Remove Advertisements

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

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Postby Varmin » Fri Mar 20, 2009 12:04 am

Am I reading it wrong or does Broken Promise with Judgement/SoR glyph out threat Last Laugh with Judgement/SoV?
Varmin
 
Posts: 172
Joined: Sun Jan 18, 2009 8:57 pm

Postby theckhd » Fri Mar 20, 2009 6:10 am

Nope, you're reading it correctly. There's a more detailed analysis in this thread.

With the 30% SoR glyph and 5/5 Reckoning, Broken Promise edges ahead slightly. However, if you don't have both of those conditions satisfied, there's not much of an advantage, except on bosses that can't parry/dodge.

It's not clear what this means, since we're not sure if the 30% glyph is still being implemented, and it's not clear from our talent trees yet whether 5/5 Reckoning will be a reasonable investment compared to other options.
"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: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby jere » Fri Mar 20, 2009 7:37 am

It might be important to note what the assumptions are for that conclusion in terms of expertise/hit values. As you increase expertise/hit, SoR does improve more than SoV (though I am not sure of the overall dps/tps effects).
User avatar
jere
 
Posts: 2960
Joined: Fri May 11, 2007 5:12 pm

Postby theckhd » Fri Mar 20, 2009 9:29 am

jere wrote:It might be important to note what the assumptions are for that conclusion in terms of expertise/hit values. As you increase expertise/hit, SoR does improve more than SoV (though I am not sure of the overall dps/tps effects).

None of these analyses are working near the melee hit cap or expertise soft cap. While SoR will get more benefit out of these stats, the relative change should be small compared to factors like weapon speed.

The Hit Rating and Expertise Rating values used in most of the simulations should be 129 and 62 respectively, since that's what's used in the default gear set.

By the way, I've updated the gear module to include the new T8 set stats. Since there was some discussion about how the default gear seat should be chosen, now would be a good time to discuss what people would like to see in the T8 default file.

Also, I didn't see you reply to my question in the 3.1 changes thread in General, so if I missed it I apologize for asking again - When you tested the AP scaling of Holy Shield on the PTR, did you test the base value and SP scaling as well, or just the AP scaling?

In other words, when I update the MATLAB code, am I safe to assume that the whole thing gets scaled by 1.3 exactly like it used to when talented, or did they just scale the AP contribution (a net nerf, in other words)?
"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: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby majiben » Fri Mar 20, 2009 9:42 am

The base was scaled up 30%. For a period of time the base was boosted by 30% but the scaling wasn't.

We might want to consider using 3 gear sets for 3.1. A threat set, an avoidance/EH set and a balanced set.
Amirya wrote:some bizarre lovechild of Hawking, Einstein, and Theck
User avatar
majiben
Moderator
 
Posts: 6999
Joined: Fri Aug 22, 2008 4:37 pm
Location: Retired

Postby Arees » Fri Mar 20, 2009 10:08 am

This may have been posted already, but I can't make heads or tails of these graphs. With the changes to SA, I was thinking of speccing like this.

I'm not picking up HotC because we have a ret pally that is always in the raid. Is 3/5 conviction and 2/3 crusade better for dps/tps than 5/5 SotP?

*edit*
also, in one rotation of the 969 rotation, what is the equivilent mp5 of 4/5 benediction? Maybe with that I could go 1/2 SA and put that last point back in crusade...
Image
User avatar
Arees
 
Posts: 546
Joined: Tue Dec 25, 2007 9:50 pm

Postby majiben » Fri Mar 20, 2009 10:12 am

Much better
Amirya wrote:some bizarre lovechild of Hawking, Einstein, and Theck
User avatar
majiben
Moderator
 
Posts: 6999
Joined: Fri Aug 22, 2008 4:37 pm
Location: Retired

Postby theckhd » Fri Mar 20, 2009 10:31 am

WATERBOYsh wrote:Is 3/5 conviction and 2/3 crusade better for dps/tps than 5/5 SotP?

Based on estimated values from the graphs, the 3 Conviction 2 Crusade spec should be 40-50 DPS and 80-100 TPS better than 5/5 SotP.

WATERBOYsh wrote:also, in one rotation of the 969 rotation, what is the equivilent mp5 of 4/5 benediction? Maybe with that I could go 1/2 SA and put that last point back in crusade...

From this post
Code: Select all
Skill  Mana  Eff. CD   m/s
ShoR   263      6       43,83
HotR   263      6       43,83
HS     439      9       48,78
Cons   966      9      107,3
Judg   219      9       24,3

devided them by their cooldown and added them toghter

268,04 mana lost per second

4/5 Benediction gives 8% reduction, so 268*0.08=21.45 mana/second, or roughly 107 MP5.
"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: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby jere » Fri Mar 20, 2009 11:04 am

theckhd wrote:Also, I didn't see you reply to my question in the 3.1 changes thread in General, so if I missed it I apologize for asking again - When you tested the AP scaling of Holy Shield on the PTR, did you test the base value and SP scaling as well, or just the AP scaling?

In other words, when I update the MATLAB code, am I safe to assume that the whole thing gets scaled by 1.3 exactly like it used to when talented, or did they just scale the AP contribution (a net nerf, in other words)?


Sorry, I totally missed your question before.

When I ran the original tests that caught the mistake, I first calculated what my HS values should be and then tested on the PTR to see if they matched. When they didn't, I did some controlled tests, where I took out my talents (except for the ones to actually get HS obviously), trying to keep anything that would multiply my damage out. I swapped in trinkets with various AP or SP values to see the effects on HS. That way, I could see in a controlled environment how only changes to AP or only changes to SP affected the results. I then talented normally again, and did the tests over and found the results the same. What I found originally was:

Base was increased by 30% (211*1.3 = 274)
SP coefficient was increased by 30% (0.91*1.3=.1183)
AP coefficient wasn't increased (stayed at 0.56)

This last build, I hopped on, unchanged (still talented) and did some tests. Holy Shield was providing the numbers I was expecting if *everything* was affected by 30% (base and coefficients). I swapped out some gear and did the calculation:

1.3*(211+0.056*AP+0.091*AP)*1.1*1.03

Which reflected my talents, and the values came out consistent with that equation (before, the AP coefficient wasn't getting multiplied like that equation would imply.

They now scale everything by 1.3x (so the finally got the AP coefficient up there with the base and SP coefficient)
User avatar
jere
 
Posts: 2960
Joined: Fri May 11, 2007 5:12 pm

Postby theckhd » Fri Mar 20, 2009 11:41 am

Thank you, the ability module should now correctly reflect 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: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby cordelia » Sun Mar 22, 2009 1:05 pm

theckhd wrote:TLDR version: If you're willing to give up Holy Shield uptime, you get far better returns by replacing Holy Shield with Exorcism than you do by trying to use seal-weaving to take advantage of the SoR glyph.


Thanks for running these. I'm going to ask you to run one more. Looking at all the analyses, If SoR or SoV are more than double HS TPS, then the following rotation should be our best non-HS tps rotation.

Scenario 5: HS-doublesub
Swap in SoR or SoV every other Holy Shield, keeping the rest of the base Exorcism HS-sub rotation identical. This results in DoT downtime, but may result in more pure TPS.

V6C6J6E6C6J6R6C6J6E6C6J6V6C6J6E6C6J6R6C6J6E6C6J6 ... V6C6J6E6C6J6R6C6J6E6C6J6

Calculations:
Consecration uptime: default 969
Judgement uptime: default 969
Exorcism uptime: default HS-sub
SoR uptime: 50%
SoV uptime: 50%
SoV DoT uptime: 80% (LL w/reck), 79% (LL w/o reck), 75% (BP w/reck), 73% (BP w/o reck)
HS uptime: 0%
LoO uptime: default 969

I would use the bolded SoV DoT uptime, as Last Laugh w/o reckoning is going to be the default situation, I think.

Math Details:
This technique requires calculating how long it will take for the DoT to get reapplied.

Assuming 10 expertise from gear/talents/racials and 8% melee hit, 11% spell hit, 6% base boss dodge, 16% base boss parry, the probability of applying a dot per swing, P(x) = 0.83*0.95 = 0.7885. In average, 5 DoT's should be applied @ 6.3 swings.

Using a 1.6 speed weapon, and 23% melee haste (typical raid buffs), 20% parry, swing speed will be ~1.2, resulting in 0.84 swings/min. Add HotR to this, and you get 1 swing/sec, or 6.3 seconds of downtime. Using a 2.5 speed weapon, swing speed is ~1.87, resulting in 0.70 swings/sec after HotR, or 9 seconds of downtime. Since this downtime also has partial DoT's running, I will decrease the downtime by 40% to simulate having partial damage running during the downtime: LL: 3.8 sec, BP: 5.4 sec.

With a swing speed of 1.2(LL) or 1.9(BP), the average re-application prior to SoR cast will be approximately: 0.6*0.7885 + 1.5*0.5*0.7885*(1-0.7885){HotR} + 1.8*(1-.7885-0.5*0.7885*(1-0.7885)) = 0.82 seconds. This means that SoV dot will fade an average of 14.2 seconds into SoR, resulting in 3.8 seconds of additional downtime per rotation. For BP, we have: 4.3 seconds of additional downtime.

This results in uptime numbers of: LL: 79%, and BP: 73%.

With 5/5 reckoning, and assuming re-apply time is halved with reckoning up, and LL@11% uptime, BP@19% uptime, we get LL: 80% DoT uptime, and BP: 75% uptime.
User avatar
cordelia
 
Posts: 630
Joined: Tue Oct 09, 2007 3:33 am

Re: Theck's MATLAB TPS analysis (A Jonesy derivative work)

Postby theckhd » Mon Mar 23, 2009 7:05 am

Ran the numbers for you. Note that all of these plots assume the Righteousness glyph is 10% rather than 30%, since I've seen conflicting reports about it's implementation lately. I'll comment on this at the end, though.

The bar chart (which uses Last Laugh) shows that while it's the best of the seal-swapping rotations, it's still slightly lagging the HS-substitution rotation.

The line plot vs. spec shows that this is true for any spec.
Image

However, swapping to Broken Promise improves things. The bar chart shows that the seal-swapping rotation beats the usual HS-sub in this case, thanks to SoR's scaling with weapon speed.

However, this is only true for the 5/5x/6 spec - once you drop SotP, there isn't any noticeable benefit over the Exorcism HS-sub:
Image

It's worth noting that if the Righteousness glyph does end up at 30%, the Last Laugh plots don't change significantly, but the Broken Promise one does - the "HS doublesub" rotation wins across the board by roughly 10 DPS in that scenario.
"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: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB TPS analysis (A Jonesy derivative work)

Postby jcamper1337 » Tue Mar 24, 2009 10:59 am

ok just a quick one

Judgement substitution is only better if both Judgement and SoV are glyphed, though it's always better than a 969 without the Judgement glyph.

with this being said. Its better to Exor then it is to judge even if both are glyphed. How. Dosent judgment have a better affect and more base damage? because If I remember correct Exor new glyph will increase the dmg by 20% but the Judgment one is only 10% and Ive never seen an Exor Hit for more then about 2-3k crit.

and from what I understand with the HS sub all you do is skip HS once then do the rotation and everytime Exor Is off cooldown weave it in insted of a HS. My only question is isnt that lowering you overall Dmg reduction because then you will not be blocking 100% of every incoming hit where that might cause some work for the healers

*EDIT* also can there be some more expelantion on Accuracy over potency. I didnt see anything in the post on it maybe I missed it
Image
jcamper1337
 
Posts: 121
Joined: Sun Aug 03, 2008 2:17 am
Location: Tanking stuff and trying Not To die

Re: Theck's MATLAB TPS analysis (A Jonesy derivative work)

Postby theckhd » Tue Mar 24, 2009 12:55 pm

jcamper1337 wrote:Judgement substitution is only better if both Judgement and SoV are glyphed, though it's always better than a 969 without the Judgement glyph.

with this being said. Its better to Exor then it is to judge even if both are glyphed. How. Dosent judgment have a better affect and more base damage? because If I remember correct Exor new glyph will increase the dmg by 20% but the Judgment one is only 10% and Ive never seen an Exor Hit for more then about 2-3k crit.

Er... which analysis are you referring to? Because both the updated rotation analysis and updated seal analysis never show Judgement substitution ahead of 969, for any glyph combination. At low gear levels, Exorcism should hit for more than Judgement, but Judgement scales better with STR, so at higher gear levels there's no reason to substitute it for any of our primary damage abilities.

jcamper1337 wrote:and from what I understand with the HS sub all you do is skip HS once then do the rotation and everytime Exor Is off cooldown weave it in insted of a HS. My only question is isnt that lowering you overall Dmg reduction because then you will not be blocking 100% of every incoming hit where that might cause some work for the healers

Yes, it is. This is why it's not suggested to use it while tanking Patchwerk, for example. If you're tanking a boss that doesn't melee though, or you're not actively tanking (i.e. you tank adds later in the fight, but lolprotdps for the early parts) then you could imagine using it to boost DPS.

jcamper1337 wrote:*EDIT* also can there be some more expelantion on Accuracy over potency. I didnt see anything in the post on it maybe I missed it

I'm not sure what more there is to say than was said here. I calculated the DPS and TPS you gain from each enchant, and found Accuracy is roughly 9 TPS better than Potency. However, Potency also gives a small amount of mitigation in the form of Block Value, so neither is clearly head-and-shoulders above the other.
"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: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB TPS analysis (A Jonesy derivative work)

Postby jcamper1337 » Tue Mar 24, 2009 1:51 pm

ahh i got ya now i musta skiped a page. I sitting in all gear at or above 211 item lvl so i see what u mean by the judgment doing less dmg then Exor with the lower gear. Ill actually have to start testing it my self now. Im currently my guilds main tank so this lolprot dps is the warriors job :-p lately ive been doing lazy tanking and only hitting holy shield on fights that are tougher and not all of them so i can weave a Exor into them now so i guess its bar rework time. thanks forthe help

*Edit* Im also a different spec then listed. Im 4/58/9 which seamed to work for me. We recently got a raiding ele shammy so im probally gonna drop my 3 points into heart of the crusader for something else. Not really sure. But i see how it scales better for the different specs with the seals now. Thanks this is great work
Image
jcamper1337
 
Posts: 121
Joined: Sun Aug 03, 2008 2:17 am
Location: Tanking stuff and trying Not To die

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: No registered users and 1 guest


Remove Advertisements

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
?php } else { ?