Remove Advertisements

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

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Postby theckhd » Fri Mar 13, 2009 8:18 am

Recalculated these just to see if anything had changed significantly. Since the last time, the gear module has been updated to be more realistic, Libram of Obstruction uptime has been added, and several other things have changed. <edit> Armor Penetration has also been added. This is all with the 5/60/6 default build, 969, JS glyphs, etc.

The graphs don't look too strikingly dissimilar though. Expertise seems a little stronger than before for DPS, but is still weak for white damage. AP snuck ahead of BV for dps by a hair, but is still behind for TPS. Armor Penetration surprised me a bit, being better than Crit/AGI for DPS, but lagging behind Expertise. I was expecting it to be far worse, but keep in mind this isn't including Sunder/FF yet, so it might be over-valuing it. It's worse than Crit/AGI for TPS either way though.

Also keep in mind that these graphs are plotted vs. unbuffed character sheet STR, so the average end-game character should be between 1200 and 1400 on these plots.
Image

Image
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Call to Arms 5.x, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7658
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby Meloree » Fri Mar 13, 2009 11:19 am

This is offtopic, maybe, but all of this stuff is amazing, I'd just like to extend my thanks. This thread is making me a better tank.

--
Meloree, <Edge>, Garona
Meloree
Maintankadonor
 
Posts: 1420
Joined: Wed Mar 12, 2008 10:15 am

Postby Arees » Fri Mar 13, 2009 2:15 pm

Meloree wrote:This is offtopic, maybe, but all of this stuff is amazing, I'd just like to extend my thanks. This thread is making me a better tank.

--
Meloree, <Edge>, Garona


I would like to say thanks too. I'm good at the math required for programming and thinking about computational complexities, but my knowledge of statistics and how to model stuff like that is pretty slim.

I don't understand what most of the graphs are representing though or what they mean. I just see a bunch of lines with acronyms attached, and wait for someone to turn the results into english :)

yeah... anyway, keep up the awesome work.
Image
User avatar
Arees
 
Posts: 546
Joined: Tue Dec 25, 2007 9:50 pm

Postby theckhd » Sat Mar 14, 2009 9:08 am

Since the Seal of Righteousness glyph has been boosted to 30% on the PTR, there was some question as to whether this would change the results of the seal comparison. The answer is involved enough that it would take a considerable re-work of that post, which might make the resulting discussion confusing, so instead I'm going to just repost the pertinent stuff.

Also worth noting that someone pointed out that I was using 22% weapon damage instead of 27% for SoB, so I fixed that, and this analysis includes LoO and all the more recent changes.


First of all, let's look at the numbers for reckoning again:

Reckoning

First, the base numbers:
Code: Select all
Seal  DPS   DPS(R)   Diff  TPS    TPS(R)  Diff
SoV   3107   3167     60   7792   7880     88
SoR   3052   3137     85   7677   7836    159
SoB   3025   3105     80   7565   7706    141

Here we notice a subtle difference - SoR with Reckoning now beats out SoV without reckoning. This is a change from the previous calculation.

This distinction is important, because while doing the re-calculation I realized that in the previous calculation, I was using 5/5 Reckoning for the SoR/SoB calcluations, and 0/5 Reckoning for the SoV calculations - in other words, a "worst-case scenario" for SoV. I did this because SoV was consistently winning anyway.

With the boost to the SoR glyph, however, this pushed SoR ahead of SoV, but only by a hair. The difference was less than the effect of Reckoning on white swings alone, which made the comparison unfair. In other words, it was OK that we were comparing apples to oranges before, because the apples were beating the oranges even with the heavy handicap - now that they were very close, we have to be much more careful.

Next, let's look at the scaling of all three seals with AP and STR:

Image
Image
As you can see, SoV still scales the best with AP and STR. The order is always SoV > SoR > SoB.


Rotation
Since we're being careful, let's look at the rotation plots again with 5/5 and 0/5 Reckoning.

5/5 Reckoning:
Image
threat graph
Again, with 5/5 Reckoning, SoV is still ahead of SoR, even when optimally glyphed. Just like before, the seal glyph always out-performs the Judgement glyph.

0/5 Reckoning:
Image
threat graph
We see that SoR falls behind even further without Reckoning, as expected. SoR DPS really hinges on getting as many procs as possible, so it suffers without Reckoning.

It's worth reiterating here that if you look at the numbers on these plots, 969 with Vengeance and 0/5 Reckoning performs worse than Righteousness with 5/5 Reckoning, purely due to the boosted white swing damage. However, if you're putting 5/5 in reckoning, you're still better off using JoV.

Talents
Since these graphs will all look pretty much the same, we'll just analyze the 969 graph. I'll link to the other ones, but there's really not much extra to be learned from them. All of these plots assume 5/5 Reckoning

969:
Image

Holy Shield substitution, HSs w/ 3 glpyhs
Exorcism on Cooldown, EoC w/ 3 glpyhs
Judgement Replacement, Judgement Replacement w/ 3 glyphs

The plot clearly shows that SoV tops SoR by around 40-50 dps, straight across the board. The old form of the glyph (10%) would drop it down roughly 50-60 DPS across the board, placing it tied or behind SoB for all specs except 5/60/6.


TLDR Version
SoV is still superior to SoR and SoB, even with the new glyph, even with reckoning, no matter what. You should be tanking with SoV.
Last edited by theckhd on Sat Mar 14, 2009 11:09 am, edited 1 time in total.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Call to Arms 5.x, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7658
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby jere » Sat Mar 14, 2009 9:44 am

Just a silly question probably, but when you are doing the comparisons based on the SoV glyph, are you using different expertise values (different by 10 expertise) for the SoV trial versus the SoR Trial?

I am sure you are, but was just checking.
User avatar
jere
 
Posts: 2951
Joined: Fri May 11, 2007 5:12 pm

Postby theckhd » Sat Mar 14, 2009 9:56 am

yup:

Code: Select all
expertise=(Base_exp + GoSoV + (Gear_Exp+extra_exp)./exp_rating_to_exp) + ComExp;


GoSoV is 10 when glyphed and 0 when unglyphed, and as you can see from the graph legends, I don't use the Vengeance glyph for SoR or SoB.

I didn't explicitly state it in this post (though I did in the previous version) for the AP and STR scaling graphs, but those use Judgement/Vengeance for SoV and Judgement/Righteousness for SoR/SoB.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Call to Arms 5.x, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7658
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby halabar » Sat Mar 14, 2009 10:00 am

What exactly is the Holy Sheild substitution?... how does that differ from the 969?
Amirya wrote:... because everyone needs a Catagonskin rug.

twinkfist wrote:i feel bad for the Mogu...having to deal with alcoholic bears.
User avatar
halabar
 
Posts: 9343
Joined: Fri Jun 08, 2007 8:21 am
Location: <in the guild that shall not be named>

Postby theckhd » Sat Mar 14, 2009 10:20 am

It just replaces Holy Shield with Exorcism every other cast. This results in a considerable amount of Holy Shield downtime, but there are some bosses that don't melee, and Holy Shield is by far our weakest damage/threat component.

By the way, I cranked out the scaling of the seals with Expertise to find out when the crossover point occurs. Results here. Looks to be at about 185 expertise rating from gear.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Call to Arms 5.x, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7658
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby Jonesy » Sat Mar 14, 2009 11:03 am

I just updated my spreadsheet and I have SoV doing more DPS and SoR doing more TPS. Double-check for me please?

(I haven't modelled Reckoning yet, though.)
Jonesy
 
Posts: 418
Joined: Fri Jul 13, 2007 1:23 am

Postby theckhd » Sat Mar 14, 2009 12:43 pm

Jonesy wrote:I just updated my spreadsheet and I have SoV doing more DPS and SoR doing more TPS. Double-check for me please?

(I haven't modelled Reckoning yet, though.)

I haven't found any situation where that's the case. But my default gear set is different from the one you started with. Since everything is coming out pretty close in these comparisons, it's possible that you've found a gear set that gives that sort of result.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Call to Arms 5.x, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7658
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby majiben » Sat Mar 14, 2009 12:51 pm

I don't remember if this has been asked and answered before but are you modeling partial resists? Also does the SoV vs SoR modeling also take into account the judements? Also it seems the SoR glyph only affects SoR damage and not it's judgements.
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 » Sat Mar 14, 2009 2:18 pm

Majiben wrote:I don't remember if this has been asked and answered before but are you modeling partial resists? Also does the SoV vs SoR modeling also take into account the judements? Also it seems the SoR glyph only affects SoR damage and not it's judgements.

Yes, all of the skills include the "boss_resist_reduce" factor, which accounts for partial resists.

The modeling calculates DPS for the full 969 rotation, which includes the appropriate Judgement. And the glyph bonus is only being applied to SoR, not JoR.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Call to Arms 5.x, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7658
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby PsiVen » Sat Mar 14, 2009 2:50 pm

theckhd wrote:... there are some bosses that don't melee, and Holy Shield is by far our weakest damage/threat component.


Surely that's true for bosses that don't melee, but have you examined the results on a scale of say, faster than Patchwerk to uninterrupted KT in terms of HS contribution to relative stat values? I apologize if this has been examined already, but I wasn't able to find something like this.

I know that on Patchwerk, my HS is often around #2 at 16% of my damage and thus by far the 2nd highest contributor for threat. Brutallus as memory serves placed HS squarely at #1, despite continuing to use other abilities while not the active tank. I wonder if this still would hold true today for a boss attacking at 2 swings / 1.2 sec.
Last edited by PsiVen on Sat Mar 14, 2009 2:58 pm, edited 1 time in total.
Gladiator Psiven, 90 Tankadin
90 Druid, 90 Mage, 85 Monk, 85 DK, 70 War, 70 Pal, 60 Priest, 60 Lock, 64 Rogue
Longtime addict of Space - Glory Through Conquest
User avatar
PsiVen
Moderator
 
Posts: 4363
Joined: Fri Jun 01, 2007 5:28 pm
Location: On a Boat

Postby majiben » Sat Mar 14, 2009 2:55 pm

Those bosses you both brought up are duel wielders the opposite of when you would want to do a holy shield subsitution. By dropping holyshield on those encounters you would increase the damage by a huge margin. The question of whether or not you get more threat is somewhat mute as you would never consider fropping holyshield.
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 PsiVen » Sat Mar 14, 2009 2:57 pm

I wasn't referencing the drop-HS rotation actually, I was referring to the effect on the relative value of stats due to the variable weighting of Holy Shield's scaling.
Gladiator Psiven, 90 Tankadin
90 Druid, 90 Mage, 85 Monk, 85 DK, 70 War, 70 Pal, 60 Priest, 60 Lock, 64 Rogue
Longtime addict of Space - Glory Through Conquest
User avatar
PsiVen
Moderator
 
Posts: 4363
Joined: Fri Jun 01, 2007 5:28 pm
Location: On a Boat

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