Enhancement Cage Match: Two specs enter, One spec leaves - Page 2

Malan · 01/07/08, 11:35 PM

Oh I don't know, I suppose our standard 10,000 hours is fine. I'm not saying anyone needs to grind those through repeatedly, just that if people are going to start talking about % increases and the like that we really need a 'base case' for everyone to establish a common understanding with.

• Binkenstein · 01/07/08, 11:51 PM

If you're talking base stats, try running one without any ele/resto talents for a baseline, so rather than seeing the difference, we see the DPS gain from each tree.

Katrael · 01/08/08, 12:02 AM

Originally Posted by Malan

Let us make some assumptions shall we?
Lets take a moderate amount of AP/Crit say 1500 AP, 30% crit. For the Elem/Enhance test, assume *exactly* enough hit rating to cap yellow attacks, no more. For the resto side, assume *exactly* the 9% from talents, and no more. No haste, no -armor. Assume that with a 6 sec shock CD you can max the #of shocks in 2 minutes, and same with a 5 sec CD. Assume Battleshout, Kings, Might, food and a flask, all the normal spell debuffs. Uncheck the box for "someone steals my SS debuff." Simulate that, and lets examine those results.

It occurs to me that it won't really be a fair showing of the differances if we give the ele spec sufficient hit rating to cap but we don't give the same hit rating to resto, because there will be a disparity with the gear in our two tests.

Edit- Going to go with Binkenstein's suggestion and run the sim with only the standard enhancement talents as a control, then see about ele and resto.
I'm altering Malan's parameters slightly so that both specs will have 48 hit rating. That's very slightly over 9% with only DW specilization. I'm also going to assume a pair of season two weapons for the MH/OH DPS and speed inputs. Is it likely that including mongoose will muddy the test results?

Malan · 01/08/08, 12:48 AM

Yah Bink has the right idea, thats what I was trying to say as well, that we need a *base case* to start with.

Is it likely that including mongoose will muddy the test results?

Not that I would think if all 3 cases included it, could always try it with no weapon enchants and no trinket procs of any sort as well.

Katrael · 01/08/08, 1:06 AM

Okay, I was running the control with bad armor values, so I'm tossing it out and starting over. Think 7700 armor minus five sunders and Faerie fire sounds good?

• Binkenstein · 01/08/08, 1:17 AM

Less variables = good.
Throwing extra effects into the mix will just muddy the waters, so to speak, and when doing algebra you can get rid of like factors on both sides to simplify the equation.

Resto/Enh + procs + other stuff vs Ele/Enh + procs + other stuff. So we don't need to consider gear questions like enchants/procs.

PSGarak · 01/08/08, 1:27 AM

Unless the procs happen to favor one side or the other. Would resto's 3% melee hit give a detectable edge to on-hit melee procs?

Katrael · 01/08/08, 1:37 AM

Originally Posted by PSGarak

Unless the procs happen to favor one side or the other. Would resto's 3% melee hit give a detectable edge to on-hit melee procs?

I see this as two ways. On the one hand we're trying to just work up a basic set of numbers from which to calculate the differances of the two specs, and thus more variables like on hit procs muddy things. But on the other hand, something as basic as mongoose on your weapons isn't really going to vary between the specs, so if we don't include it in the control as well we may end up undervaluing the extra hit% that resto provides.

As an aside, I just finished working up a nice set of numbers based on the control, mongoose included, when I realized I hadn't checked off Leader of the Pack. So starting over again. It's worth noting that in the ten sets of 10k hours that I ran with LotP on, eight returned a DPS of 1229, with two yielding 1231. So, seems like there's very little variance in the output.

Edit- Finished up running the sim over again without LotP. There was a bit more variance this time, but not much. I did 15 iterations of 10,000 hours. Input used was 1500 attack power, 30% crit, 9% hit, a pair of 97.5DPS 2.6 speed weapons with mongoose, no spellhit, the default 5% spellcrit. Boss armor was set to 4490(7700 armor boss with 5xSunder and FF). Buffs used were Might, Kings, Mark of the Wild, Flask, Haste Potion, 20 strength food, Battle Shout and SoE.

Averaged DPS was 1177.8
The fifteen individual tests returned the following: 1177, 1177, 1177, 1179, 1177, 1178, 1179, 1177, 1179, 1179, 1178, 1178, 1177, 1178, 1177

Illundai · 01/08/08, 3:42 AM

I'm just confused as to why I got higher results yesterday evening. I ran it a couple of times now and I get about 150 dps less from each spec

Weird, just ignore my previous results then :P. Might have input something funky somewhere while being tired...

Mulgero · 01/08/08, 4:04 AM

Have to consider also that these are simulator based values for pretty much optimum situation. However to get most out of 1sec faster shocks it requires you to be able to shock every 5sec, we got GCD kicking on pretty often, waiting WF cd before using stormstrike, re-dropping totems and if we talk about raids Binks shock rotation won't work due raid eating least second SS debuff. Even with 6sec CD I'm not always able to shock right away if I want to use SS debuff myself or I'm about to wait SS free'd for use.

3% melee/spell hit won't suffer if you can't shock every 5sec. Ele spec is better when you have to move and are out of range so can get more damage with shocks and fire totem.

If simulations will end up showing about same or 10-40 more dps for ele then due things above it'll be waste to spec for faster shocks in my opinion.

Melthar · 01/08/08, 4:36 AM

Originally Posted by Malan

For the Elem/Enhance test, assume *exactly* enough hit rating to cap yellow attacks, no more. For the resto side, assume *exactly* the 9% from talents, and no more.

Aren't these exactly the same amount of +hit? (9% chance to miss on specials on bosses) hence the elem build is essentially getting 3% +hit from gear for "free" to bring it up to the same 9% that resto gets?

Katrael · 01/08/08, 5:48 AM

Originally Posted by Melthar

Aren't these exactly the same amount of +hit? (9% chance to miss on specials on bosses) hence the elem build is essentially getting 3% +hit from gear for "free" to bring it up to the same 9% that resto gets?

This is correct. The resto sim runs should include the 3% hit from gear as well, to be a fair comparison.

testthewest · 01/08/08, 7:56 AM

I already posted this in a wrong thread, so here the repost. I wanted to keep things simple, so mistakes are easier to point out.

"I didnt find any satisfying reasons as to why Ele/Enh is worse than Resto/Enh.

Asuming just from the Enhance tree w/o any other points spend:
white dmg 40%
WF 35%
SS 15%
Shocks 10%

Restotree: +3 hit to melee/spells:
40% * 1,03 = 41,2% (the hit will increase the white dmg for about 3%)
35% * (1+(0,8*0,03) = 35,84% (about every fifth hit is from SS, so we have 2,4% more hits to procc WF)
15% = 15% (SS is capped w/o restotree)
10% * 1,03 = 10,3% (Shocks will hit 3% more)

41,2+35,84+15+10,3= 102,34% compared to nothing but Enh-tree

Eletree: +5% shockdmg, *20% faster CD
90% will be unchanced
10% *1,05*1,2=12,6

90% +12,6= 102,6% compared to only Enh-tree.

If you factor crit in both examples (which I'm to lazy for, since point is proven for me now), the impact of hit will lessen even more, since in a one-roll system it wont raise the dps by one 1%.

Please prove me wrong, or help me to find a post in this overly huge thread which shows me better math (than my bad try at it)."

Malan · 01/08/08, 9:41 AM

Originally Posted by Katrael

Edit- Finished up running the sim over again without LotP. There was a bit more variance this time, but not much. I did 15 iterations of 10,000 hours. Input used was 1500 attack power, 30% crit, 9% hit, a pair of 97.5DPS 2.6 speed weapons with mongoose, no spellhit, the default 5% spellcrit. Boss armor was set to 4490(7700 armor boss with 5xSunder and FF). Buffs used were Might, Kings, Mark of the Wild, Flask, Haste Potion, 20 strength food, Battle Shout and SoE.

Averaged DPS was 1177.8
The fifteen individual tests returned the following: 1177, 1177, 1177, 1179, 1177, 1178, 1179, 1177, 1179, 1179, 1178, 1178, 1177, 1178, 1177

Ok, lets use this as our baseline then. Did you happen to write down the breakdowns of AutoAttack/WF/SS/Shock percentages of total damage?

tuurgan · 01/08/08, 12:38 PM

This is silly. Losing the hit will hurt you more than upping the damage of your totems or your shocks....shocks are like 15% of your damage tops...TOPS!

Just run recap or recount or something you dont need to run some sims to show which is better. Shaman threads already state 'capping yellow dmg is all you need' which is way wrong already. Stop this insanity and get some sleep; please.

Toots Hepcat · 01/08/08, 12:42 PM

Originally Posted by PSGarak

Unless the procs happen to favor one side or the other. Would resto's 3% melee hit give a detectable edge to on-hit melee procs?

I would not have thought so. However, many of the simulations I've run are showing less benefit from elemental talents than predicted and one reason for that could be that the reduced proc rate of on-hits.

Indeed, with the default stats/buffs, shocks at 20/24 and my procs (WF/WF, Exec MH, Mongood OH, Dragonstrike, Hourglass), Resto (1195) is roughly equivalent to Elemental (1186 + 10 from 15% searing).

Incidentally, when did Yo add in the dps of the lowly searing totem? I love that wooden imp.

Malan · 01/08/08, 12:59 PM

Originally Posted by tuurgan

This is silly. Losing the hit will hurt you more than upping the damage of your totems or your shocks....shocks are like 15% of your damage tops...TOPS!

Just run recap or recount or something you dont need to run some sims to show which is better. Shaman threads already state 'capping yellow dmg is all you need' which is way wrong already. Stop this insanity and get some sleep; please.

You seem to have completely missed the point of this thread and the fact that we specifically have asked not to post this kind of garbage. You've also given yourself away to being misinformed.

Toots Hepcat · 01/08/08, 1:11 PM

Originally Posted by tuurgan

This is silly. Losing the hit will hurt you more than upping the damage of your totems or your shocks....shocks are like 15% of your damage tops...TOPS!

Just run recap or recount or something you dont need to run some sims to show which is better. Shaman threads already state 'capping yellow dmg is all you need' which is way wrong already. Stop this insanity and get some sleep; please.

This is the sort of ignorance this thread was created to clarify. Shocks are only 15% of your damage BECAUSE YOU ARE RESTO. Once you go elemental, they are substantially more than 15%.

Let's say you do 1000dps. You have more than 12% to-hit. Resto spec.

50% of your damage is white damage (500), 15% is shock damage (150) for a total of 650.

We abandon NG. 500 dps is reduced by 1.03, 150 dps is reduced by 1.03. 485 dps and 145 dps.

Now we activate Concussion -- 5% additional shock damage -- and Reverberation, 20% faster (and thus one assumes 20% MORE) shocks. 145 dps is first increased by 5% to 152.25 then by 20% to 182.

485 + 182 = 664, > 650 without even considering the affect on fire totems. Shocks are now ~18% of our damage, which has increased by 1.4%.

Keep in mind:
- 50% is a high estimate for white damage.
- Removing NG is not a 3% loss of white damage. It's actually less, as crits are unaffected by its removal.
- Stormstrike is unaffected by losing hit.
- On-crit procs are unaffected by losing hit
- Windfury is not directly affected, but the lag between procs outside of the cooldown may suffer slightly due to 3% fewer hits.
- On-hit procs are affected by losing hit, and will have a decreased uptime. Depending on how many of these you have -- and some shaman have a lot -- you may see a flat or even a slight negative by switching to Elemental.
- You probably won't want to totem twist with an elemental offspec. It's expensive and frenetic.
- 11.6% higher damage from Searing Totem is a nice bonus.
- To see an increase in dps demands shocking more times in a minute than you do now. If your play style or lag will not permit more procs per minute, your DPS will go down.
- Maintaining a 6s shock cycle with a 10s stormstrike cycle means you will enter a race condition fairly often, meaning one or the other of SS or a shock will need to wait. A 5s Reverberation cycle, however, can be optimized so that you never enter a race condition. This is harder to codify but should not be neglected -- you will find it easier to shock more and will never have to decide whether to shock or stormstrike.

• sordee · 01/08/08, 2:09 PM

Originally Posted by Toots Hepcat

Now we activate Concussion -- 5% additional shock damage -- and Reverberation, 20% faster (and thus one assumes 20% MORE) shocks. 145 dps is first increased by 5% to 152.25 then by 20% to 182.

Answering one of my earlier questions, yes Yo's Simulator does properly handle the clipping that a 5s shock cycle will introduce.

Using the base stats (ie. default):

20 shocks per 2 minutes, No Reverberation : 199DPS (shock damage)

Theoretical
24 Shocks with Reverberation and not counting clipping (i.e. ASSUME +20% damage) = 199 * 1.2 = 239DPS

Actual:
24 shocks per 2 minutes, Reverberation: 216DPS

Net: In the optimal case (24 shock per two minutes) Reverberation is only a net 8% DPS Increase because of clipping.

Malan · 01/08/08, 2:36 PM

Ok Katrael's baseline (Enhance talents only) test was 1177 DPS on average.

I just ran the same stats (1500 AP, 30% Crit, S2 wpns w/ mongoose, 9% hit, 0% spell hit, kings/might/motw/haste pots/flask/food/shout) this time selecting Reverbration and Concussion and setting shocks to 24 per 2 min. (I did not select drums, Katrael didn't mention them either way so I'm assuming his baseline test assumed no drums)

My results for the Elemental sub spec at 10,000 hours averaged 1205 DPS, we can call that a 30 DPS increase over the base.
Shocks were on average 20% of total damage in these runs.

Toots Hepcat · 01/08/08, 2:40 PM

So the ideal rotation, then, would be FS + SS + ES + ES ...

Pitbuller · 01/08/08, 2:47 PM

Fs/es vs. fs/es/es

24 shocks per 2 minutes, Reverberation fs/es: 216DPS

8 shocks per 2 minutes fs only: 68DPS
16 shocks per 2minutes es only 118DPS
total:
24 shocks per 2minute fs/es/es 186DPS

Even if you can get SS debuff for every es its only +22dps and total 208DPS.
Even without scorch + CoE and with SS c fs/es is better than fs/es/es. 204dps vs 194dps

Edit: numbers dont include conclusion but its change nothing for results.

• Daler · 01/08/08, 5:47 PM

It might be worthwhile to note that Alliance shaman have a slightly different baseline case that can't really be avoided (i.e., no racial ability to trigger since it's passive). I'll run 15 tests at 10k hours using Katrael's numbers and post my results shortly to set an Alliance baseline.

One thing I noticed using Yo's sim, though, is that CoF isn't available as an option. Should we add that in manually as an average increase to Searing Totem DPS when calculating the Enh/Elem totals?

EDIT: Baseline testing done. 15 runs, average DPS = 1182

• Daler · 01/08/08, 8:22 PM

Originally Posted by Malan

My results for the Elemental sub spec at 10,000 hours averaged 1205 DPS, we can call that a 30 DPS increase over the base.

Using alliance shaman, my results for the Resto sub spec at 10k hours averaged 1213 DPS on an 1182 baseline, a 31 DPS increase.

My results for the Elemental sub spec averaged 1213 DPS as well. Shocks averaged in at ~20% of total DPS with 254, Searing Totem contributed ~76 DPS on average.

However, it does not appear that Yo's sim allows for including the 15% searing totem increase from spec'ing 16/45/0. Adding that 15% increase to the Searing Totem DPS portion for the Elemental sub spec yielded a total average DPS of 1224, bringing the Elem DPS increase to 42, or 3.56%, over the baseline enhancement talents alone.

If Yo's sim does automatically include the damage increase from CoF when you check off concussion and/or reverberation, then the two seem to be in a dead heat for total damage from my initial results for alliance shaman.

Beowolf · 01/10/08, 2:22 AM

They ironic thing is that I was discussing this in detail not to long ago on the WoW Shaman forums. It quickly degenerated into a flame fest, most directed at me as I'm in the Ele/Enhance camp, but a few gems came out of this!

I generally used 1500 DPS, 30% crit, 12%/9% hit (on a side not, Ele/Enhance is even better for Draenei due to the 1% spell hit aura). My own back of the napkin calculations came up with a ~1.5% dps increase, including totems, when using Ele/Enhance over Resto/Enhance. I used the following ratio:

45% White
30% WF
10% SS
15% Shocks

Resto/Enhance

675 White DPS
450 WF DPS
150 SS DPS
225 Shock DPS
------
1500 DPS

Totem DPS: 156.07

Ele/Enhance

655 White DPS 675*.97
443 WF DPS 450*.985 (I used a 1.5% less WF damage due to the 3% loss in white hits)
150 SS DPS
251 Shock DPS 225 * 1.1479 * .97 (Concussion and Reverb mathematically came out to ~12.02% increase to shock damage while using a FS/ES/FrS rotation)
-------
1499 DPS

Totem DPS: 173.22 (+17.15)
Net Gain: 16.15 DPS (~1.08%)

Of course, if we dont lose any WF DPS, its around a 23.15 DPS increase over all, or ~1.54% overall dps increase. Now I know this is a different way of looking at the two specs, but I don't believe its necessarily a bad one. Thoughts?