Elemental Shaman info

Started by Nasanna, December 03, 2008, 07:12:24 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Nasanna

What stats should you stack as elemental? Which is better, haste or crit? What rotation should you use? Post any information you find about elemental shamans that can help answer these questions here.

What I've found so far:

Crit is important for elemental shamans because crits procs elemental oath and clearcasting. However, with a guaranteed crit every 8 seconds with lava burst, it seems to be less important. At this point, I do not know if lightning bolt scales better with spellpower, haste, or crit. The problem with haste is that you are sort of "capped" with haste by the cooldown on lava burst. I'll try to do some math and see what I can figure out. If I make any progress, I'll post it here.

This thread is not totally updated and there is little math in it but there are some good ideas: http://elitistjerks.com/f79/t36811-elemental_ttt_entry_work--progress/

Urmyboyblue

i would say crit...let me run some numbers as well

Iliessa

I have been reading the Elitest Jerks forums for about 5 hours now and there are a couple of things that everyone there seems to agree on so far. Unfortunately it doesn't answer the Crit vs. Haste question.

As is true with most DPS, Hit > all. That seems especially important with Lich King. Imagine if Flame Shock missed. Tragedy! And for trash mobs and PvP, you sure don't want Hex to go on that 45 second cooldown without some lolribbitâ,,¢ results to show for it. According to EJ forums the level 80 Hit cap is 341.

So I've looked around on Wowhead and would you believe there is not a single piece of level 77-80 mail armor with Hit Rating on it pre T7? Ugh. Since we're going to have to find our Hit elsewhere, I thought I would post the items that came up from a filtered search:

Pre-Naxx Hit Cloaks and Accessories
Rings:                           Source:
[item]The Prospector's Prize[/item] - Heroic Halls of Stone - Krystallus - 24% drop rate
[item]Ring of Northern Tears[/item] - Jewelcrafting - 2x[item]Titanium Bar[/item], [item]Eternal Water[/item], 4x[item]Northsea Pearl[/item]
[item]Bat-Wool Signet[/item] - Icecrown Group Quest - Basic Chemistry (currently bugged according to comments, and I wouldn't suggest this ring anyway due to the spirit and low hit..)

Neck:                            Source:
[item]Necklace of Taldaram[/item] - Heroic Ahn'kahet: The Old Kingdom - Prince Taldaram - 17% drop rate
[item]Emeline's Locket[/item] - Icecrown Group Quest - The Admiral Revealed

Trinket:                          Source
[item]Mark of the War Prisoner[/item] - Heroic The Violet Hold - Cyanigosa - 20% drop rate

Cloak:                            Source:
[item]Cape of Seething Steam[/item] - Regular or Heroic Halls of Lightning - Volkhan - 13% drop rate
[item]Dark Soldier Cape[/item] - Reputation Reward - Honored with Knights of the Ebon Blade - 20g 2s

Getting these items should put you at roughly 225 Hit Rating. Searching for Epics using the same item filters returns 0 results.


Now as far as rotation goes, the general consensus appears to be (Pre-Shamanism):

Flame Shock, Lava Burst, Chain Lightning, X*Lightning Bolt, Lava Burst, Chain Lightning, Y*Lightning Bolt

Post-Shamanism (I believe) you would remove Chain Lightning from the rotation. The problem with our rotations is that we run into the dilemma of that last lightning bolt either going over the Lava Burst cooldown, or we sit and wait for the cooldown to end. It doesn't seem like it's that big of a deal to wait half a second for the cooldown to finish but according to the numbers, even .1 seconds of lag can noticeably hurt your dps so half a second would be a killer in a 10 minute boss fight. It means you would have been standing there doing nothing for 37.5 seconds of the fight. A similar DPS loss would happen if you decided to cast the Lightning Bolt anyway and basically turned Lava Bursts 8 second cooldown into 9. Not good either.

What makes sense at first glance would be to determine the haste required for a specific rotation. For example:

FS, LvB, CL, 3xLB, LvB, CL, 2xLB = ? Haste Rating
FS, LvB, CL, 4xLB, LvB, CL, 3xLB = ? Haste Rating


The problem with that is being inbetween Haste Ratings/Rotations is going to make up 90% of your raiding time so you will be gimped until you are done, unless you can effectively juggle your Haste Rating by including or excluding your buffs. And you'd still probably need backup equipment. Maybe that is what Blizzard is expecting us to do, I don't know. And even figuring out the numbers for this now will (might) be irrelevant as soon as Shamanism is added to our trees, and will need to be reworked again due to the difference in cast time between CL and LB.

I'll add more on this later, I need to go take an Excedrin...  ;)

Nasanna

I was going to start doing my math by figuring out a rotation, then figuring out how it changed with haste, crit, and spell damage. I assumed that the shaman had [item]Glyph of Flame Shock[/item]. I wrote out all the spells, lengths and cooldowns, and like you figured out Illiessa, there are little gaps of half a second where you should be casting but are waiting for lava burst to cool down.

I talked to Panzer about how he rotates his enhancement attacks and he uses a priority system, based on what is on cooldown and what is available, which seems to make sense here. It would be ridiculously hard to not have any haste through totems, heroism, and raid buffs.

So your priority would look something like:
1) Flameshock if it's about to wear off and Lava Burst is nearly cool
2) Lava Burst
3) Chain lightning
4) Lightning Bolt

Meaning that you want to use in this order, Flameshock if it's about to fall off, Lava Burst whenever it's not on cooldown, chain lightning when its not on cooldown, if everything else is cooling down, lightning bolt. If there is CC around your target or you have mana problems, take out the chain lightning.

Another way to look at the haste vs. spellpower vs. crit problem is this addon I found called Dr. Damage (Shadow was kind enough to host it for us here: http://forum.twilightonalex.com/index.php?topic=6067.0 ). I started using it to figure out how much to downrank my heals back in Kara. It does a lot of good things. It calculates and shows on your buttons how much damage or healing each spell is expected to do. It accounts for your buffs, figures out when you use your trinkets, and looks at all your talents to determine what is going on.

Additionally, when you hover over a spell, the tooltip pops up and tells you how much more damage/dps your spell will do, on average, with 1% hit, 1% crit, 1% haste, and 10 spell damage so you can look at it and see what benefits the most from each stat. However, I'm resto at the moment and really don't want to pay the 100g to respect just to look at some numbers. I'll take a look later next week when I'll have some free time if you guys don't.

Arcdelad

while you are looking up that info nas, you should also find a schexy shaman avatar like illiesa...yowza!

Iliessa

Quote from: Nasanna on December 04, 2008, 05:17:27 AM
Meaning that you want to use in this order, Flameshock if it's about to fall off, Lava Burst whenever it's not on cooldown, chain lightning when its not on cooldown, if everything else is cooling down, lightning bolt. If there is CC around your target or you have mana problems, take out the chain lightning.

There were two guys on the EJ forums arguing this exact point, focusing on using CL twice in a single 8 second Lava Burst cooldown. I have been using this 'priority system' as I went through dungeons to level up and it is decent dps, but it causes issues in timing of the rotation. I kept trying to figure out why my rotation was never the same twice, and it ended up being the second Chain Lightning cooldown that was messing it up.

I have very little spell haste, only 24. That's less than 1%. Regardless of how much haste you have, haste does not affect the length of a cooldown. Lets say you are using the main rotation I listed above, but throwing in CL every time it's off cooldown. You'd end up with having to change your spell cast order every 8 second rotation. Example with the cast time and cooldowns above and below the respective spells:

Cooldown:   8     3.........      3...........
             FS, LvB, CL, 2xLB, CL, LvB, LB, CL, 2xLB
Cast time:          1.5   4      1.5


What this is saying is because of the second CL cast in the first 8 second LvB cooldown, you will not be able to cast CL right after you refresh LvB because CL will still be on cooldown. You would have to cast a LB first, which is no big deal, really, but at the end of my above example you will have both FS and CL ending at about the same time, which may cause wait time for that final tic of FS to go off before refreshing since that has priority over CL, and the entire rotation only adds up to 17.5 seconds (assuming no haste and no lag, which will not be the case, but is hard to account for right now). Also the first 8 second cooldown only has 7 seconds of cast time in it. You will need enough haste to shave another second off the cast time to safely squeeze in another LB before LvB and at that point the priority system starts to gain a little bit of stability. You may still end up with the problem of CL and FS being ready at the same time. It's hard to say since I'm not personally good with the number crunching and not sure of the best way to account for lag and other things that can make a difference over an 18 second rotation. The other argument against the priority system is human error. You are already trying to keep an eye on two cooldowns that will be showing on your screen in 2 separate places - LvB and FS debuff. Throw the CL cooldown in there and your eyes are going to be flicking between your spell bar and the target's debuff list quite a bit, and may become overwhelming to try and keep up with. One last thing to keep in mind with CL in the rotation is that by casting it immediately after LvB, you will be guaranteeing that is gains the benefit of Clearcasting, as opposed to throwing in additional CLs when it may not be. But that only matters if you have mana issues.

And bear in mind that Shamanism may end up making CL obsolete. Right now CL is higher DPS than LB because it's roughly the same damage output in a shorter amount of time, but when Shamanism buffs LB by 10%, it might not matter anymore.

Quote from: Nasanna on December 04, 2008, 05:17:27 AM
Another way to look at the haste vs. spellpower vs. crit problem is this addon I found called Dr. Damage (Shadow was kind enough to host it for us here: http://forum.twilightonalex.com/index.php?topic=6067.0 ). I started using it to figure out how much to downrank my heals back in Kara. It does a lot of good things. It calculates and shows on your buttons how much damage or healing each spell is expected to do. It accounts for your buffs, figures out when you use your trinkets, and looks at all your talents to determine what is going on.

Additionally, when you hover over a spell, the tooltip pops up and tells you how much more damage/dps your spell will do, on average, with 1% hit, 1% crit, 1% haste, and 10 spell damage so you can look at it and see what benefits the most from each stat. However, I'm resto at the moment and really don't want to pay the 100g to respect just to look at some numbers. I'll take a look later next week when I'll have some free time if you guys don't.

I used to have Dr. Damage a long time ago, but let it slip through the cracks because it seemed to be incorrect sometimes. I'll download it again and check out what it has to say about our spells so you don't have to respec, but it will also not be until next week that I'll have the chance to do this. My only nights off work that I have to play is Monday and Tuesday night.

Quote from: Arcdelad on December 04, 2008, 05:19:28 AM
while you are looking up that info nas, you should also find a schexy shaman avatar like illiesa...yowza!

Hehe thanks Arc. :)

Valerian


Valerian

 Nvm found it in previous post  :)

Iliessa

#8
Quote from: Iliessa on December 04, 2008, 04:22:36 AM
What makes sense at first glance would be to determine the haste required for a specific rotation. For example:

FS, LvB, CL, 3xLB, LvB, CL, 2xLB = ? Haste Rating
FS, LvB, CL, 4xLB, LvB, CL, 3xLB = ? Haste Rating


The problem with that is being inbetween Haste Ratings/Rotations is going to make up 90% of your raiding time so you will be gimped until you are done, unless you can effectively juggle your Haste Rating by including or excluding your buffs. And you'd still probably need backup equipment. Maybe that is what Blizzard is expecting us to do, I don't know. And even figuring out the numbers for this now will (might) be irrelevant as soon as Shamanism is added to our trees, and will need to be reworked again due to the difference in cast time between CL and LB.

I read up about Shamanism and its effect on Lightning Bolt damage versus the currently higher Chain Lightning DPS. They ran some numbers and found out that in order for Shamanism to scale your spell damage to the point that it equals Chain Lightnings DPS, you will need over 11K spell power. Looks like Chain Lightning will stay in our rotations.

I also found some info on the EJ forums that answers the question of Haste values for rotations similar to the ones I posted above. It's a MUCH greater gap than I thought it would be. In general, attempting to solve the rotation problem by juggling your haste is not an option. We will have to figure out what rotation is best for the amount of haste you currently have. Fortunately that has already been done: Mithrandir's graph of rotations.

Others say Haste is more linear, but shows little bumps in the road. Here is Binkenstein's Haste scaling spreadsheet. If you look at the 25% haste and then 25.5%, your dps apparently decreases by 31. You will need to get your haste to 27.5% before your DPS will increase from what it had been when you were at 25%. Here's a link to Bink's chart of Haste areas to avoid. I wish I could tell you which one was more correct. Such is theorycrafting.

Another topic they touched on was the effect of using the Priority System that Nasanna had brought up. They've determined doing so, depending on haste, would add approximately 3.69% more damage output at the cost of a 67% increase in mana consumption. In a long, stationary fight it may be difficult to keep up that type of mana cost. If you ended up having to use Thunderstorm to keep your mana pool high enough to use that rotation, the 1.5 seconds of wasted time every 45 seconds will almost nullify the benefit. That's assuming that you're not going to be standing within 12 feet of the raid boss. If it's a mobile fight though, while you're running around you can cast Thunderstorm without taking any more of a DPS hit than you already are.

Just bear in mind that this rotation is a little more complicated than only casting Chain Lightning once per 8 second cooldown. I would only use this system if you were comfortable enough with your character and the boss mechanics that you could pay attention to Lava Burst cooldown, Flame Shock debuff, Chain Lightning cooldown, Omen, and keep on eye on what the boss is doing so you can react to situations appropriately. Plus you know you're gonna be itching to look at Recount during the fight as well. Like Geos had hinted at in a previous thread, player skill has a lot to do with DPS. If you choose a rotation that causes you to consistantly make mistakes, you'll do less DPS then if you chose the easier cycle.

If you choose to go with a cycle that only has 1 CL per 8 second cooldown, then you may find Mithrandir's graph of simple rotations a little easier to decipher.

Now to figure out Critical Strike Rating and Spell Power by creating examples that prove the cake is a lie...

My idea is that Spell Power will always be a linear increase in DPS, but Critical Strike will start off slow and increase in importance as your overall damage increases, because the actual Crit damage will be higher. That statement is completely of my own idea and is not based off any math whatsoever unlike everything else I have posted so far. If Blue is running numbers on Crit, maybe he can let us know how Crit actually affects our DPS.

Iliessa

Dr. Damage numbers

Lightning Bolt






+10 Dmg: +11.4
+1% Crit: +22.5
+1% Hit: 22.5
+1% DPS (Cr|Ht|Ha|Dmg): 57.3/32.7/37/24.6
+1 Dmg (Cr|Ht|Ha): 2.33/1.33/1.5

Chain Lightning






+10 Dmg: +20.2
+1% Crit: +49.8
+1% Hit: 56.5
+1% DPS (Cr|Ht|Ha|Dmg): 57.2/28.8/37/30.7
+1 Dmg (Cr|Ht|Ha): 1.86/0.94/1.2

Flame Shock






+10 Dmg: +8.1
+1% Crit: +7.9
+1% Hit: 20.2
+1% DPSC (Cr|Ht|Ha|Dmg): 125.7/28.1/37/26.7
+1 Dmg (Cr|Ht|Ha): 4.7/1.05/1.38

Lava Burst






+10 Dmg: +5
+1% Crit: +20.2
+1% Hit: 20.2
+1% DPS (Cr|Ht|Ha|Dmg): 53.9/30.8/37/47.4
+1 Dmg (Cr|Ht|Ha): 1.14/0.65/0.78

Nasanna

Thanks for the numbers :)

So after hit capped, it looks like spell crit is the way to go in gems and enchants, on paper.

However, if you look at gems and enchants, 1% crit doesn't necessarily equal 10 damage. Since item values are not equal, I'm trying to figure this out by looking at blue quality level 80 gems, which all have the same item level/value:
[item]Quick Autumn's Glow[/item], [item]Smooth Autumn's Glow[/item], [item]Runed Scarlet Ruby[/item]

so 16 haste=16 crit=19 spellpower.

Converting to percents:
% Spell Haste at level 80 = (Haste Rating / 32.79)
% Spell Crit at level 80= (Critical Strike Rating/ 45.91)  (from http://www.wowwiki.com/Combat_rating_system)

From this, the item values convert to
0.488% haste= 0.349% crit= 19 spellpower

Converting to whole numbers, in terms of item value (not value to elemental shama):
1% haste= 0.72% crit, can't convert SP since it's not a %.

So this basically shows that with gems, even though you are getting less than 1% haste or 1% crit, the values work out such that you're getting about 0.5% haste or 0.35% crit. Even though the %crit is lower than the %haste, it looks like the crit is still better.


As nice as it is to do math on paper, I thought of a couple ways to test crit, haste spellpower in game on target dummies, without changing gems and enchants, which seems expensive/excessive. I'll test this next week when I'm done with final exams but I was thinking of something along the lines of:

Put all talent points in elemental and save 13 points for later. Put 5 points in Ancestral Knowledge for now. Don't spend the other 8 yet.
-If I can figure out how ancestral knowledge increases crit, I'll look at that too.
-Have 8 unspent talent points, DPS the target dummies for a couple minutes, figure out number of casts, dps, etc. DO NOT use flametongue weapon.
-Same as above, use flametongue weapon to see how DPS increases. Can also use flametongue totem and TOW for more trials.
-Spend 1/2/3/4/5 points in Thundering strikes, DPS for the same amount of time, figure out the same stats for each trial. This will determine to a point, how crit scales and how much it affects DPS. Additionally, try having/not having glyph of flametongue weapon if possible. Can also use TOW for more trials.
-Spend 1/2/3 points in Elemental weapons, figure out the spell damage increase on flametongue weapon (with and without flametongue glyph) and how it affects DPS.
-Finally, for haste, drop Wrath of Air to figure out how haste affects casting. Unfortunately, this is the only way I can find to increase haste.

Controls:
-DPS same target dummy (same level and next to same number of dummies so that chain lightning jumps won't matter)
-DPS for the same amount of time or number of casts
-Take off "chance on damaging spell" equipment to prevent any kind of interference
-Don't use "on use" trinkets to be consistent and get accurate numbers

Urmyboyblue

aite, if i have spent forever looking for accurate coefficients, and all that i have found are out dated, so ima have to get off my lazy but and do this one of these days and stop relying on my estimations in my head >.< rofl ill probibly do it friday after work, maybe sooner if i find some spare time. anywho...ill do em, just busy and when i get on most the time i dont wanna run numbers and think, i wanna kill stuff pew pew style! :P



Urmyboyblue

afterplayin with the dummies a little, i would have to say SP wins. i have a 15.02% chance to crit on my char screen, i was pulling an easy 20 to 25% crits with all my spells thanks to ele oth and other Crit stuff, on top of that, your Flame shock gets an incredable amount from it, which doesnt seem like much...but you would be supprised, i found the lowest numbers as far as crit chance was in my shocks, im sure with Totems and my Imp FT wep on that will change to tho, i wouldnt pass up Crit, but i wouldnt make it a priority, i will probibly run real numbers later, but i just wanted to give a heads up on the tendenciys i saw, and what it looks like to me, you will get high damage output with SP i think

Iliessa

This is the Chart from Elitest Jerks Elemental Shaman WotLK TTT.









SpellRank   Mana Cost   Range   Cast Time   Cool Down   Damage   Co-Efficient   School
Lightning Bolt1410%30(36)2.5(2)715-81571.43%Nature
Chain Lightning   826%30(36)2(1.5)6(3.5)973-111157.14%Nature
Lava Burst210%30(36)2(1.5)81192-151857.14%Fire
Flame Shock917%20(35)Instant6500 + 556/12 sec   21.40% + 40.00%   Fire
Thunderstorm40%10(12)Instant451450-1656UnknownNature

Nasanna

For these trials, I was in the Exodar, casting on the level 60 dummy to ignore the effects of hit since I am not hitcapped. I did not see any "misses" on any spells during the trials. I stood with my back to the dummy next to me in order to avoid chain lightning jumps. I casted for 2 minutes for each trial, using a priority system by casting spells in the following order: If flameshock was not on the target, I put it up, if lava burst was cool, I cast it, if chain lightning was cool, I cast it. If none of the other spells was needed, I cast lightning bolt. I did not use Thunderstorm at all. I had 2 "on use" trinkets and did not use either at all. I also start with fairly low numbers for spellpower and crit as I haven't enchanted all of my elemental gear yet. I was specced 54/5/0 for the first couple trials, and put 5 points into Thundering Strikes in enhancement for later trials, as noted below. The actual numbers from each trial are attached in a spreadsheet below if you really want to look at them or check my math.

Abbreviations:
-FT-flametongue
-LB-Lightning Bolt
-LvB-Lava Burst
-CL-Chain Lightning
-FS-Flame Shock
-FSD-Flame Shock Dot
-TP-talent points
-TOW-totem of wrath

The trials I've done so far (all used 284 haste (8.66%) mana spring, water shield, Glyph of Lava, and Glyph of Flameshock):

Spellpower trials:
Trial 1:
-12 TP left
-no totems
-no weapon imbue
-no glyph of FT
1230 SP
12.8% crit

Trial 2:
-12 TP left
-FT totem
-no weapon imbue
1350 SP
12.8% crit

Trial 3:
-12 TP left
-FT totem
-FT weapon (no FT glyph)
1561 SP
12.8% crit

Crit Trials-all include 5 points spent in Thundering Strikes (+5% crit):
Trial 4:
-7 TP left (5 spent in Thundering Strikes)
-no totems
-no weapon imbue
1206 SP (had a food buff for trial 1 that I forgot about)
17.8% crit

Trial 5:
-7 TP left
-FT weapon
-FT totem (not TOW)
1417 SP
17.8% crit

Trial 6:
-7 TP left
-FT weapon
-TOW
1697 SP
20.8% crit (including TOW)


Some math:

Spellpower:

So basically, the first 3 trials, I'm looking at the effect of spellpower. In the first trial, I have no bonus spellpower from totems or FT weapon, and I added those over the next two trials, using FT totem to avoid the extra crit from TOW. Looking at the  average damage done by each spell (% dmg increase calculated by dividing Trial 2 or Trial 3 Avg Dmg Value by Trial 1 Avg Dmg Value):

For 120 extra spell damage:
-LB sees a 9.5% dmg increase
-LvB sees a 4.9% dmg increase
-CL sees a 7.2% dmg increase
-FSd sees a 5.7% dmg increase
-FS sees a 4.0% dmg increase

For 331 extra spell damage:
-LB sees a 14.2% dmg increase
-LvB sees a 12.6% dmg increase
-CL sees a 12.8% dmg increase
-FSd sees a 13.7% dmg increase
-FS sees a 9.9% dmg increase

These numbers show that LB sees the biggest damage increase per added spellpower. This is expected since LB will have the largest coefficient due to the longest base cast time. You can also see that LvB seems to do more dmg as more spell power is acquired. This makes sense since LvB has a shorter cast time than LB but is a guaranteed crit, so it gets a larger effect as more SP is added. CL also sees a damage increase as SP is increased, but since it does not have a high crit chance and has a short cast time, it does not benefit as much.


Crit:

Trials 1 and 4 can be used to compare the effect of 5% crit. Unfortunately, I don't have multiple trials to base this off of. I also think that my small sample size since I only tested for 2 minutes on each trial may not be completely accurate for crit. Average DPS increased by 245 with the addition of 5% crit. Since the number of casts of each spell was about the same for trials 1 and 4, the total damage of each spell can be used to compare how crit affects DPS:

For 5% extra crit:
-Total LB damage increases by 5.2
-Total LvB damage remains the same (since it always crits, this makes sense)
-Total CL damage decreases (unexpected, may be due to low sample size)
-FS and FS dot remain the same (makes sense since the dot cannot crit and you are not casting FS very frequently for the initial portion of the spell to crit)


What does it all mean? (or I don't care about the math, tell me what stat to stack after I'm hit capped)

Since LB makes up the most of total damage for all of the trials (38-44% total damage done), and crit doesn't matter for LvB damage, I will consider LB for these calculations, using Trials 1,2, and 4. I will also look at total DPS using Trials 1, 2, and 4.

From Trial 1-2, damage done by each non crit LB increases by 9.5%. However, total damage done by LB stays about the same and total damage done by LvB increases by about 3%.

From Trial 1-4, LB total dmg increases by about 3% and LvB damage increased slightly.

From Trial 1-2, overall DPS increased by 245.
From Trial 1-4, overall DPS increased by 253.
These numbers are pretty similar, especially considering the small sample pool size, so I would consider 120 SP to be about equal to 5% crit in terms of DPS.

Using the conversion I talked about in my previous post, 19 SP= 0.35% crit for the same item level (ie considering you are going to be gemming or enchanting your gear for either of the two)
120 SP= 2.2% crit (from gear)

This shows that it is MUCH cheaper and easier to get more spellpower than to get crit rating and because of this, spellpower will increase your damage more than crit. However, crit also affects clearcasting, which affects your mana regen, which is not considered here.