Simulationcraft results for WoD

[shaithis]

I was under the impression that Stone of Fire was passive mastery with an agi/int/str proc, just like the blue version, and the version currently on ptr.wowhead is passive vers with an agi/int/str proc, so... is it really that much of a nerf?

I mean, it wouldn't have had passive agi in the first place since it was designed to be a one-for-all type of trinket, so I'm really lost as to what you guys are talking about.

Yeah he meant to type Mastery, not Agility.

Well it was way too strong.

Well it was Mythic tier, but not stronger than other top trinkets, now it is barely Heroic tier...

[Toroq]

Yeah he meant to type Mastery, not Agility.



Well it was Mythic tier, but not stronger than other top trinkets, now it is barely Heroic tier...

Isn't the proc the main strength of that trinket? Just how much of a nerf is it to change the passive stat from mastery to vers? I know it's not a good stat for any rogue spec, but it seems to me like people are overreacting.

[fierydemise]

Just throwing this out there before everyone gets really excited about DfA, it whiffs probably more often than you think.

This is from some testing I was doing on the swing timer last night, 58 DfA casts only 48 eviscerates, ~18% of the eviscerates whiffed.
https://www.warcraftlogs.com/reports...-done&source=1

Celestalon says they have a fix in this PTR build or the next one (https://twitter.com/Celestalon/statu...69205899554817) but until the whiff rate is near 0 I'm not sure I'd take it.

[shaithis]

Isn't the proc the main strength of that trinket? Just how much of a nerf is it to change the passive stat from mastery to vers? I know it's not a good stat for any rogue spec, but it seems to me like people are overreacting.

You know, that is actually a good question, back when I originally crunched the numbers on it, I was getting that the proc normalized to 368 agi. Of course, I was using same mechanics as Draenic Philosopher Stone (55sec ICD), so 1350 * 15/55, is this not the correct thought process? So, 253 vers 368 agi, using Subtlety values from first page, 253 * 3.43 + 368 * 7.79 = 3,734.51 points. Compare to M Scales of Doom, 265 agi, 267 multistrike (normalized, 15.3% uptime * 1743), 267 * 4.07 + 265 * 7.79 = 3,151.04 points. So according to this, SoF should be ~18.5% ahead of Scales for Subt...what is wrong in my math / assumptions, or are the sims not working properly?

By the way, small nitpick, but would be nice if you guys could add normal versions of Beating Heart, Meaty Dragonspine and Humming Blackiron to the Trinket analysis, looking like they may end up ranking highly for the respective specs on which they are BiS.

[Imotbh]

Is it worth doing a cycle like SS > Evi > SS > Evi to make the 4-Part Combat set procs? I thought it was 20% chance to proc with w/e Combo Point you had but it seems it's only 4% per CP (https://twitter.com/WarcraftDevs/sta...28183534829569). Nevertheless, I've seen really high numbers on Darmac testing streams (so mainly AE fight) with the 1CP spamming method.

[gahddo]

Is it worth doing a cycle like SS > Evi > SS > Evi to make the 4-Part Combat set procs? I thought it was 20% chance to proc with w/e Combo Point you had but it seems it's only 4% per CP (https://twitter.com/WarcraftDevs/sta...28183534829569). Nevertheless, I've seen really high numbers on Darmac testing streams (so mainly AE fight) with the 1CP spamming method.

Likely that it's actually 20% flat right now and that was just him trying to save face and they're fixing it.

- - - Updated - - -

@ryme Trying to figure out how you got the first graph in the OP. Assassination at 43k o.o? The profile you have posted does ~38k on default settings.

- - - Updated - - -

Yea...there might be a major bug right now with assassination energy generation. I'll update when I have more information.

- - - Updated - - -

Reia pushed something 2 days ago that fixed a bug where DfA was generating 2x relentless strikes energy. In the process, envenom also accidentally got nuked. They applied a check to see which spells were affected by relentless strikes, and for whatever idiotic reason, envenom is the only spell not listed in the spell data. Should get fixed by tomorrow morning.

[Deleted]

Is Multistrike that good for combat cleave, that you would actually go down from 685 pieces with mastery/crit and haste/vers, to 665 and 660 with Multistrike/x?

[Puffah]

The BiS list for combat lists Meaty and Heart of the mountain, while on the analysis Humming ranks higher than Heart. Is Heart considered BiS simply because it has an on-use?

[Generical]

Any reason why Haste would start pulling back ahead of Multistrike for 2 target cleave as Combat? Multistrike is still ahead on 3 targets, but Simcraft, and Shadowcraft are showing that haste is ahead on 2 targets now.

[Ryme]

The BiS list for combat lists Meaty and Heart of the mountain, while on the analysis Humming ranks higher than Heart. Is Heart considered BiS simply because it has an on-use?

Thanks for the catch, should be using Humming - updated.

@ryme Trying to figure out how you got the first graph in the OP. Assassination at 43k o.o? The profile you have posted does ~38k on default settings.

Are you disabling set bonuses?

[Imkilo]

I'm curious as to why the Multi/crit belt/neck is pulling ahead of the multi/mastery counter-parts that have more multi for combat?

[Ryme]

Combat is using haste/multistrike belt, and the multistrike crit neck offers better returns than the multistrike/mastery neck.

[shaithis]

You know, that is actually a good question, back when I originally crunched the numbers on it, I was getting that the proc normalized to 368 agi. Of course, I was using same mechanics as Draenic Philosopher Stone (55sec ICD), so 1350 * 15/55, is this not the correct thought process? So, 253 vers 368 agi, using Subtlety values from first page, 253 * 3.43 + 368 * 7.79 = 3,734.51 points. Compare to M Scales of Doom, 265 agi, 267 multistrike (normalized, 15.3% uptime * 1743), 267 * 4.07 + 265 * 7.79 = 3,151.04 points. So according to this, SoF should be ~18.5% ahead of Scales for Subt...what is wrong in my math / assumptions, or are the sims not working properly?

By the way, small nitpick, but would be nice if you guys could add normal versions of Beating Heart, Meaty Dragonspine and Humming Blackiron to the Trinket analysis, looking like they may end up ranking highly for the respective specs on which they are BiS.

Bumping this, has anyone had time to check my calcs on SoF / verify that it is being simmed correctly?

[Deleted]

Multistrike is by far your best AoE stat as combat. Fast/Slow is the set up you want for multiple targets, as per the work from firey:


The reasoning is that poisons can multistrike when they do they go through as non-distinct damage through blade flurry where they can then proc more poison hits. Alongside shifting more damage from yellow hits to white which are unaffected by the BF penalty.



Not really needed if it's all there already I don't like these kinds of tables, we set up the SimC APL to take into account the effect of each tier set, so it's never accurate to just disable or shift things around. With that set, doing a simple disable_set_bonuses=1 will tell you that it's roughly a 9% gain for sub and an 11% gain for combat and assassination.

Apologies for asking but, after clicking on the link to the work from Fiery, I was left wondering where those numbers came from.

Do we have actual maths for this or are we just taking his word for it that the numbers are correct?

[Mac223]

Fiery specifically stated that he got the numbers from simC.

[Deleted]

Where can I find the maths behind Simulationcraft's methods?

[Mac223]

https://code.google.com/p/simulation...ource/checkout

[Deleted]

I'll have a look when I get the time. Never keen on just taking peoples' word for things.

[fierydemise]

Apologies for asking but, after clicking on the link to the work from Fiery, I was left wondering where those numbers came from.

Do we have actual maths for this or are we just taking his word for it that the numbers are correct?

For these numbers I used the my gear ~ilvl 670 with multistrike gems and enchants, ilvl 670 Phemos for slow and ilvl 670 boot knife for fast and I simply ran SimC (450 seconds) varying the number of targets from 1 to 10. Also to make sure I was only looking at BF based effects I removed the CT lines from the APL.

Full data spreadsheet is here: https://www.dropbox.com/s/bl14y404qh...gets.xlsx?dl=0

You can see a very similar set of results in ShadowCraft, its a bit annoying with the webUI doesn't list fast MHs in the combat dropdown but you can import a fast MH or select one while in another spec. Similarly you can check the results and dump them into a nice pretty CSV using a quick python script with just the ShC engine.

You can find the ShC engine source here: https://github.com/dazer/ShadowCraft-Engine

Obviously its a good thing to check how things were computed to make sure nothing wonky is going on but when two separately developed tools both come to the same conclusion that gives me a lot of confidence in the result.

[Deleted]

ANyone got a updated list of TIER BONUSES? The updated list I can see only shows SUb and Assa as 2,3 % gain from 4p?