Theck's MATLAB thread - MoP/5.x

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Re: Theck's MATLAB thread - MoP/5.x

Postby lifeonmars » Wed May 08, 2013 12:50 pm

I don't think we should necessarily jump to conclusions yet, because as far as I know as of right now, the 5% SotR damage reduction nerf has only been discovered by datamining at MMO-Champion and has not appeared in any official patch notes for the PTR.

It's not uncommon for changes to be rapidly released and reverted on the PTR, and it seems strange to me that there has been zero official commentary on this change, versus the huge amount of text generated on the Grand Crusader change.
lifeonmars
 
Posts: 20
Joined: Wed Mar 28, 2012 9:52 am

Re: Theck's MATLAB thread - MoP/5.x

Postby Kerriodos » Wed May 08, 2013 2:23 pm

Of course not--PTR changes come and go all the time. That said, sometimes it's fun to speculate (though, admittedly, this thread may not be the place for it).
User avatar
Kerriodos
 
Posts: 168
Joined: Tue Dec 21, 2010 8:38 am

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Thu May 09, 2013 1:06 pm

Fetzie wrote:Not sure why that would pull us away from haste stacking.

Because any reduction to the baseline mitigation of SotR nerfs haste, hit, and expertise. You're lowering the effectiveness of SotR while it's up, which necessarily hurts stats that improve its uptime. This should be obvious if you consider the extreme limit: if you drop the baseline mitigation to 0% (and ignore mastery scaling), hit/exp/haste become useless.

That change also leaves mastery more or less unchanged.

lifeonmars wrote:I don't think we should necessarily jump to conclusions yet, because as far as I know as of right now, the 5% SotR damage reduction nerf has only been discovered by datamining at MMO-Champion and has not appeared in any official patch notes for the PTR.

It's not uncommon for changes to be rapidly released and reverted on the PTR, and it seems strange to me that there has been zero official commentary on this change, versus the huge amount of text generated on the Grand Crusader change.

That said, this is one of the potential nerfs they mentioned for 5.2 before they decided on the Grand Crusader change, if my memory is correct. Given gear inflation and higher mastery levels on gear, SotR may just be getting too strong to begin with. This nerf would simultaneously weaken SotR a little to bring us back into line and devalue haste slightly, both of which are likely good things.

I don't think it's going to be enough to reverse the haste-gearing paradigm though.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby Klaudandus » Thu May 09, 2013 1:21 pm

theckhd wrote:I don't think it's going to be enough to reverse the haste-gearing paradigm though.


But does the change closes the gap between C/H and C/M?
The Element of Forum Hyperbole
Image
---
Flüttershy - Draenei Protection Paladin, Aerie Peak
Klaudandus - BE Protection Paladin, Feathermoon (Semi-retired)
User avatar
Klaudandus
 
Posts: 9508
Joined: Thu Apr 02, 2009 7:08 am
Location: Texas' Armpit

Re: Theck's MATLAB thread - MoP/5.x

Postby Fetzie » Thu May 09, 2013 4:18 pm

theckhd wrote:
Fetzie wrote:Not sure why that would pull us away from haste stacking.

Because any reduction to the baseline mitigation of SotR nerfs haste, hit, and expertise. You're lowering the effectiveness of SotR while it's up, which necessarily hurts stats that improve its uptime. This should be obvious if you consider the extreme limit: if you drop the baseline mitigation to 0% (and ignore mastery scaling), hit/exp/haste become useless.

That change also leaves mastery more or less unchanged.


However, in a real world scenario, reducing a (for arguments sake) 100k hit to 50k instead of 45k isn't going to make that much of a difference. What counts is that you have the ShoR buff active in the first place, whether you have a bit more or a bit less reduction isn't going to do much other than reduce the amount the healer overheals you by with their Divine Light/Greater Heal. It is more a change that would be noticeable in the long run, not necessarily about when in the 5-8 second window you have the buff active, and if we cared about the long run over the short-term burst phases that can actually kill us then we'd be gearing for TDR, not for damage control.

While it would, technically, devalue haste slightly, I don't believe that anything much will actually change for the player. I don't think it will mean a change in rotation, it won't mean a change in gearing strategy and it won't mean that our survivability will change much, if at all. If it makes Mastery slightly stronger then that won't change much for most people, as that is the stat they are reforging to if they are already at the hit and expertise caps and the item already has haste on it.

In my opinion it is a very clever and subtle nerf, one that will increase our damage taken slightly, but not one that will mean significant changes to the way we play. And that is what I meant by my comment you responded to.

Anyway, it has yet to be documented in a patch note. The only indication we have that it might be changing is a partial tooltip alteration (that contradicts itself - the base reduction is 25%, but the minimum limit in the new wowdb tooltip is still 30%) that could simply be a parsing error, it definitely wouldn't be the first time that has happened.
Fetzie | Protection/Holy Paladin | EU-Kazzak
Former Author of the TankSpot Protection Paladin Guide
Image
Sagara wrote:You see, you need to *spread* the bun before you insert the hot dog.

bldavis wrote:we are trying to extend it as long as we can...it just never seems to last very long
User avatar
Fetzie
 
Posts: 2033
Joined: Sat Feb 07, 2009 9:43 am
Location: Karlsruhe, Germany

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Fri May 10, 2013 9:20 pm

While I agree with most of your logic, it doesn't change the fact that in smoothness simulations haste will lose some ground to mastery.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby Klaudandus » Sat May 11, 2013 5:31 am

theckhd wrote:While I agree with most of your logic, it doesn't change the fact that in smoothness simulations haste will lose some ground to mastery.


I am guessing we will hopefully see those graphs come 5.3?


Also, read someone on twitter saying that Pally TDR was pretty bad, that what propped pally tanks up was the AM control we have over the flow of the battle. Is this assertion correct?
The Element of Forum Hyperbole
Image
---
Flüttershy - Draenei Protection Paladin, Aerie Peak
Klaudandus - BE Protection Paladin, Feathermoon (Semi-retired)
User avatar
Klaudandus
 
Posts: 9508
Joined: Thu Apr 02, 2009 7:08 am
Location: Texas' Armpit

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Sat May 11, 2013 9:04 am

I'll have smoothness simulations done this week, probably. I've been stuck on SoI modeling for a while. With any luck I'll manage to write that blog post this weekend and get it posted Monday, then follow it up with some data later in the week.

I can't really speak to pally TDR vs. other classes, because I haven't looked at the other classes. That said, I'm not sure anyone who understands tanking takes TDR seriously in the first place, so I don't see any reason we'd need to "prop up" our poor TDR. If anything, the entire haste paradigm is about trading useless TDR for very useful AM control.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby Klaudandus » Sat May 11, 2013 9:22 am

On the tdr, yeah... I get that, just was wondering if it was a valid assertion or not.
The Element of Forum Hyperbole
Image
---
Flüttershy - Draenei Protection Paladin, Aerie Peak
Klaudandus - BE Protection Paladin, Feathermoon (Semi-retired)
User avatar
Klaudandus
 
Posts: 9508
Joined: Thu Apr 02, 2009 7:08 am
Location: Texas' Armpit

Re: Theck's MATLAB thread - MoP/5.x

Postby Derrickster » Wed May 15, 2013 3:59 am

I'm extremely confused as to how in the single target rotations CS>J>AS+>HW>cons>AS sims out as slightly more DPS and HPGS then the suggested CS>J>AS>HW>cons. I just can't seem to find any reason the holy power gains would be higher doing this seeing as how HW and cons can't proc grand crusader according to the tooltip. I can't say I've payed very much attention to how hard AS hits compared to HW but the only I can think of here is that it has to do with how the GCDs work out maybe? Can anyone please help me wrap my head around this?
Derrickster
 
Posts: 11
Joined: Thu Dec 23, 2010 3:15 pm

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Wed May 15, 2013 5:16 am

The holy power gain does look a little fishy, but keep in mind it's a tiny amount: 0.4428 vs. 0.4444. That's a difference of less than 0.5%, which might be about the level to which we can trust the accuracy of the model. I'll look at it quick and see if I can pinpoint where that difference is coming from.

As far as the DPS, it's a little surprising because AS does more damage than Cons or HW. But sometimes prioritizing the short-cd ability over the long-cd ability nets higher DPS if it eliminates empty GCDs. I think that's what's happening here - if you notice, the CS>J>AS>HW>Cons rotation has 8.9% empty GCDs, while CS>J>AS+>HW>Cons>AS has only 7.5%. Utilizing 1.5% more GCDs is a pretty big DPS gain.

That said, the advantage will probably disappear once you include SS and L90 talents, which eat up a good number of GCDs.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby Derrickster » Wed May 15, 2013 2:44 pm

After looking over the defensive queues portion of the single target rotation it just so happens just by throwing SS at the very bottom of the queue that your open GCDs drop down to 0%. The part of me that loves to be mathematically sure of things though would still like to see what happens when you throw SS and lv90s in to the CS>J>AS+>HW>cons>AS queue. Even if just as simple as throwing AS>Lv90>SS at the end to eat up the open GCDs and see where it falls then.
Derrickster
 
Posts: 11
Joined: Thu Dec 23, 2010 3:15 pm

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Wed May 15, 2013 3:44 pm

Looking at it, the Grand Crusader proc rate seems incredibly low for both of those queues (as in, one every 144 seconds or more). It looks like something in the C# got borked or commented at some point, because the proc rate from CS was defaulting to zero. I've fixed that now, so I'm going to re-run the entire suite tonight and tomorrow to update the numbers. Hopefully the updated data will look a little more sane.

Throwing SS at the end of the queue will generally eat up the remaining GCDs, because it's only got a 6-second cooldown. That gives you enough fillers to automatically fill every X in CS-J-X-CS-X-J-CS-X-X- every time. That doesn't really mean much for DPS though, because those GCDs don't contribute to DPS.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby Schroom » Wed May 15, 2013 11:47 pm

Worldbreaker's Stormscythe tends to pull ahead thanks to its hit/expertise itemization, but the two sword offerings (Qon's Flaming Scimitar and Do-tharak, the Swordbreaker) manage to match it in the hands of a Human.


what if you only gem full yellow hastegems. Do-tharak is the only one who will give you a socketbonus of 60 strength.

would this let it pull ahead of Worldbreaker's Stormscythe?
User avatar
Schroom
 
Posts: 670
Joined: Mon Aug 15, 2011 2:03 pm
Location: Luxembourg

Re: Theck's MATLAB thread - MoP/5.x

Postby Schroom » Thu May 16, 2013 9:55 am

GC twitter:

Going to try Vengeance cap at 30% health (down from 100%) in 10s and 50% of health in 25s. This is a big change, so might be 5.4, not 5.3.
User avatar
Schroom
 
Posts: 670
Joined: Mon Aug 15, 2011 2:03 pm
Location: Luxembourg

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Thu May 16, 2013 6:13 pm

Schroom wrote:
Worldbreaker's Stormscythe tends to pull ahead thanks to its hit/expertise itemization, but the two sword offerings (Qon's Flaming Scimitar and Do-tharak, the Swordbreaker) manage to match it in the hands of a Human.


what if you only gem full yellow hastegems. Do-tharak is the only one who will give you a socketbonus of 60 strength.

would this let it pull ahead of Worldbreaker's Stormscythe?

Probably, though it's already getting a haste/stam gem, so the difference is only 160 haste. That's still something like 400 DPS though.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Fri May 17, 2013 7:53 am

Updated the single-target rotations post with corrected data. I had to disable some of the queues (ones that combined ES with SS) because they cause the FSM code to fail now thanks to increased state size. They'll have to wait until I have time to decide how/if to fix them.

The oddity with the CS>J>AS+>HW>Cons>AS queue still exists - it has a small increase in HPG. I still have some thinking to do about this; theoretically it shouldn't happen, but I may just be missing a really strange edge case.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby Derrickster » Sat May 18, 2013 6:15 pm

Looking over the sims again it looks like any time AS+ is designated over just AS the HPGs goes up so the only thing I can think of that would cause this is if separating AS+ from AS someone makes the grand crusader usable instantly either through resetting the GCD or possibly making the proc'd AS off the GCD? I'm not sure I buy this explanation but it seems to be the only thing that could possibly be wrong without really knowing how the sim (or sims in general really) work.
Derrickster
 
Posts: 11
Joined: Thu Dec 23, 2010 3:15 pm

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Mon May 20, 2013 8:55 am

That can't happen in the sim, as far as I'm aware. But I'll double-check the code.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby Klaudandus » Thu May 23, 2013 1:14 pm

Ok, I might have some time off tomorrow, is there anything that needs to be tested since 5.3 came out?
The Element of Forum Hyperbole
Image
---
Flüttershy - Draenei Protection Paladin, Aerie Peak
Klaudandus - BE Protection Paladin, Feathermoon (Semi-retired)
User avatar
Klaudandus
 
Posts: 9508
Joined: Thu Apr 02, 2009 7:08 am
Location: Texas' Armpit

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Thu May 23, 2013 6:35 pm

I don't think there's anything pressing, no.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby theckhd » Wed Jun 12, 2013 7:34 am

Today's blog post is of some relevance to this thread:
http://www.sacredduty.net/2013/06/12/th ... uture-sim/
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 6.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 6211
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Theck's MATLAB thread - MoP/5.x

Postby Promdates » Sat Jul 20, 2013 6:54 am

Do you have an easy way of determining the Holy Power per minute of the new Holy Avenger and Sanctified Wrath talents in 5.4?
User avatar
Promdates
 
Posts: 121
Joined: Thu Apr 12, 2007 4:10 pm

Re: Theck's MATLAB thread - MoP/5.x

Postby Thels » Sun Jul 21, 2013 2:49 am

Promdates wrote:Do you have an easy way of determining the Holy Power per minute of the new Holy Avenger and Sanctified Wrath talents in 5.4?


Holy Avenger didn't change, did it?

Either way, I did some napkin math earlier:

Thels wrote:Lessee... Let's assume 0% haste for a moment.

HA: Lasts 18 seconds. During these 18 seconds, we'll press CS 4 times, and J 2.66 times on average. This would generate 13.33 extra HoPo per use, or 6.66 extra HoPo per minute.

SW: Lasts 30 seconds. During these 30 seconds, we'll press J 10 times and CS 5 times. This would generate 25 total HoPo per use. Normally, in that 30 second timespawn, we'd press CS 6.66 times and J 4.44 times on average, so we'd generate 11.11 total HoPo. Therefor, SW would generate 13.88 extra HoPo per use, or 4.62 HoPo/minute.

HA also buffs the HoPo from GC, while our SW rotation actually reduces the number of GC procs while increasing the chance we waste procs from avoidance. Ergo, not only does HA generate more HoPo per minute, it also generates more HoPo per use.



To make sure Haste doesn't mess things up too much, let's check the other end of the spectrum, 50% haste:

HA: Lasts 18 seconds. During these 18 seconds, we'll press CS 6 times, and J 4 times. This would generate 20 extra HoPo per use, or 10 extra HoPo per minute.

SW: Lasts 30 seconds. During these 30 seconds, we'll press J 15 times and CS 7.5 times on average. This would generate 37.5 total HoPo per use. Normally, in that 30 second timespawn, we'd press CS 10 times and J 6.66 times on average, so we'd generate 16.66 total HoPo. Therefor, SW would generate 20.83 extra HoPo per use, or 6.94 HoPo/minute.

Again, HA also buffs the HoPo from GC, while our SW rotation actually reduces the number of GC procs while increasing the chance we waste procs from avoidance. Ergo, not only does HA generate more HoPo per minute, it also generates more HoPo per use.


Of course, that was pre-GC change, but it doesn't change much. SW comes out a little better, since the less CS presses don't lead to less GC procs. Still, HA provides extra HoPo per AS+, and the tight SW rotation still has a bigger chance of wasting GC procs, so HA still wins out on HoPo generation per use. Obviously, with a shorter CD, it wins out on HoPo generation per minute as well.

However, the generation per use is really small, so if you have more than 20% haste, and are saving it for tight moments, SW's extra healing received might top that instead. Then again, do you really need the extra healing at that moment.

Under 20% haste, SW doesn't give you 100% SotR coverage, so HA clearly wins as a defensive cooldown.
ImageImage
User avatar
Thels
 
Posts: 1233
Joined: Tue May 15, 2007 7:30 am
Location: The Netherlands

Re: Theck's MATLAB thread - MoP/5.x

Postby Promdates » Mon Jul 22, 2013 8:35 am

I'm never looking at those talents as "defensive" cooldowns, as I would rarely use them for that (mostly because I've been running Divine Purpose since mid last tier). So in general, HA is roughly 10 HP/min. That's better than the random proc of 2-15 HP/min of Divine Purpose. Using HA with the 4P bonus and some "interesting" mechanics can make you get ~100% sotr uptime.
User avatar
Promdates
 
Posts: 121
Joined: Thu Apr 12, 2007 4:10 pm

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: Bing [Bot] 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: Bing [Bot] and 1 guest