A Call to Arms - Cataclysm Mechanics testing

[Kierly]

Here is a naked data set and a set with only spell power difference. If I find more time tonight or tomorrow ill get more sets.

http://www.worldoflogs.com/reports/02z031pau14uday0/details/0/
AP: 522
SP: 178

http://www.worldoflogs.com/reports/0md26i0i8b6wmmg1/details/0/
AP: 522
SP: 978

Weapon Speed: 1.4
Weapon Damage: 42-78

Target Dummy: 83 Heroic Training Dummy with 2 level 60 Expert Training dummies for holy wrath and consecration

Spec: All Prot damage talents minus HG with 3/3 crusade in Ret
Glyphs: Truth, Hammer of The Righteous, Shield of the Righteous, Focused Shield

Items:
Dalaran Sword
Small Shield

[Fetzie]

here is my log of last night's icc25 semi-pug.

http://www.worldoflogs.com/reports/q9grie7wuz5hri0s/

unfortunately I forgot to log marrowgar.

if you need the raw text file I can upload it.

[theckhd]

Here is a naked data set and a set with only spell power difference. If I find more time tonight or tomorrow ill get more sets.

http://www.worldoflogs.com/reports/02z031pau14uday0/details/0/
AP: 522
SP: 178

http://www.worldoflogs.com/reports/0md26i0i8b6wmmg1/details/0/
AP: 522
SP: 978

Weapon Speed: 1.4
Weapon Damage: 42-78

Target Dummy: 83 Heroic Training Dummy with 2 level 60 Expert Training dummies for holy wrath and consecration

Spec: All Prot damage talents minus HG with 3/3 crusade in Ret
Glyphs: Truth, Hammer of The Righteous, Shield of the Righteous, Focused Shield

Items:
Dalaran Sword
Small Shield

That log is very interesting. I think I have everything nailed down except for Holy Wrath. It seems that they've changed the base damage and scaling fairly significantly; it's now scaling by 50% of spellpower and the base damage is down to about 2200. I'd need another data point with higher spellpower to feel certain.

The only other updates that seem necessary are:
Censure: 0.035*AP + 0.016*SP per stack (up from 0.025 and 0.013)
JoTruth: 0.2*AP + 0.28*SP (up from 0.14 and 0.22 respectively)

[Kierly]

Here is log of just holy wrath at 542 AP, 3049 SP hitting 3 targets.
http://www.worldoflogs.com/reports/sviirx9fa915kzo9/details/2/

Looks like normal hits of 2303 and critical hits of 3455.

Might be useful hitting 2 targets it was doing normal hits of 2861 with critical hits of 4291.

[theckhd]

Thanks, that seems to nail down Holy Wrath. They didn't change the (bugged) formula, they just increased the coefficient from 0.3 to 0.39 and nerfed the base damage to 2156.

The formula now seems to be (2156+0.39*SP)/N+0.39*SP damage done to each target for N targets.

Now I just have to track down what's wrong with Judgement of Truth.

<edit>Just for my own sanity here:

Code: Select all

SP    JoT
178  372.5
978  901.7

(901.7-372.5)/(1.5*1.6)=220.5 more damage due to 800 spellpower, or 0.2756*SP. Probably 0.2755 after error is considered.

[theckhd]

Cross-post from the matlab thread:

Something funny is definitely going on with Judgement. Here's the data I have so far, from both here and the CtA thread:

Code: Select all

AP     SP    Dmg
522   178   372.5
522   978   901.7
3980 1216  2487.7*   corrected for Judgement glyph (2736.5 observed)
3822 1168  2473.6
3822 1168  2389.3*  corrected for Judgement glyph (2628.2 observed)

The fourth data set (Rabs' unglyphed set) looks a little suspect; a nonlinear fitting algorithm (nlinfit in MATLAB) gives coefficients of 0.17 and 0.295 for SP and AP respectively, which fits data sets 2, 3, and 5 almost perfectly, but results in a significant error for 1 and 4. My current suspicion is that data set 1 (Kierly's naked w/ DS set) has some error in it, which would explain why the simple hand calculations don't match the fit.

I'm checking in with Redcape at EJ to see what he's using for coefficients; in the meantime I'm hoping to get a little more data for JoT at various AP and SP values to nail down this discrepancy.

[Rabs]

• 50 judges.
• Stats: 1152 AP, 2160 SP, level 83 target dummy.
• Spec: http://wowtal.com/#k=-kfF-NSL.a6d.paladin .
• Glyphs: Truth, Shield of the Righteous

http://www.worldoflogs.com/reports/wb894h6bfc1srqss/

Are there any abilities (and/or at certain stat levels) in particular that you need in addition to this?

[theckhd]

....

OK, so strangely enough, that data set agrees with both 0.2*AP + 0.2755*SP and 0.175*AP+0.29*SP

Basically, I can make that data set agree with Kierly's data or with your other (glyphed) data sets, but not both simultaneously.

There has to be something else going on here that I'm missing - either a phantom talent effect, stray buff or debuff, or something.

Code: Select all

AP     SP    Dmg
522   178   372.5
522   978   901.7
1152 2160  1988.1
3980 1216  2487.7   corrected for Judgement glyph (2736.5 observed)
3822 1168  2473.6   doesn't fit with either model, some sort of error likely
3822 1168  2389.3   corrected for Judgement glyph (2628.2 observed)


[theckhd]

Are there any abilities (and/or at certain stat levels) in particular that you need in addition to this?

Do you think you could repeat Kierly's data sets (same spec you currently have, naked w/ Dalaran Sword and then equipping some INT gear to raise SP)?

I also could use some Hammer of Wrath data, but that's not very pressing since I'm fairly certain that formula is correct.

Thanks for all of your help.

[Rabs]

Do you think you could repeat Kierly's data sets (same spec you currently have, naked w/ Dalaran Sword and then equipping some INT gear to raise SP)?

Nakedlog:

• 50 judges.
• Stats: 522 AP, 178 SP, level 83 target dummy.
• Spec: http://wowtal.com/#k=-kfF-NSL.a6d.paladin .
• Glyphs: Truth, Shield of the Righteous
• Weapon: Dalaran Sword

http://www.worldoflogs.com/reports/279fzwqjqs8zi4y1/

Parse with higher spellpower in progress

Edit: Parse with some int-gear:

• ~50 judges.
• Stats: 522 AP, 1739 SP, level 83 target dummy.
• Spec: http://wowtal.com/#k=-kfF-NSL.a6d.paladin .
• Glyphs: Truth, Shield of the Righteous
• Weapon: Dalaran Sword

http://www.worldoflogs.com/reports/y2cso3cpcoi2iuwa/

Would a 0/0/0-specced parse help figure out if any talents (such as WotL) are behaving oddly?

[Kierly]

I think I may have found part of the problem, from your adjusted data for the glyph it appears you are expecting it to be 1.1*(unglyphed dmg). I did some tests and it appears to be around 1.0625*(unglyphed dmg).

All tests were down with same spec as before just swapping out the judgement glyph.
They can all be found under Oct 25 on http://www.worldoflogs.com/guilds/93724/

Code: Select all

AP   /   SP    UnGlyphed  Glyphed   ratio
522  /  178      352.6     374.7    1.0627
522  /  942      877.0     931.7    1.0623
1860 /  580     1213.0    1288.7    1.0624
1860 / 1344     1737.3    1846.0    1.0626


[theckhd]

Those give me a 0.286 coefficient for SP, consistent with the feedback I got from Redcape (0.2861). If I trust him on his AP coefficient (0.1823), I get pretty good agreement on all but three of the data sets.

With all of the data sets together, the nonlinear fitting algorithm is giving 0.1751 and 0.2898 for the two coefficients. Using those, here are the errors (the diff column is nonlinear fit minus observed, the Red column is the Redcape's model minus observed):

Code: Select all

  AP          SP         DMG         Diff     Red
522         178         373         -29      -19
522         978         902          -2        1
1152        2160        1988          -1        1
3980        1216        2488          30       91
3822        1168        2474         -55        3
3822        1168        2389          29       87
522         178         353         -10        0
522        1739        1425           5        0

My current guess is that something happened in those two high traces and Kierly's naked set, though who knows what it was. The incredible agreement with every other trace is enough to convince me that his coefficients are correct though.

[theckhd]

I think I may have found part of the problem, from your adjusted data for the glyph it appears you are expecting it to be 1.1*(unglyphed dmg). I did some tests and it appears to be around 1.0625*(unglyphed dmg).

What the hell.....

That means it's acting additively with WotL....

1.7/1.6=1.0625

Weird.

[theckhd]

I really, really, really need Hammer of Wrath data. That's the last thing I need to nail down before I can re-run all of the simulations. This should only depend on AP/SP, so weapons, talents, and glyphs are irrelevant.

Ideally I'd like to have two sets with the same AP but vastly different SP to nail down the SP scaling, and a couple more at whatever AP/SP values I can get my hand on. If anyone has time for this, please parse around 50 HoW's in as many different gear configs as you have time for.

[Klaudandus]

I'll try to start working on parses again, just been very out of it and sick -- problem is that I dont have any SP items, but I know of 3 I can get easily. Expect me to hit Grol later today.