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

Warning: Theorycraft inside.

Moderators: Worldie, Fridmarr, Aergis, theckhd

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

Postby Fridmarr » Wed Jun 24, 2009 8:51 pm

Does it even proc off of a full block now? I think that was just a tooltip clarification with zero change to what caused the ability to proc. IE, if you take non self inflicted damage it can proc.
Fridmarr
Global Mod
 
Posts: 7649
Joined: Sun Apr 08, 2007 8:03 am

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

Postby jere » Wed Jun 24, 2009 10:23 pm

Honestly I have no clue. I never bothered to test it after the tooltip changed because it didn't make any difference to boss fights. I assumed from his post it did, but it very well may not.
Image
User avatar
jere
 
Posts: 2726
Joined: Sat May 12, 2007 12:12 am

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

Postby Dread » Wed Jun 24, 2009 10:53 pm

Jasari wrote:
Dread wrote:Theck's original estimates for 3.2 numbers were provided before the coefficients for the new seals and judgements were released. Now they're available for use in the formulas. That might change the result (although from first glance it seemed that it widened the gap between 53/18 and a SotP build).

Can you post the new numbers, or link me to them? It'd be easy to update, but yeah it shouldn't change the rankings of anything since it'll relatively lower SotP which is still behind Crusade builds.


Seal of Vengeance now deals [ 13% of AP + 6.5% of Spell Power ] Holy damage (down from [ 15% of AP + 7.8% of Spell Power ]) and has a new effect - Once stacked to 5 times, each of the Paladin's attacks also deals 33% weapon damage as additional Holy damage.

From http://www.mmo-champion.com/index.php?topic=67249.0

Seems like the coefficients for SoV have gone down in 3.2, which is what lead me to believe the gap would widen further between SotP and Conv/Crusade. I believe Theck's previous numbers were based on the 3.1 coefficients.

Although it won't matter for SotP, I wonder if the new formula for SoC will make it anymore interesting: Seal of Command has been redesigned - All melee attacks deal [ 36% of mw ] to [ 36% of MW ] additional Holy damage. Lasts 30 min. Unleashing this Seal's energy will judge an enemy, instantly causing [ 8% of AP + 13% of Spell Power + 19% of mw ] to [ 8% of AP + 13% of Spell Power + 19% of MW ] Holy damage.
Dread
 
Posts: 89
Joined: Sun Nov 02, 2008 12:36 am

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

Postby steadypal » Wed Jun 24, 2009 10:58 pm

5 tick dot dmg is exactly the same as live...


on ptr 3/5 conviction 3/3 crusade 5 stack melee hit.. hits for roughly 80 more dmg than 5/5 sotp


me thinks sotp isnt affecting the 5stack hit dmg
steadypal
 
Posts: 967
Joined: Sat Mar 15, 2008 7:28 pm

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

Postby Ascendant » Thu Jun 25, 2009 4:59 am

I'm thinking here that if we end up seeing a lot of humanoid, undead, demon, or elemental bosses in the next tier of content, crusade will win hands down. is that correct?
Ascendant
 
Posts: 106
Joined: Mon Apr 14, 2008 6:27 pm

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

Postby WATERBOYsh » Thu Jun 25, 2009 5:17 am

Ascendant wrote:I'm thinking here that if we end up seeing a lot of humanoid, undead, demon, or elemental bosses in the next tier of content, crusade will win hands down. is that correct?


yes
User avatar
WATERBOYsh
 
Posts: 454
Joined: Wed Dec 26, 2007 4:50 am

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

Postby culhag » Thu Jun 25, 2009 10:52 am

Jasari wrote:• Each utility talent (either PoJ or Vindication) cost about 50 DPS each.

Is that per talent point or fully talented ?
RatingBuster - Tankadin Edition - Easily compare the threat and EH gains of your equipement ! *updated 25/06/10*
Image
User avatar
culhag
Maintankadonor
 
Posts: 1006
Joined: Wed Aug 06, 2008 2:50 pm
Location: France

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

Postby Jasari » Thu Jun 25, 2009 10:59 am

Culhag wrote:
Jasari wrote:• Each utility talent (either PoJ or Vindication) cost about 50 DPS each.

Is that per talent point or fully talented ?

That's full talented.

I'm going to redo all the calculations today though, with the following changes that I've read about. Feel free to confirm/deny any of these, it'll be another few hours before I get the chance to make the changes and run the tests:

  • 5 stack SoV procs cannot miss (or be dodged/parried)
  • 5 stack SoV procs do not scale with SotP
  • BV from items increased y 100%
  • Duration of LoO increased by 100% (giving it 100% uptime)
Neither of the BV changes should really make a difference since we haven't been testing BV sets, but they're worth making "just in case"

I will do a recalc of the estimated relative value of each talent and then a comparison of the same viable tanks specs as before. I'll only make changes to the specs if the order of best->worst talents change.
User avatar
Jasari
 
Posts: 2195
Joined: Sun Apr 27, 2008 7:32 am
Location: Boston, MA

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

Postby Jasari » Thu Jun 25, 2009 2:11 pm

Updated Analysis.

I made the changes that I listed before, increasing BV from items except shields, making the SoV 5-stack proc unavoidable and making it no longer benefit from SotP. This had dramatic changes on the relative value of talents, essentially plummeting SotP to the bottom of the list in DPS even though it remains ahead of Reckoning in TPS.

Relative DPS/TPS Value of Talents

Code: Select all
Talent   DPS       TPS
Reck     16.5359   30.9
SotP     14.2900   39.6
Conv     19.2554   45.9
Crus     30.4033   76.3
TbtL     45.9125   127
1HWS     90.3808   227


This put me in an in an interesting position - how should I value the talents in my example specs? I chose to go with the same specs that I used earlier which favor SotP over Reckoning because it's only a loss of about 2 DPS per point, it's a gain of about 9 TPS. So when chosing talents, I always chose that which would give me the most TPS which meant: Crusade>Convition>SotP>Reckoning. That said, the specs ended up staying exactly the same.

Goal:

We're trying to find the highest DPS/TPS spec based on how many utility talents you pick up. I'm defining utility talents as: Divine Sacrifice + Divine Guardian, 2/2 Pursuit of Justice, 2/2 Vindication. In order to pick up DSac+DGuard, we'll be taking the three points from Reckoning. In order to pick up PoJ, and/or Vindication, we'll be taking points from Conviction and/or SotP.

Setup:

Rotation: 969
Seals: Seal of Vengeance (SoV)
Glyphs: Judgement+Vengeance (JV)
Gear: Majiben's T8 Progression set - WITH TITANGUARD

Base Spec: 0/53/6

Key:

Spec:
S=SotP
V=Cov
C=Crus
R=Reck

Utility:
D=DSac+DGuard
P=PoJ
V=Vindication

Note: On specs that take only "P" for utility, it can be swapped for "V".
DPS/TPS of Viable Tanking Specs
Code: Select all
Utility  Spec      DPS       TPS
         5S3V3R    3.3152    8.3445
         5V3C33R   3.3693    8.4531
D        5S3V      3.2636    8.2482
D        5V3C3     3.3153    8.3522
P        5S1V3R    3.2743    8.2468
P        3V3C3R    3.3278    8.3543
DP       5S1V      3.2237    8.1522
DP       3V3C      3.2747    8.2551
PV       4S3R      3.2372    8.1518
PV       1V3C3R    3.2863    8.2554
DPV      4S        3.1870    8.0581
DPV      1V3C      3.2342    8.1580


Graph of TPS:
Image

Conclusions:

No real surprises here. Assuming SotP does not affect the SoV proc, it remains the same DPS/TPS as before since the SoV DoT damage was not changed at all. Reckoning received a very large boost in DPS from the SoV proc (a 47% increase) but it wasn't enough to catch anything besides SotP since all our other talents received buffs from our overall increased DPS/TPS. Additionally, we can clearly see that despite the ShoR nerf we're ahead of what we were at in 3.1. I've seen talk of ShoR getting a threat increase, possibly to compensate for the damage decrease but I haven't made any changes to account for that and it's not like we need additional threat at this point either. The gear set I'm using was the default one when I downloaded Theck's code and should possibly be changed. I updated the BV changes to gear and LoO, but still have LotSS equipped even though it's currently worse than LoO. I assume it will get buffed slightly which will buff all the talents except for Reck and SotP by a little bit. The only gear that I swapped was equipping Titanguard instead of Last Laugh.


tl;dr Summary:

  • Crusade specs are universally ahead of SotP specs (assuming SotP doesn't affect the SoV 5-stack Proc)
  • In terms of DPS, the order of talents goes: Crus > Conv > Reck > SotP
  • In terms of TPS, the order of talents goes: Crus > Conv > SotP > Reck
  • Reckoning's DPS per point has improved by about 47% in 3.2 (11.2 -> 16.5)
  • 0/53/18 builds remain the highest in DPS/TPS when compared to other specs that take the same utility talents
User avatar
Jasari
 
Posts: 2195
Joined: Sun Apr 27, 2008 7:32 am
Location: Boston, MA

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

Postby Dread » Thu Jun 25, 2009 2:35 pm

Thanks for running the analyses, Jasari. I know it can be quite time consuming. It's almost disappointing that the results came out the way they did. I was hoping Blizzard was trying to offer people flexibility on their spec (ie SotP for a maximum threat build or Crusade for a utility build). Even if it stays this way, I'm sure we'll still see those Paladins that insist SotP is the best use of their talents. :cry:
Dread
 
Posts: 89
Joined: Sun Nov 02, 2008 12:36 am

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

Postby Jasari » Thu Jun 25, 2009 2:48 pm

Most of the thanks should be directed at Theck, he really did all the development work. All I have to do is go back in a tweak a few numbers to reflect the latest changes.

As for variation in specs, I think the main differences you'll see (among people who want to maximize TPS) will be what utility talents they choose to pick up. It would be mathematically very hard for SotP to be better for TPS without utility and Crusade to be better for TPS with utility simply because you end up with so many left over talent points after picking up the necessary tanking talents that you can actually pick up several points in Conviction even with a SotP build. And once you're that low in the tree, it only take a couple more points to get Crusade. Basically, I think either Crusade will come out ahead in every situation, or SotP will come out ahead in every situation. The only way to change that would be if you didn't have enough left over talent points to go to the 3rd tier ret talents AND pick up SotP. I guess that might be the case if you want to also pick up Imp HoJ or some points in Divinity in addition Reck or DSac+DGuard.
User avatar
Jasari
 
Posts: 2195
Joined: Sun Apr 27, 2008 7:32 am
Location: Boston, MA

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

Postby Soldierxx » Thu Jun 25, 2009 3:04 pm

How come you only test 3 points in reckoning? For maximum TPS couldnt you take 2 points out of judgement of the just and put them in reckoning instead?

I thought judgement of the just would be considered a utility talent.
Soldierxx
 
Posts: 48
Joined: Thu Jun 04, 2009 6:37 am

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

Postby trellian » Thu Jun 25, 2009 3:06 pm

It has been tested extensively years (?) ago that 3/5 and 5/5 reckoning doesn't provide a noticable difference in it's procrate.
User avatar
trellian
 
Posts: 621
Joined: Wed Jul 16, 2008 10:02 am
Location: Silvermoon, EU

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

Postby Soldierxx » Thu Jun 25, 2009 3:10 pm

trellian wrote:It has been tested extensively years (?) ago that 3/5 and 5/5 reckoning doesn't provide a noticable difference in it's procrate.


A lot of the talents only provide a tiny bit of an upgrade but I thought the tests were trying to find the maximum TPS and DPS specs, so even if its a tiny upgrade, surely its worth testing 5 reckoning? Since a tiny upgrade, is still an upgrade.

Still dont understand why Vindication is considered a utility talent but judgement of the just isnt considered one.
Soldierxx
 
Posts: 48
Joined: Thu Jun 04, 2009 6:37 am

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

Postby Jasari » Thu Jun 25, 2009 3:31 pm

Soldierxx wrote:How come you only test 3 points in reckoning? For maximum TPS couldnt you take 2 points out of judgement of the just and put them in reckoning instead?

I thought judgement of the just would be considered a utility talent.


True, JotJ could be considered a utility talent. But if we wanted to maximize TPS we'd put those points into SotP rather than Reckoning. In the end though, I think there's enough of a consensus that JotJ is mandatory that I didn't bother with those tests.

I personally consider Vindication almost as valuable as JotJ so I will definitely be picking it up as well. The reason I classified it as a utility talent was to look at the TPS loss in picking it up since it will be a change from how we spec currently in 3.1.
User avatar
Jasari
 
Posts: 2195
Joined: Sun Apr 27, 2008 7:32 am
Location: Boston, MA

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 Wed Oct 15, 2008 1:28 am

Users browsing this forum: No registered users and 1 guest

Remove Advertisements