Elemental....Time for a New Stat Priority?

[thegempolice]

Nice work Gendori, It's great to see some new Shaman theorycrafting that doesn't come from Bink or Endus, and isn't unsupported opinion like so many posters here. I've been uneasy about the prevailing wisdom on stat weights for a long time simply because it's all simmed using the Patchwerk model which, as we all know, is virtually irrelevant in modern raid encounters. However I had only considered the movement aspect and not the 100% DPS time on a single target with no switching.

I suppose the practical application of this testing is that we should be reforging for each encounter, and end up favouring Crit for the majority.

[EruptorNZ]

Don't see any of the elemental shamans in the top guilds eg. leeds, brightleaf, reforging into crit even when using primal elementalist.

Would be interesting to see if this could be applicable to reforing into crit for council, as I see this as the only fight this could be useful on, might try a crit reforge and a mastery reforge during attempts next week to see if theres a difference.

[Recom]

Huh ? Council only ? What about Tortos, Horridon, Lei shen, Ji-kun ?

This setup is really interesting, I've been running an aoe heavy setup for those bosses but I'll give this a shot since I don't see any downside to it.

[Moshne]

What settings did you use when you were simming? Did you have 3 mobs up the entire fight? How long did you make the fight? Did you run stat plots on the results to see if your results are an artifact of hitting some minor threshold? Default BIS profile or your personal one?

It is helpful to link the result so others can see what you did rather than relying on just your interpretation of the results. (using the GUI version of SimC there is a field at the bottom of the web page result where you can save it. Rename it something and upload it.)

Also, your premise that Mastery is worth less on CL is very flawed. It is worth relatively less if you are using it on less than 3 targets, sure. It is identical when you use it on 3, and when you use it glyphed and hit more targets you actually have increased benefit. In this tier you have several multitarget fights where you'll use CL: Horridon (>3 targets), Tortos on bats (>3), Ji-Kin (3-5, depending on platform), Lei Shen Ball Lightning (>3). In fact there are no fights (on 25m) where you'll be using CL frequently and be hitting less than three targets. Depending on your strat I suppose there is COE, but you'll be hitting 3 most of the time if you are hitting more than one.

Dark Animus, at least on 25, you'll be a corpse if your add is close enough to be able to cleave.

[Gendori]

What settings did you use when you were simming? Did you have 3 mobs up the entire fight? How long did you make the fight? Did you run stat plots on the results to see if your results are an artifact of hitting some minor threshold? Default BIS profile or your personal one?

It is helpful to link the result so others can see what you did rather than relying on just your interpretation of the results. (using the GUI version of SimC there is a field at the bottom of the web page result where you can save it. Rename it something and upload it.)

Also, your premise that Mastery is worth less on CL is very flawed. It is worth relatively less if you are using it on less than 3 targets, sure. It is identical when you use it on 3, and when you use it glyphed and hit more targets you actually have increased benefit. In this tier you have several multitarget fights where you'll use CL: Horridon (>3 targets), Tortos on bats (>3), Ji-Kin (3-5, depending on platform), Lei Shen Ball Lightning (>3). In fact there are no fights (on 25m) where you'll be using CL frequently and be hitting less than three targets. Depending on your strat I suppose there is COE, but you'll be hitting 3 most of the time if you are hitting more than one.

Dark Animus, at least on 25, you'll be a corpse if your add is close enough to be able to cleave.

Did you even read my original post? I put up all the numbers I got for 2 mobs....3 mobs patchwerk and later 4 and 5 mobs

The numbers dont lie and I dont even though why it is even a debate thst mastery is horrible on chain lightning....it has a 1/3 chance to proc, and tge chance is for each jump.....it doesnt matter if it hits 3 4 or 5 tsrgets it is significantly behind crit and haste. If chain lightning had the same chance to proc as every other spell it would only make mastery only as good.

Look at it this way. Lets say I have around 17% crit and 47% mastery.....on the initial damage on each jump I have a 17% chance to crit (for a whopping 2.5x the damage) and yet only a 16% chance to proc mastery and when it procs only does 75% of the damage. Just that simple napkin math alone should tell you how much better crit is

[Moshne]

Mastery works just fine on CL if you actually hit the right number of mobs. The reason it has a reduced chance to proc is you'd get triple the value otherwise. The correct way to look at Mastery chance is per cast. With one target, you have 1/3 chance per cast, with 2, 2/3 chance, 3 you get the full chance as well as the full damage potential. Glyphed, with 4 you have a 33% INCREASED chance, and with 5 a 66%. Additionally, Mastery is upping the chance for your T15 2pc and 4pc to proc.

There is a debate because it isn't as simple as you are making it out to be, and the actual fights don't occur the way in which you are simming them. I did see your original post, and I question the validity of your results as I don't know what Sim variables you ran or what criteria you used, good practice is to link to the results themselves, not just your interpretation of them.

[Justusy]

Sure mastery scales with more targets, but the other stats do too and don't have a 1/3rd penalty.. so they are better . So your calculation of 3 targets means 3/3 mastery is like saying "Yeah mastery is worth 100% on 3 targets, but crit is worth 300% of its normal value".

I've known all this a while being an avid simcraft user.
But note that the aoe priorities in the simulation are suboptimal with the buff to chain lightning, EG it's using earthquake at greater than 4 targets for instance. In fact if you change it to a bit more chain lightning spam your aoe dps simulation will go up a bit. It has a minimal impact on the scale values you're getting, but it does matter a bit.

[Gendori]

Sure mastery scales with more targets, but the other stats do too and don't have a 1/3rd penalty.. so they are better . So your calculation of 3 targets means 3/3 mastery is like saying "Yeah mastery is worth 100% on 3 targets, but crit is worth 300% of its normal value".

I've known all this a while being an avid simcraft user.
But note that the aoe priorities in the simulation are suboptimal with the buff to chain lightning, EG it's using earthquake at greater than 4 targets for instance. In fact if you change it to a bit more chain lightning spam your aoe dps simulation will go up a bit. It has a minimal impact on the scale values
you're getting, but it does matter a bit.

You said it better than me, thank you =)

Mastery works just fine on CL if you actually hit the right number of mobs. The reason it has a reduced chance to proc is you'd get triple the value otherwise. The correct way to look at Mastery chance is per cast. With one target, you have 1/3 chance per cast, with 2, 2/3 chance, 3 you get the full chance as well as the full damage potential. Glyphed, with 4 you have a 33% INCREASED chance, and with 5 a 66%. Additionally, Mastery is upping the chance for your T15 2pc and 4pc to proc.

There is a debate because it isn't as simple as you are making it out to be, and the actual fights don't occur the way in which you are simming them. I did see your original post, and I question the validity of your results as I don't know what Sim variables you ran or what criteria you used, good practice is to link to the results themselves, not just your interpretation of them.

Youre Missing the Point, The Point I am trying to make is that anytime there are 2 or more targets, crit and haste surpass mastery By far, and the gap gets wider as you add more targets. I understand that there isnt a 7 minute patchwerk fight where youre constantly AoEing tons of adds. If there was a fight like this youd see a little more than a 10% DPS gain on this fight going crit. Im going to say the average fight with a decent amount of chain lightning Id say you can probably realize a 3% gain in Total DPS. But the important part is that when theres lots of adds to kill, its usually really important to burn them down QUICKLY, so Play to your strengths which is your AoE, and burn those adds down ASAP!

The Greater Point I'm trying to make is that even On a straight Patchwerk 1 target fight, even if you reforge off all of your Mastery into crit it will turn out to be a less than 1% difference....Less than 1%! That's nothing. So while you lose almost Nothing on Patchwerk going crit, you have a Lot to gain when its Not Patchwerk.

Of course if you want to Reforge for EVERY SINGLE fight, Continue to stack mastery on Patchwerk and Then crit when youre going to be using CL a fair amount.

[Moshne]

Sure mastery scales with more targets, but the other stats do too and don't have a 1/3rd penalty.. so they are better . So your calculation of 3 targets means 3/3 mastery is like saying "Yeah mastery is worth 100% on 3 targets, but crit is worth 300% of its normal value".

This is incorrect. This idea of a "penalty" is a complete fallacy. The only reason CL has a reduced chance to proc on fewer targets is to normalize it when you are using it at max potential. You are getting zero penalty at 3 targets, and in fact an added benefit at more. When Mastery procs, you get a duplicated cast (though reduced base value, as noted above) effect. When you crit, you get the benefit for that jump only. If Crit had the effect of causing the entire chain to crit if an individual jump did, you'd have the same behavior as Mastery and your analogy would be sound. You actually get more Mastery procs per cast with 4-5 targets and less per cast with 1-2. You wouldn't say this about crit, it is completely linear with the number of targets, as is Haste.

I'm not discounting the fact that CL scales better with Crit, but what I am disputing is the validity (and the magnitude) of the results in the OP. They are skewed heavily in favor of 2-3 targets, and without seeing which talents and such you are using it is hard to adequately discuss them with you further. I don't understand what is so controversial about wanting you to post the actual output of your sim.

[Gendori]

This is incorrect. This idea of a "penalty" is a complete fallacy. The only reason CL has a reduced chance to proc on fewer targets is to normalize it when you are using it at max potential. You are getting zero penalty at 3 targets, and in fact an added benefit at more. When Mastery procs, you get a duplicated cast (though reduced base value, as noted above) effect. When you crit, you get the benefit for that jump only. If Crit had the effect of causing the entire chain to crit if an individual jump did, you'd have the same behavior as Mastery and your analogy would be sound. You actually get more Mastery procs per cast with 4-5 targets and less per cast with 1-2. You wouldn't say this about crit, it is completely linear with the number of targets, as is Haste.

I'm not discounting the fact that CL scales better with Crit, but what I am disputing is the validity (and the magnitude) of the results in the OP. They are skewed heavily in favor of 2-3 targets, and without seeing which talents and such you are using it is hard to adequately discuss them with you further. I don't understand what is so controversial about wanting you to post the actual output of your sim.

Youre Missing the Point about your other stats NOT having a Penalty on Chain Lightning or its Jumps while Mastery does, this is why Mastery is so weak on it

Ill repost My results from 4 and 5 targets and then Im going to resim this and show you, you might have missed this post

4 Targets Light Movement

INT 9.39
Crit 5.77
Haste 4.93
Mastery 1.72

5 Targets Light Movement

INT 11.21
Crit 6.77
Haste 5.47
Mastery 1.93

[Michelangelo]

Youre Missing the Point about your other stats NOT having a Penalty on Chain Lightning or its Jumps while Mastery does, this is why Mastery is so weak on it

Ill repost My results from 4 and 5 targets and then Im going to resim this and show you, you might have missed this post

4 Targets Light Movement

INT 9.39
Crit 5.77
Haste 4.93
Mastery 1.72

5 Targets Light Movement

INT 11.21
Crit 6.77
Haste 5.47
Mastery 1.93

Thank you, but we need additional sim information as was requested a bit further up on this page. There are way too many variables to be taken into account for us to take this information at face value.

I'll also sim on my character with different specs. It is important that we confirm not only that Crit significantly trumps mastery on a specific type of fight where adds are constantly up, but that the sims can match up to what our strategies per encounter are. Example: if you're trying to maximize your DPS on Horridon, AoE on Horridon adds may not make up for single-target damage on Horridon (200% damage taken). We're asking for all the variables so we can be sure that we can use a Crit build viably.

Also remember that every-time your stats change, so do your statweights. Do a double-check once you reforge/gem to crit and see what your new statweights are.

[Moshne]

I ran your tests myself and it appears SimC is handling Chain Lightning incorrectly. It IS giving CL a penalty when it doesn't do that in-game, a flat 1/3 chance to proc no matter the targets. I opened a ticket with SimC and it should be fixed later. You'll have to rebuild via the SVN to get the new version or wait until a new release is pushed.

(You can easily verify it isn't behaving by going and spamming CL on the dummies in Seven Stars/Two Moons. You should get higher than Mastery proc counts, but the sim is giving a fraction of it.)

edit: It is updated on the SVN now.

[Azlo]

I ran your tests myself and it appears SimC is handling Chain Lightning incorrectly. It IS giving CL a penalty when it doesn't do that in-game, a flat 1/3 chance to proc no matter the targets. I opened a ticket with SimC and it should be fixed later. You'll have to rebuild via the SVN to get the new version or wait until a new release is pushed.

(You can easily verify it isn't behaving by going and spamming CL on the dummies in Seven Stars/Two Moons. You should get higher than Mastery proc counts, but the sim is giving a fraction of it.)

edit: It is updated on the SVN now.

Can someone also hand-wave to me why Crit is better than Mastery for CL?

Let's assume the best case. I'm just CLing 5 targets continuously, and I have 2pc T15.

From my understanding, each jump of the CL has a chance to proc an overload. The overload does 75% of the initial damage, and still chains to 5 targets, each of these 5 additional hits also has a chance to proc lightning strike.

Crit, on the other hand, at 5 targets, has linear growth for 250% damage each jump.

I don't understand how Crit scales so exponentially when CLing, while apparently Mastery is garbage.

[Azlo]

I ran your tests myself and it appears SimC is handling Chain Lightning incorrectly. It IS giving CL a penalty when it doesn't do that in-game, a flat 1/3 chance to proc no matter the targets. I opened a ticket with SimC and it should be fixed later. You'll have to rebuild via the SVN to get the new version or wait until a new release is pushed.

(You can easily verify it isn't behaving by going and spamming CL on the dummies in Seven Stars/Two Moons. You should get higher than Mastery proc counts, but the sim is giving a fraction of it.)

edit: It is updated on the SVN now.

Can someone also hand-wave to me why Crit is better than Mastery for CL?

Let's assume the best case. I'm just CLing 5 targets continuously, and I have 2pc T15.

From my understanding, each jump of the CL has a chance to proc an overload. The overload does 75% of the initial damage, and still chains to 5 targets, each of these 5 additional hits also has a chance to proc lightning strike.

Crit, on the other hand, at 5 targets, has linear growth for 250% damage each jump.

I don't understand how Crit scales so exponentially when CLing, while apparently Mastery is garbage.

[Moshne]

Well, for starters, it isn't, SimC was just modeling CL wrong and the OP had a misunderstanding about how Mastery works on CL. Previously it was making it so CL only had a 1/3 chance to proc no matter whether you targeted 1 or 5 targets. That isn't the way the spell works.

With the fixed SimC the numbers reflect what I was saying, Mastery becomes the BEST stat for AEing with >3 targets, though Crit is still strong. Haste, as expected, tapers off as it caps. But central premise of this thread was that because Crit was so much better, we should be stacking Crit even for single target. That is just incorrect. At three targets Crit is best, but I'm not confident it is good enough to skew our gearing (especially considering how rare it is we'd be consistently CLing three targets this tier.)

3 targets: http://www.somethingwickedguild.com/simc/3targets.html
5 targets: http://www.somethingwickedguild.com/simc/5targets.html

If you want to duplicate the sim, I recommend turning off Earthquake in the default settings. It was questionable before with the incorrect handling, it is just plainly wrong now. The multitarget/Lava Burst logic probably needs to be retested as well.

[Azlo]

Well good. I was a bit uncomfortable thinking that my intuition was so far off in this case.

Does EB still increase 5 target cleave damage with the increased CL damage?

[Moshne]

I honestly haven't tested it in a while, but I'm inclined to say it won't, as it wasn't worth it before.

I don't know when CL rebroke in SimC either, I know I had them fix this very issue at some point back in Cata. I'm guessing some wires got crossed during all the MoP changes. I'm also not certain if they have it behaving properly with EoE right now. I'll play with the AE priority lists again when I get some time to run several tests.

[Gendori]

******THIS IS A LONG POST WITH A LOT OF NAPKIN MATH SO BE WARNED!*******

I Also put hours worth of work into this to make sure I had the formula right and such as I am pretty good at math, but Not wicked awesome, so be gentle =)

After doing some incredible Napkin Math, and checking the Numbers, The Results are well, Close

When you cast a chain Lightning that Hits 5 targets, You are essentially casting 5 Different spells on 5 different Mobs, Each is a separate event that has an independant chance for the following to happen

1)Your spell crits
2)It Procs your mastery and casts another 5 chain lightnings for 75% of the damage
3)If #2 happens these 5 Chain Lightnings again have a chance to crit, But not to again overload

I have 2 reforges I can reforge into

1)3695 crit 16.16%
9541 Mastery 49.14

2)6523 crit 20.87%
6523 Mastery 38.38

Lets Say we have 5 targets, and we cast a Chain Lightning on Target 1 for 40 Damage, It will then also Jump to the other 4 targets for 30 damage. Heres what you can expect from Reforge 1

I used the following formulas

For crit

x+y=Total Extra Damage from Crit

61.2*Crit Chance/100*5=x
x=49.44 extra damage from crit Per Chain Lightning Cast on 5 Targets

Now heres the result for Extra Damage from Mastery

49.14% Mastery gets Cut to 16.38% Chance to proc

40 Damage now does 30 Damage when it procs, However The proc will do a total of 190 Damage spread over 5 targets(the initial 40+the 150 mastery proc), so the formula is almost the same. Essentially your Mastery on 5 targets becomes a critical damage for 3.75x the damage

150*Mastery Chance/100*5=122.85 Extra damage per cast from your mastery

However, were not done yet, Your Crit Double Dips from your mastery so we have to calculate the damage crit gives you per cast off the mastery Jumps....the formula is the same


(122.85*2.53-122.85)*Crit Chance/100=y

y=30.37

Now add this to x (48.48) and you have a total of Crit adding 79.81 Damage per cast and Mastery adding 95.25 damage per cast

So On Paper it looks like Mastery is king right? Not so fast.....

To Recap so Far for Reforge 1 16.16% Crit will add 78.25 Damage Per cast While 49.14% Mastery adds 122.85 Damage per cast

Lets do the formula for reforge 2.......6713 Mastery(38.38%) 6523 Crit (20.87%)

x+y=Total Extra Damage for each cast from crit

61.2*20.87/100*5=x

x=63.86

For Mastery

38.38% Mastery becomes a 12.79% chance to overload on each jump

so 150*12.79/100*5=95.92 extra Damage From your mastery for each Cast of chain lightning

Now we have to calculate that those jumps also have a chance to crit

(95.92*2.5-95.92)*20.87/100=y

An extra 30.62 damage per cast from crit

Add that to x and we have 94.49 extra damage per cast coming from your crit

So which reforge Did more damage on that Single Cast?

I reforged out the same amount of crit and reforged in the same amount of mastery

Reforge 1 (Mastery Build)

Crit Adds 78.25 Damage per cast
Mastery adds 122.25 damage Per Cast (200.5)



Reforge 2 (Crit build)

Crit adds 94.49 Damage per cast
Mastery adds 95.92 Damage per cast (190.4)

So On 5 targets...for chain lightning only, Mastery wins? Perhaps! From a Pure One hard cast perspective, it appears so, however Crits also effect things like trinket procs, and gives you the 15% spellpower Clearcasting buff. This also doesnt take into account Damage from Magma Totem, earthquake and Fire elemental, and LAVA BEAMS, which all scale with crit, but NOT mastery.

Lets now do 4 targets

Reforge 1

61.2*16.16/100*4=39.56 Extra damage from crit per cast

120*16.38/100*4=78.62 Extra Damage per Cast from Mastery

(78.62*2.53-78.62)*16.16/100

78.62 Damage from Mastery will also add 19.44 to your crit damage

Crit 59
Mastery 78.62

137.62 total extra damage from Mastery build

Reforge 2

61.2*20.87/100*4=51.09 damage per cast from crit

120*12.78/100*4=61.34 damage per cast from Mastery

(61.34*2.53-61.34)*20.87/100=19.59

51.09+19.59

70.68 Damage per cast from Crit

61.34 Damage per Cast from Mastery

132.02 Extra Damage per cast from crit build
137.62 Extra Damage per Cast from Mastery build

Closer this time.....lets do 3 targets (wont bother with formulas just the numbers)

Mastery build

84.88 Extra damage per Cast

Crit build

83.84 Extra Damage per Cast

So there, Simcraft might be bugged, its good to find stuff like this, even if it takes 3 pages to get there, Now that its closer than it seems NOW we can argue!

Have at it!

[orderschvank]

did you also calculate our clearcasting w/e buff that we get from critting? every crit gives our next 2 spells more dmg, the more crit we have the more this spell is up.

[Gendori]

did you also calculate our clearcasting w/e buff that we get from critting? every crit gives our next 2 spells more dmg, the more crit we have the more this spell is up.

No, And I'm well aware of the clearcasting buff, thats why I said the results are inconclusive, its beyond my scope to factor in the clearcasting buff whereas taking everything else out and calculating Pure Crit vs. Pure Mastery for chain lightning is much more simple