetch base stats (mention race) and base dodge.
Last edited by tlitp on Tue Sep 21, 2010 9:03 am, edited 1 time in total.
User avatar
tlitp
 
Posts: 487
Joined: Mon Jul 27, 2009 3:25 pm

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Tue Sep 21, 2010 9:02 am

Avenger's Shield tooltip has no numbers in it yet.
Ditto Judgement.
HotR says (paraphrase) Hammer the current target for 31% weapon damage, causing a burst of light which hits all enemies within 8 yards for 973 holy damage. (Biggest splash I observed was 125...)
Shield says 20%/60%/120% of AP.
Holy Wrath says 2260 divided among the targets.
CS tooltip has no numbers.
Last edited by Arcand on Wed Sep 22, 2010 6:23 am, edited 1 time in total.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby theckhd » Tue Sep 21, 2010 9:31 am

Interesting data here. First of all, it looks like SoT scales with Censure as expected - roughly linearly, probably 3.2% per stack from 0 to 16%.

The interesting part is the damage of the on-swing portion. It's considerably higher than expected. With SotP, one would expect:

weapon_swing*0.16*1.15 = weapon_swing*0.184

Instead, Towelliee's parses show closer to 30% swing damage. My guess at this point is that SotP is additive instead of multiplicative, i.e.

weapon_swing*(0.16+0.15) = weapon_swing*0.31

To test this, I need someone to parse a dummy for me. Untalented, and then with 1 to 3 points in SotP. No other talents, gear is mostly irrelevant as long as it doesn't generate procs, no buffs.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6205
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Tue Sep 21, 2010 9:34 am

And you need big enough samples that a 5% difference isn't lost in the noise, I guess. Ok, will do.

Edit: Just did some vigorous failing with the combat logs - think I have a plan now but had to go respec and now I'm waiting for my hearthstone to cool off.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby theckhd » Tue Sep 21, 2010 10:38 am

That shouldn't be too hard though, I really just need to compare 0/3 to 3/3 and see how much of an increase we get. The other two sets (1/3 and 2/3) are just sanity checks.

Also, try and take note of Censure damage if you could. It scales with SP and AP, so performing the test naked (except for weapon) might be better for that. I think SotP might be affecting Censure based on Kysen's data in another thread, this test would confirm or refute that.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6205
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Tue Sep 21, 2010 11:59 am

Censure damage is definitely up as I add points to Seals of the Pure. I'm lame and don't know how to parse a log (went to worldoflogs.com and didn't see options to do anything but look at other peoples' raid logs) but here's the raw log data:

http://www.29tharmor.com/resources/Truth-0SotP.txt
http://www.29tharmor.com/resources/Truth-1SotP.txt
http://www.29tharmor.com/resources/Truth-3SotP.txt

In the first log I don't start doing anything for quite a while; I was taking gear off and confirming buffs and all that. For all three of these logs I had no buffs except Crusader Aura and Seal of Truthgeance. Attack Power 522, Spell Power 178, weapon was a Hatchet with 7.4 DPS (weirdly, it's 7.2 according to WoWhead but I double-checked) and 2.50 speed.

(I don't really want to hang the parsing work on you; if you can tell me how to do it or link to some idiot-proof instructions I'll do it. I just put the logs up in case parsing is easier than teaching others to parse.)

Update: I did the pencil-and-paper thing - Seals of the Pure affects Censure, without a doubt.

Without SotP, in my standard gear, Censure ticks for 419 every time unless it crits.
One point of SotP, ticks for 440.
Two points of SotP, ticks for 461.
Three points of SotP, ticks for 482.
Last edited by Arcand on Wed Sep 22, 2010 9:05 pm, edited 11 times in total.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Tue Sep 21, 2010 12:02 pm

tlitp: Wearing a shirt, tabard and a grey axe, my human 80 PTR guy has...

90 agility
2.38% dodge
5.00% parry
21.00% block

Kings brings the agility up to 94. That takes my dodge to 2.46% and, as expected, has no effect on parry or block.

Oh, and his health is 10271 naked with 233 stamina.
Kings makes it 10391 with 245 stamina.

Kinda weird - I'm used to gaining a couple, three thousand health when I cast Kings...
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby tlitp » Tue Sep 21, 2010 1:50 pm

Hm. I'd be interested in some (pure) agility scaling - at both 80/85. Base dodge was certainly lowered.
User avatar
tlitp
 
Posts: 487
Joined: Mon Jul 27, 2009 3:25 pm

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Tue Sep 21, 2010 2:34 pm

tlitp wrote:Hm. I'd be interested in some (pure) agility scaling - at both 80/85. Base dodge was certainly lowered.


I'll see what I can do with gemming. I've got plenty of socketed gear but as usual people are Striving! With! All! Their! Might! to get rich on the PTR, charging crazy prices for stuff...
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Bastien » Wed Sep 22, 2010 3:48 pm

Arcand wrote:Censure damage is definitely up as I add points to Seals of the Pure. I'm lame and don't know how to parse a log (went to worldoflogs.com and didn't see options to do anything but look at other peoples' raid logs) but here's the raw log data:

http://www.29tharmor.com/resources/Truth-0SotP.txt
http://www.29tharmor.com/resources/Truth-1SotP.txt
http://www.29tharmor.com/resources/Truth-3SotP.txt

In the first log I don't start doing anything for quite a while; I was taking gear off and confirming buffs and all that. For all three of these logs I had no buffs except Crusader Aura and Seal of Truthgeance. Attack Power 522, Spell Power 178, weapon was a Hatchet with 7.4 DPS (weirdly, it's 7.2 according to WoWhead but I double-checked) and 2.50 speed.

(I don't really want to hang the parsing work on you; if you can tell me how to do it or link to some idiot-proof instructions I'll do it. I just put the logs up in case parsing is easier than teaching others to parse.)



I don't speak MATLAB, but I do speak perl.

I just happened to be working on a very rudimentary log parser, so maybe this helps?

From Arcand's posted logs:
Truth - 0SotP
Code: Select all
Total Damage      Spell Name   Cast Count      Average Damage      Casters
      7826      Censure          110           71.15      Arcand
      7560      Melee            110           68.73      Arcand
      1912      Seal of Truth    108           17.70      Arcand
==================================================================


Truth - 1SotP
Code: Select all
Total Damage      Spell Name   Cast Count      Average Damage      Casters
      8369      Censure          118           70.92      Arcand
      7648      Melee            110           69.53      Arcand
      2088      Seal of Truth    108           19.33      Arcand
==================================================================


Truth - 3SotP
Code: Select all
Total Damage      Spell Name      Cast Count      Average Damage      Casters
      7250      Censure                89           81.46         Arcand
      6062      Melee                  89           68.11         Arcand
      1839      Seal of Truth          89           20.66         Arcand
==================================================================
Last edited by Bastien on Thu Sep 23, 2010 8:32 am, edited 1 time in total.
ImageImage
Image
User avatar
Bastien
 
Posts: 40
Joined: Thu Dec 18, 2008 10:05 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Wed Sep 22, 2010 7:14 pm

The 0 and 1 logs give identical numbers? That's too weird; I hope I didn't screw
up the naming or something. Are those two logs identical? Or is there any
chance you hit the same link twice, maybe?

Edit: They're not...and yet the total melee and Censure damage are exactly
the same? That's so improbable I can't even imagine, and it's also bizarre because
I'm positive the Censure numbers went up when I added points to Seals of the Pure.
Time for some more note-taking...


PS: Bastien, thanks for doing that.

PPS: Theck, see my first post on this page: sure as sure can be, Seals of the Pure increases Censure ticks.
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby theckhd » Thu Sep 23, 2010 6:43 am

A few notes:

1) I made a new WoL account and guild for hosting beta parses. The guild is named maintankadin and it's listed under US-Blackrock. Anyone who's willing to help out can make a WoL account and request to be added to the guild - I'll add everyone and give them upload rights.

Uploading a log is as easy as downloading the client, running it (it's Java), putting in your user/pass, and then opening the log file and uploading.

2) I uploaded arcand's original parses:
0/3: http://www.worldoflogs.com/reports/6jqa ... etails/13/
1/3: http://www.worldoflogs.com/reports/oyib ... details/8/
2/3: http://www.worldoflogs.com/reports/8b6c ... details/7/

Average Censure ticks were 75.9, 78.1, and 88.4 respectively. The 1/3 seems a little lower than it should be (should be around 79.7), but the 3/3 is 16.4% higher than 0/3, which supports Arcand's findings in his regular gear.

3) I'm fairly certain Bastien has a copy/paste error in his post. Every single number in the 1/3 section is identical to the 0/3 section.

4) Looking at the logs, you saw a bunch of misses and dodges. However, upon closer inspection, it looks like they're all failed re-applications. The DoT ticks never seem to miss or be dodged. So this behavior is consistent with what we know about SoV on live. I assume you didn't see any parries because you were attacking from behind.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6205
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Bastien » Thu Sep 23, 2010 8:36 am

theckhd wrote:3) I'm fairly certain Bastien has a copy/paste error in his post. Every single number in the 1/3 section is identical to the 0/3 section.


Looks like that was the case. I ran a diff on the 0SotP and the 1SotP files and they were identical. I thought that a little strange myself when I first ran it.

I downloaded the correct 1SotP file and re-ran it. I updated my last post with the new numbers, but they don't reflect/confirm Theck's findings from the WoL parse, so it's entirely possible my script is missing something. :-/
ImageImage
Image
User avatar
Bastien
 
Posts: 40
Joined: Thu Dec 18, 2008 10:05 am

Re: Theck's MATLAB thread - Cataclysm/4.x

Postby Arcand » Thu Sep 23, 2010 9:24 am

theckhd wrote:Average Censure ticks were 75.9, 78.1, and 88.4 respectively. The 1/3 seems a little lower than it should be (should be around 79.7), but the 3/3 is 16.4% higher than 0/3, which supports Arcand's findings in his regular gear.

I wasn't consistent about logging the ramp-up time; in fact, I'm pretty sure on one trial I mistyped /combatlog and didn't realize my mistake until a minute later, so we might have some runs containing ticks from 1-4 stacks while others don't. I was running each trial for about three minutes so hacking off the first 20% of each log might make the discrepancy go away.

3) I'm fairly certain Bastien has a copy/paste error in his post. Every single number in the 1/3 section is identical to the 0/3 section.

Same thing I thought, and it looks like he's made a correction now.

4) Looking at the logs, you saw a bunch of misses and dodges. However, upon closer inspection, it looks like they're all failed re-applications. The DoT ticks never seem to miss or be dodged. So this behavior is consistent with what we know about SoV on live. I assume you didn't see any parries because you were attacking from behind.

The way my threat feels right now on the PTR, I thought it most realistic to model my abilities with me attacking the bad guys from behind while they're busy killing somebody else. :roll:
Arcand
Moderator
 
Posts: 3108
Joined: Fri Mar 02, 2007 11:15 am

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

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

Who is online

In total there are 3 users online :: 2 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], Google [Bot] and 1 guest