Remove Advertisements

Theck's MATLAB thread - MoP/5.x

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

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

Postby theckhd » Sun Nov 04, 2012 5:31 pm

Yes and no. The default cycle takes 13.5 seconds without haste:
CS-J-X-CS-X-J-CS-X-X

So provided we line up our HA cast with CS exactly, we get one full cycle plus CS-J-X within our 18-second window. In theory, any amount of haste would be enough to compress the rotation enough to include the next CS (i.e., 1 haste rating would be enough). It takes another 8.333% haste to fix the next X, another 8.3333% (16.67% total) to fit the next J, another 8.3333% (25% total) to get the next CS in, and so on. GC procs will make that noisier, but won't change the steady-state average.

So in theory, we get plateaus at <1%, 8.33%, 16.67%, 25%, 33.33%, etc.

In practice, there are a number of other factors that will blur that out a bit. The point in your cycle at which you cast HA matters a lot, for example - we've assumed you hit it exactly in conjunction with CS, but if you cast it half a second early that shifts the haste plateaus accordingly (they'll still be evenly spaced, but will happen at different numbers). Since human reaction time varies, it's fair to say that unless you're using an HA+CS macro to activate HA every time, it will blur the edges of the stepwise function fairly significantly.

And while it's "nonlinear" in the sense that it's a stepwise function, it's a stepwise function that approximates a linear function. So while your gains come in discrete chunks (before "blurring" effects), the overall scaling is linear (one way to think of it is that it's linear but rounded to the nearest GCD).
"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: 7655
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

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

Postby Klaudandus » Tue Nov 06, 2012 2:50 pm

In 5.1, we’re migrating four additional procs over to the Real PPM system. Dancing Steel and Jade Spirit are 2 Real PPM, River's Song is 4 Real PPM, and Colossus is 6 Real PPM.
The Element of Forum Hyperbole
Image
---
Flüttershy - Draenei Protection Paladin, Aerie Peak
Klaudandus - BE Protection Paladin, Feathermoon (Semi-retired)
User avatar
Klaudandus
 
Posts: 10805
Joined: Thu Apr 02, 2009 7:08 am
Location: Texas' Armpit

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

Postby Lastwolf » Wed Nov 07, 2012 7:19 am

That makes colossus an 800HPS absorb, before haste?
Lastwolf
 
Posts: 73
Joined: Thu Nov 01, 2012 3:58 am

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

Postby Klaudandus » Wed Nov 07, 2012 7:23 am

Lastwolf wrote:That makes colossus an 800HPS absorb, before haste?


No idea... I am not the math genius, but that tidbit was pretty buried in a bunch of blue posts from the last few days, So I thought it was worth mentioning.
The Element of Forum Hyperbole
Image
---
Flüttershy - Draenei Protection Paladin, Aerie Peak
Klaudandus - BE Protection Paladin, Feathermoon (Semi-retired)
User avatar
Klaudandus
 
Posts: 10805
Joined: Thu Apr 02, 2009 7:08 am
Location: Texas' Armpit

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

Postby theckhd » Wed Nov 07, 2012 8:00 am

It's an 8k absorb every 10 seconds (on average), so yes, about 800 HPS. Exact value will differ slightly due to probabilities, but that rough estimate should be close enough.
"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: 7655
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

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

Postby Nooska » Thu Nov 08, 2012 5:03 am

800 HPS that scales with haste, from the previous explanation of RPPM.
Main Characters:
Nooska, Blood Elf BM/SV Hunter on Argent Dawn (EU)
Morosin, Bloody freezing orc death knight on Argent Dawn (EU)
Niisca, Shady forsaken "priest" on Argent Dawn (EU)

Keeper Emeritus of the BM hunters guide on Elitist Jerks and the wowhead version untill patch 5.3.
User avatar
Nooska
 
Posts: 1348
Joined: Sun Nov 27, 2011 10:55 am

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

Postby theckhd » Thu Nov 08, 2012 7:42 am

To be fair, he did say "before haste." :P

But yes, in our case it'll be 800*spell_haste absorption per second.
"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: 7655
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

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

Postby Nooska » Thu Nov 08, 2012 7:43 am

I fail at reading today (also). :D
Main Characters:
Nooska, Blood Elf BM/SV Hunter on Argent Dawn (EU)
Morosin, Bloody freezing orc death knight on Argent Dawn (EU)
Niisca, Shady forsaken "priest" on Argent Dawn (EU)

Keeper Emeritus of the BM hunters guide on Elitist Jerks and the wowhead version untill patch 5.3.
User avatar
Nooska
 
Posts: 1348
Joined: Sun Nov 27, 2011 10:55 am

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

Postby Brokenone » Wed Nov 21, 2012 6:16 am

With Valor Point upgrades coming soon in 5.1, can anyone run the numbers for which piece is best to upgrade? A mage in our guild found for mages that weapon > chest/legs/helm/trinket >> everything else. I suspect that will hold true for us as well, but it'd be nice to see some math backing it up. Here is his (rough) math:

nathanbp wrote:As everyone knows, coming in 5.1 is the ability to upgrade your raid items with Valor Points. Following is a consideration of the best items to upgrade. Specifically, how much better is it to upgrade a weapon than any other item?

Fire Mage T14N Scale Factors (from Simulationcraft):
Int 4.8
SP 3.46
Hit 3.44
Crit 2.83
Hast 2.29
Master 2.11

We'll consider Regail's Crackling Dagger (Normal). Upgrading by 8 item levels increases all stats by 8% (approx.). This gives
Int +36 (172.8 dps)
Mastery +27 (56.97 dps)
Crit +21 (59.43 dps)
Spellpower +561 (1941.06 dps)
Total: 2230.26 dps

As you can see, the majority of the dps gain comes from the additional spellpower. Compare this to upgrading Leggings of the Burning Scroll. 8 extra item levels only nets 709.6 dps. And it's only that high because the item slot multiplier for legs is 1. Upgrading Ring of Wretched Flames only nets 437.67 dps.

Upgrading trinkets looks to be about as good as upgrading pants/chest/helm (the 1.0 multiplier slots). Relic of Yu'lon +8 item levels is a 716 dps increase. Light of the Cosmos +8 is a 847 dps increase.

In conclusion, upgrading your weapon at is least 3 times as good as upgrading any other slot. For the maximum dps upgrade to help with progression, it likely makes sense to spend valor on upgrading your weapon, even if it is not BiS, assuming that it's the best weapon that will be available to you in the near term. Second in priority should be trinkets as they are likely to be your next best upgrade slot (and also since you have 2 trinket slots), however at some point it's better to just consider which items you have are BiS and upgrade those.
Image
Brokenone
 
Posts: 150
Joined: Fri Sep 07, 2007 11:56 pm

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

Postby theckhd » Wed Nov 21, 2012 6:33 am

It depends on what you're trying to increase. Helm/Chest/Legs all have itemization multipliers of 1 (everything else is lower), so you'll generally get more STR/STA and secondary stats by increasing one of those three. Trinkets are close behind in stamina, but generally not secondary stats. Everything else falls behind, relatively.

So I think the most logical approach is to figure out which of those three slots has the best itemization (i.e. use it on a hit/exp/haste/mastery item, not on one with dodge/parry) and boost that one.

I don't think the weapon slot will be as powerful for us. It may be the best raw DPS upgrade due to weapon damage, but I'd need to verify that to be sure because weapons have less than half of the STR/STA itemization of chests. So it may be that the extra 1k strength you get from upgrading a chest offsets or exceeds the boost from increased weapon damage, especially since the majority of our DPS is independent of weapon damage in the first place.
"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: 7655
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

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

Postby daishan » Wed Nov 21, 2012 7:10 am

Where would the shield come in that list do you think?
Low on stats but has a massive armour budget.
Daishan of <Nidor Amo Nex>
"I'll be more enthusiastic about encouraging thinking outside the box when there's evidence of any thinking going on inside it." -- Terry Pratchett
User avatar
daishan
Maintankadonor
 
Posts: 551
Joined: Sun Mar 06, 2011 4:04 pm
Location: UK

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

Postby Brokenone » Wed Nov 21, 2012 7:56 am

The shield makes me sad because it's dodge/parry. I think I'd rather upgrade a chest/helm/legs first.
Image
Brokenone
 
Posts: 150
Joined: Fri Sep 07, 2007 11:56 pm

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

Postby Hespherus » Wed Nov 21, 2012 8:03 am

Weapon should be first for us also. Might not be much, but if you are still progressing, more often than not i think, its gonna matter if you will be able to pull 1k more dps, then just having 70 more stamina + 50 strength and some secondary stats. I have no math to back that up, but i feel that tank dps matters a lot(10 man PoV).
Hespherus
 
Posts: 158
Joined: Fri Sep 02, 2011 1:03 am

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

Postby theckhd » Wed Nov 21, 2012 1:36 pm

daishan wrote:Where would the shield come in that list do you think?
Low on stats but has a massive armour budget.

Good question. Armor is a big stat for us, so shield might be a decent choice. I'll look at the math for that later.

Hespherus wrote:Weapon should be first for us also. Might not be much, but if you are still progressing, more often than not i think, its gonna matter if you will be able to pull 1k more dps, then just having 70 more stamina + 50 strength and some secondary stats. I have no math to back that up, but i feel that tank dps matters a lot(10 man PoV).

Except you might get more DPS from upgrading your chest, because weapon DPS is fairly weak for us compared to most other classes. Remember that the only thing in our arsenal that scales with weapon DPS is Crusader Strike, while everything scales with AP.
"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: 7655
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

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

Postby Fetzie » Wed Nov 21, 2012 6:09 pm

I'll probably be doing the chest, helmet or the legs first, as they have the biggest stat budget. As you said, depending on what stats are there. No real point wasting 1500 VP on a dodge/parry item.
Fetzie | Protection Paladin | EU-Anub'arak
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: 2130
Joined: Sat Feb 07, 2009 9:43 am
Location: Karlsruhe, Germany

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