Theck's MATLAB TPS analysis - WotLK/3.x

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Re: Theck's MATLAB TPS analysis

Postby theckhd » Wed Apr 14, 2010 7:47 am

Nadir wrote:Meaningless nitpick, but it's actually the Judgement of Light/Wisdom/Justice debuff rather than HotC which provides an additional Mote of Anger proc on Judgement.

Good to know, and makes sense given what we know about enchant proc triggers.
"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: 7972
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB TPS analysis

Postby theckhd » Wed Apr 14, 2010 8:04 am

steadypal wrote:2 heroic saurfang 25 attempts


279 hit rating, with tiny abom and 26 expertise

http://www.worldoflogs.com/reports/tjrdco5mtebg9401/


For Try 2, you got in 122 weapon swings (51 hits, 26 crits, 3 blocks, 36 glances, 6 parries) in 195 seconds, putting your hasted swing speed at 1.6.

Expected rate: 1.52 motes/sec

Try 1: 85 motes in 81 seconds, for an average of 1.05 motes/sec
Try 2: 213 motes in 195 seconds, or 1.09 motes/sec

Both seem a little low. However, they would be roughly consistent with the expectation if HV applications did not proc motes (in which case the expected is 0.998 motes/sec). Alternatively, I noted you weren't keeping up a perfect rotation (20 ShoR vs 15 HotR), probably due to Blood Beast spawns. So the variance could be due to that.

Just in case, if HV apps don't proc motes in the same fashion that they don't proc enchants, here's how the numbers work out: it would narrow the DPS variance with hit and expertise, such that the proc would account for a resonably reliable 132 DPS and 266 TPS regardless of your hit and expertise stats. So the trinket would give you around 240 DPS under ideal conditions (well under hit-cap, keeping perfect rotation on the boss).
"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: 7972
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB TPS analysis

Postby Marblehead » Wed Apr 14, 2010 8:44 am

It may be the judgement debuff that procs TAiaJ instead of HotC. I don't have a spec w/o HotC at the moment to check it, but if that's consistent with the enchant procs, then maybe I should say that I was wrong about that.

As for the HV applications, I hope the following 2 screenshots of the combat log will indicate that I was right.

Image

Image

On both of them, I was just auto-attacking the dummy with SoV. The first one is a triple proc before having 5 stacks of HV, while the second one is a triple proc with 5 stacks.
Last edited by Marblehead on Fri May 20, 2011 1:42 am, edited 1 time in total.
Image
Life is not difficult, people are.
User avatar
Marblehead
 
Posts: 202
Joined: Sun Aug 02, 2009 9:28 pm
Location: Bloodhoof (EU)

Re: Theck's MATLAB TPS analysis

Postby theckhd » Wed Apr 14, 2010 9:20 am

Marblehead wrote:On both of them, I was just auto-attacking the dummy with SoV. The first one is a triple proc before having 5 stacks of HV, while the second one is a triple proc with 5 stacks.

Yup, those certainly seem to indicate that HV applications can trigger it. That would suggest that the original estimates are right, and steadypal's data deviates mostly because of rotation adjustments.
"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: 7972
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB TPS analysis

Postby badgermonkey » Wed Apr 14, 2010 2:41 pm

http://wowwebstats.com/dtp5pvih2hlzu

Fights logged are Marrowgar, Saurfang, Festergut and Rotface (I was tanking boss on Rotface, so is probably the one with most consistent 969).


Was wearing the gear currently on my armoury: http://eu.wowarmory.com/character-sheet ... eedonboots


36 Expertise (164 Rating + Glyph)
239 Hit Rating (+7.29% chance to hit)
Image
badgermonkey
 
Posts: 329
Joined: Tue Apr 28, 2009 6:36 pm

Re: Theck's MATLAB TPS analysis

Postby Marblehead » Wed Apr 14, 2010 3:37 pm

One more log from a different log parser.

http://www.worldoflogs.com/reports/hy80ph4k8z9gzwzf/

I was wearing TAiaJ on every boss except Rotface (kiting) and all trash. On Valithria I was using SoC. Hit capped and 27 exp with glyph.
Image
Life is not difficult, people are.
User avatar
Marblehead
 
Posts: 202
Joined: Sun Aug 02, 2009 9:28 pm
Location: Bloodhoof (EU)

Re: Theck's MATLAB TPS analysis

Postby lythac » Wed Apr 21, 2010 5:48 am

Should a melee hit capped paladin substitute HoW every other Judgement? Looking at the 3.3 Paladin Ability Damage/DPS/Threat/TPS Breakdown it looks like it should if the Judgement damage listed includes the proc from JotJ?
Ryshad / Lythac of <Heretic> Nagrand-EU
User avatar
lythac
Moderator
 
Posts: 2685
Joined: Wed Sep 24, 2008 8:10 am

Re: Theck's MATLAB TPS analysis

Postby Boyfriend » Wed Apr 21, 2010 5:52 am

lythac wrote:Should a melee hit capped paladin substitute HoW every other Judgement? Looking at the 3.3 Paladin Ability Damage/DPS/Threat/TPS Breakdown it looks like it should if the Judgement damage listed includes the proc from JotJ?


Theck wrote:...which puts HoW in an odd place as far as substitutions. While it's a 6-second cooldown, we don't want to use it in place of our 6's. That means that we either substitute it for Consecration (if we're mana-starved) at a slight TPS loss, or substitute for Judgement or Holy Shield for a slight DPS gain if we want to go all-out DPS on a boss below 20%.
Image
User avatar
Boyfriend
Maintankadonor
 
Posts: 588
Joined: Wed Feb 24, 2010 8:16 am
Location: Sweden

Re: Theck's MATLAB TPS analysis

Postby lythac » Wed Apr 21, 2010 5:56 am

I failed -.-
Ryshad / Lythac of <Heretic> Nagrand-EU
User avatar
lythac
Moderator
 
Posts: 2685
Joined: Wed Sep 24, 2008 8:10 am

Re: Theck's MATLAB TPS analysis

Postby theckhd » Wed Apr 21, 2010 11:51 am

The section Boyfriend quoted is before including seal damage as well. If you look near the end of that same post:

Theck wrote:On the first graph, we see that the seal damage boosts Judgement up to be nearly as good as HoW for net damage per cast - with a little more expertise (STR?) on the gear set they would be tied. This means that as our gear progresses a little more, we will likely reach the point where HoW is only a valid substitution for Holy Shield, at least until cataclysm gives us new ranks to play with.

I didn't notice until now that I typed expertise there. I likely meant to type "STR" instead. Expertise doesn't make any sense (neither ability can be dodged or parried), nor does hit (because at that gear level the raw damage of HoW exceeds JoV+SoV). However, STR scaling will push Judgement clearly ahead.

JoV: 1.5*(0.22*SP + 0.14*AP) = 0.33*SP + 0.21*AP
SoV: weapon_base_speed*0.33*AP/14 = 0.0377*AP for a 1.6-speed weapon, around 0.06 for a 2.6-speed

HoW: 0.15*SP + 0.15*AP
"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: 7972
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB TPS analysis

Postby badgermonkey » Fri Apr 30, 2010 10:26 am

Do we have a threat read-out for TAiaJ yet? :P
Image
badgermonkey
 
Posts: 329
Joined: Tue Apr 28, 2009 6:36 pm

Re: Theck's MATLAB TPS analysis

Postby theckhd » Fri Apr 30, 2010 11:14 am

badgermonkey wrote:Do we have a threat read-out for TAiaJ yet? :P

I'll try and get around to looking at the parses you and marblehead provided soon, but I'm sorta swamped at work right now, so no guarantees on a time frame.
"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: 7972
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB TPS analysis

Postby badgermonkey » Fri Apr 30, 2010 7:00 pm

That's np. Who'd have thunk people have real lives!
Image
badgermonkey
 
Posts: 329
Joined: Tue Apr 28, 2009 6:36 pm

Re: Theck's MATLAB TPS analysis

Postby theckhd » Tue May 04, 2010 7:09 am

Marblehead wrote:One more log from a different log parser.

http://www.worldoflogs.com/reports/hy80ph4k8z9gzwzf/

I was wearing TAiaJ on every boss except Rotface (kiting) and all trash. On Valithria I was using SoC. Hit capped and 27 exp with glyph.

DBS:
3:16 (196 seconds)
122 melee swings (51 hit, 37 crit, 2 blocked, 29 glancing, 3 parry)
255 Mote generations

swing timer is thus 1.6066 (196 seconds / 122 swings)
p1=1 (hit-capped)
p2=1-.0725 (27 expertise)
Plugging those in gives me an expected mote generation rate of 1.5106
Observed value: 255/196 = 1.3010

Not bad, since that's an ideal value and assumes perfect rotation (which you didn't have, since you had 27 ShoR and 28 HotR in 196 seconds rather than 32 of each).

Here's an attempt at an exact calculation based on what you cast:
27 ShoR
28 HotR
19 Judgements
122 melee swings (and HV apps)
36 Manifest Anger procs (melee + HV app + SoV proc)

0.5*(4*19*p1 + 2*(27+28)*p2 + (27+28)*p2^2)/196 + 0.5*p2*(2+p2)*(122+36)/196 = 1.6588

Still higher than observed. If you ignore Manifest Anger, it comes out to around 1.4, but I thought someone said that the procs could trigger motes. Note that I'm not directly counting SoV apps here - you have 219 SoV hits, but only 190 successful melee+hotr+shor+judgements. However if you count MA procs, you have 225 expected SoV procs. Somehow you're either missing some procs, or something else is going on.

I'm pretty sure something's not right in the formula. Either one of the proc triggers we're counting on isn't right, or my math is wrong. Interesting note: if you ignore HV applications, the expected value comes out to 1.20 (closer, but a little low).

Here's the breakdown of that formula in text in case anyone wants to troubleshoot:

Code: Select all
(50% proc)*[(4 chances per Judgement)*(# Judgements)*(prob Judgement succeeds) +
            (2 chances per HotR/ShoR)*(# HotR/ShoR) *(prob HotR/ShoR succeeds) +
            (1 chance per HV app from H/S)*(# HotR/ShoR) * (prob HotR/ShoR and HV app both succeed)]/time
+ (50% proc)*[(chance of melee success)(2 procs per successful melee + chance at HV app success)]*(# melees)/time


Manifest Anger presumably acts exactly like a melee, which is why I added it in there.

Just one more check with your Festergut data:
237 seconds long
274 motes
138 melees (132 successful)
34 HotR (32 success)
35 ShoR (all succeeded)
24 Judgements (hit-capped)
39 Manifest Anger procs (38 success)

0.5*(4*24*p1 + 2*(34+35)*p2 + (34+35)*p2^2)/237 + 0.5*p2*(2+p2)*(138+39)/237= 1.60 expected mote generation rate

Observed rate: 274/237 = 1.16.

Again, lower than expected. Curiously, if you ignore HV procs, the expected rate is exactly 1.16. Marblehead's screenshots seem to indicate this isn't the case though, unless there's another effect in those screenshots we're not considering.
"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: 7972
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB TPS analysis

Postby theckhd » Tue May 04, 2010 7:10 am

badgermonkey wrote:http://wowwebstats.com/dtp5pvih2hlzu

Fights logged are Marrowgar, Saurfang, Festergut and Rotface (I was tanking boss on Rotface, so is probably the one with most consistent 969).


Was wearing the gear currently on my armoury: http://eu.wowarmory.com/character-sheet ... eedonboots


36 Expertise (164 Rating + Glyph)
239 Hit Rating (+7.29% chance to hit)

Sorry, this log expired before I could get to it. However, now that I've put the math out there pretty explicitly in the last post, you can repeat this yourself for any log you have available. I stuck to DBS and Fester just because they're pretty reliable, but you can use the "actual" formula for any fight since it doesn't rely on an expected rotation.
"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: 7972
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: daishan and 1 guest

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: daishan and 1 guest