The One Stop Elemental Shop - Page 63

Totec · 04/10/08, 11:07 PM

I noticed the stat weights have changed over on the OSES: Stats page and I was wondering what the lower leveled and Karazhan/Heroic stat weights were? I noticed this post at Daidalos's stat weights, and I was wondering if these still held true? Just want to establish that, mana usage excluded, at all raiding levels spell haste rating is always better than spell critical strike and spell damage.

SiknuII · 04/10/08, 11:31 PM

@ daler:
to fix your math you just got to change the 7.5 secs of the 3/1 rotation into: 1,5sec - haste + 6 seconds, cause the cl got benefit of the haste but the cd will not decrease.

i calculated this with the fixed time, but there is only a slight increase:

3/1 to 4/1 rotation:
breakthrough at 1,8% haste or 28,26 haste rating

3/1 to LB spam:
breakthrough at 8,2% or 128,74 haste rating

used your stats to calculate this.

tufy · 04/11/08, 3:21 AM

Originally Posted by Nizari

Is there any maths on TLC since the 2.5 hidden cool down?

The only time you'll lose a charge is when you crit 2x in a row on third charge. So that's [crit chance] x [crit chance] / 3

So, TLC calculation is:

(750 x [1- crit rate with nature spells] + 1125 x [crit rate with nature spells]) / 3 x [crit rate with your main nuke] x (1 - ([crit rate with your main nuke] x [crit rate with your main nuke] /3)) / [average bolt cast time] x [chance to hit with nature spells]

At 40% crit (of which 10 come from talents for lightnings only) at LB spam and 10% haste (note: LB spam = 2s average cast, with 10% haste that's 1,8s and with 20% LO proc chance, that's 1,5 seconds)

(750 x (1 - 0,3) + 1125 x 0,3) / 3 x 0,4 x (1- (0,4 x 0,4)/3) / 1,5 x 0,99 = 71,85

Moshne · 04/11/08, 4:07 AM

Originally Posted by Nizari

Is there any maths on TLC since the 2.5 hidden cool down?

My main concern is trinkets. I have pretty most (except Skull and Hex) and looking towards stacking damage trinkets come more haste.
If the internal cool down on Quags eye is actually 30 seconds I have it and Hex as the two best trinkets for stacking haste (excluding skull which I will never get) for shaman, with crusader close behind.

I still think Totem of Ancestral Guidance trumps the rest. But... haven't done any maths on the matter.

The cooldown on the TLC is negligible, the math hasn't changed substantially on it. Aside from the two you mentioned, its still one of the best trinkets we have available to us.

Chant · 04/11/08, 4:57 AM

Originally Posted by tufy

The only time you'll lose a charge is when you crit 2x in a row on third charge. So that's [crit chance] x [crit chance] / 3

That's assuming every third crit makes the trinket proc, which isn't the case anymore as a result of the cooldown. I think you need an equation to arrive at the correct procchance, something like

crit * (1- charge/3) = charge

which states that you get a charge if you crit and that crit wasn't preceded by a proc (which happens on every third charge). Solving for crit=40% yields approximately 35% as the chance to get a charge.

Moshne · 04/11/08, 5:09 AM

Originally Posted by Chant

That's assuming every third crit makes the trinket proc, which isn't the case anymore as a result of the cooldown. I think you need an equation to arrive at the correct procchance, something like

crit * (1- charge/3) = charge

which states that you get a charge if you crit and that crit wasn't preceded by a proc (which happens on every third charge). Solving for crit=40% yields approximately 35% as the chance to get a charge.

He has that built into the equation above, albeit with the order of operations completely missing. His method looks pretty close to identical to what I've been working at:

=(((750*(1-(Crit-0.1))+(1125*(Crit-0.1)))/((avgcast/(Crit))*3))*(1-((Crit)*(Crit)/3)))*hitchance
Where: Crit = LB/CL crit rate, and avgcast is the average of your hasted rotation, and hit chance is your chance to hit (hopefully .99.) Note that this only yields a DPS equivalent for the trinket, which isn't that useful for comparing to other trinkets, as you'll need to find some coefficient to convert into spell damage, which is a bit more controversial.

Aramund · 04/11/08, 5:54 AM

If the calculations I'm seeing over the last couple of pages are correct it seems Quag's Eye may actually be superior to the Hex Shrunken Head. I wish TLC was easier to model as I'm not sure at what point it becomes superior to the Hex Shrunken Head and therefore have no idea which of the two (if either) I should drop in favor of the Eye.

Also, wouldn't haste rating, as it stacks higher and higher, have a large effect on the uptime of the Skycall totem? Am I wrong in thinking that as you approach haste values in the 150 to 200 range that the uptime would approach if not surpass the 80% mark?

Kaideq · 04/11/08, 6:40 AM

Originally Posted by Daidalos

I disagree with a perfect LB/CL rotation being impossible. With the unavoidable haste on sunwell gear a 4LB 1 CL rotation will be required and I have never found it difficult to maintain (I already have alot of ZA haste gear and always use 4 LB 1CL). That being said. I do think that if switching to destro pots instead of mana pots for LB spam allows you to dps for the same amount of time (i.e. you have to mana pot to maintain your CLs) then LB spam is superior. I would also note that skycall is with a 60% uptime a stat weight of 1.45 DEP for haste puts it very even with Ancestral (87 vs 85).

For me the question boils down to dps and sustainability.
Highest dps to lowest (using best in slot sunwell stats and constant 80 haste battle drum rotation):
I assume 150ms between casts and calculate dps for the entire rotation by total dmg/ total time. 150ms is added to the last cast as well.
1. 4 LB 1CL using destro pots + haste drums : 2201dps : approx sustainable time with spirest and JoW 200s
2. 4 LC 1CL with mana pot +haste drums : 2170 dps: approx w/ sp and JoW 334s
3. LB spam with destro + haste drums : 2149dps : approx w/ sp and JoW 1140s
4. LB spam with mana pot + haste drums : 2115 dps: approx w/ sp and JoW infin

My dps estimates are fairly rough so don't take the dps at actual values (trinkets were just added in as spell dmg) but more for getting an idea of the dps difference between the different rotations.

I think the main difference between the 2 rotations is progression raids VS farm raids, on new content there is always the possibility of something going wrong and you needing to addapt.
My shadow priest brainfarted and died to god knows what at Brutalus 90+% yesterday, we still treat this guy as a progression raid, eg going for the mana efficient rotations, being able to cope with not getting normal amounts of mana.
By using a 4/1 rotation in encounters were stuff might go wrong to your mana battery and/or JOW bot doesn't justify the extra 50/60 dps used by burning all your mana.

Mox · 04/11/08, 8:15 AM

Originally Posted by tl.

@daler: You actually figured out when 4/1 or LB spam are better versus 3/1 without haste, while keeping same sp.dam and crit. So what you actually did is say:
I do more damage if I have 100 more haste and spam LB then using 3/1 with same stats. Where are you going to get 100 haste for free? :P

After reading the OP and just about just about the whole thread this is something I've been trying to wrap my head around.

The stat weights posted in the OP assume that you have X spell dmg and Y crit, 1000 and 40% respectively. So if I understand it right, once you reach these 'baseline' figures haste becomes very good but only if you can add haste while maintaining the other two? The reason I ask is using those stat weights you get tons of gear being thrown up thats just haste and spell dmg so by using that gear your crit is going to get nuked to hell down to like 30% which just doesn't seem right to me.

Sorry if I'm missing a very obvious point, only just switched to elemental. Been enhancement for along time.

tl. · 04/11/08, 8:33 AM

Are TLC, Eye and Hex Head actually the best trinkets for us? Where does Sextant go with all this?
I can get myself around stat weights and shit on other gear slots, but this trinkets just confuse me.

Niarb · 04/11/08, 9:03 AM

Originally Posted by tufy

Same as you, Evovi. I consistently did less damage than I should have according to spreadsheets yesterday and when checking stuff out, I noticed consistently lower crit. And it wasn't just any random crit rate, it was ALWAYS around 34%, no more, no less.

been noticing the same thing as well, either rather incredibly unlucky rng or I just dont get it..

KrietoR · 04/11/08, 9:35 AM

Originally Posted by Moshne

He has that built into the equation above, albeit with the order of operations completely missing. His method looks pretty close to identical to what I've been working at:

=(((750*(1-(Crit-0.1))+(1125*(Crit-0.1)))/((avgcast/(Crit))*3))*(1-((Crit)*(Crit)/3)))*hitchance
Where: Crit = LB/CL crit rate, and avgcast is the average of your hasted rotation, and hit chance is your chance to hit (hopefully .99.) Note that this only yields a DPS equivalent for the trinket, which isn't that useful for comparing to other trinkets, as you'll need to find some coefficient to convert into spell damage, which is a bit more controversial.

with ur current haste/crit/spelldmg/trinket choices, how much dps are u pushing out? lets say on brutallus fight, or Dr. Boom (raidbuffed)?

post a WWS if possible

thx in advanced

Daidalos · 04/11/08, 11:13 AM

Originally Posted by innocent

I'v been running a few test on these a part of the afternoon with this gear:

No buff,
no totem,
no pots,
all test running 2 minutes,
spam lightningbolt rank 12,
target = Dr Boom.

Skycaller:
1269dps
1289dps
1289dps
1302dps

Ancestral:
1322dps
1281dps
1251dps
1246dps

Edit:
Shadenbonus = spellpower
Kritisch = guess what? Spell critical
Tempowertung = Spell haste
(at least i hope so, my german is terrible)

So the average of skycall LB spam was:
1287.25

And the average of ancestral LB spam was:
1275

This shows skycall with a small lead. Lets assume that skycall had a 66% uptime 66*1.45 would be 95.7 DEP putting it a good deal ahead of ancestral. I think it comes down to raid situations vs testing situations for the strength of skycall. In test situations and theorycrafting it seems a good deal better than ancestral however in inpractice it may not always be so. I think its certainly something to consider. No longer mana potting and switching to destro pots and using skycall might be the highest sustainable dps many people can manage. I'm still at a loss to explain why some people are reporting a dps loss using a LB CL rotation and others are reporting a gain. I am wondering if it is due to latency but I'm really not sure.

• Daler · 04/11/08, 11:41 AM

Originally Posted by Mox

After reading the OP and just about just about the whole thread this is something I've been trying to wrap my head around.

The stat weights posted in the OP assume that you have X spell dmg and Y crit, 1000 and 40% respectively. So if I understand it right, once you reach these 'baseline' figures haste becomes very good but only if you can add haste while maintaining the other two? The reason I ask is using those stat weights you get tons of gear being thrown up thats just haste and spell dmg so by using that gear your crit is going to get nuked to hell down to like 30% which just doesn't seem right to me.

Sorry if I'm missing a very obvious point, only just switched to elemental. Been enhancement for along time.

This is exactly why I'm trying to do the math for myself. There's got to be a sweet spot given my gear where I can trade X crit/dmg for Y haste and come out on top. I just haven't quite figured it out yet. I may be reduced to plain plug and chug with multiple item sets and Bink's spreadsheet.

Thanks for the corrections. Trying to find my error was driving me up a wall.

Moshne · 04/11/08, 12:45 PM

Originally Posted by KrietoR

with ur current haste/crit/spelldmg/trinket choices, how much dps are u pushing out? lets say on brutallus fight, or Dr. Boom (raidbuffed)?

post a WWS if possible

thx in advanced

My guild has not yet reached Brutallus, we are 5/5, 7/9. As such, I don't have a Brutallus parse. Here are the two Gorefiend parses I could find that I didn't get SoD.

Wow Web Stats
Wow Web Stats

I have gotten a couple upgrades since these, but the numbers are fairly correct. I may go do a Dr. Boom test one of these days if Gorefiend isn't friendly to me this week.

Daidalos · 04/11/08, 12:47 PM

Originally Posted by Mox

After reading the OP and just about just about the whole thread this is something I've been trying to wrap my head around.

The stat weights posted in the OP assume that you have X spell dmg and Y crit, 1000 and 40% respectively. So if I understand it right, once you reach these 'baseline' figures haste becomes very good but only if you can add haste while maintaining the other two? The reason I ask is using those stat weights you get tons of gear being thrown up thats just haste and spell dmg so by using that gear your crit is going to get nuked to hell down to like 30% which just doesn't seem right to me.

Sorry if I'm missing a very obvious point, only just switched to elemental. Been enhancement for along time.

No haste gets good before that. I simply picked stats that many people like going for. The 40% crit is mostly for the mana not the dmg gained from it. Also I didn't pick entry level stats for haste since at the time all haste was from za or epic badge gear. Now that it seems bliz is introducing haste on blues in Magisters Terrace and I assume it will be on greens and blues for lich king I may provide some entry level haste values.

For example with only 999 lightning spell dmg (including the 55 dmg relic) 0 haste and 30% crit (including talents) for LB spam
1 crit rating = .65 spell dmg
1 haste rating = 1.12 spell dmg

CL is worth more due to it getting less from its spell coef than haste and crit however haste is still a bigger dps gain the crit.
for CL with the same stats
1 crit rating = .82 spell dmg
1 haste rating = 1.41 spell dmg

Crit is mostly good for the mana return. Point for point haste is far superior in pure dps. Figuring out your dps time if you need the mana return etc is up to you. These things can all be modeled however is too difficult to tell everyone what to do which is why I make spreadsheets for this and show how to do the math.

Moshne · 04/11/08, 12:51 PM

Originally Posted by tl.

Are TLC, Eye and Hex Head actually the best trinkets for us? Where does Sextant go with all this?
I can get myself around stat weights and shit on other gear slots, but this trinkets just confuse me.

The list, in order, that I have come up with is as such:

The Skull of Gul'dan
Hex Shrunken Head
Quagmirran's Eye
Darkmoon Card: Crusade
The Lightning Capacitor
Sextant of Unstable Currents
Eye of Magtheridon
Icon of the Silver Crescent

The Sextant is just a weak trinket, too long of a CD, too unpredictable of a proc. These "rankings" were arrived at with 1525 raid buffed damage, 35% crit, and roughly 18% haste. Bumping the crit up and the haste down doesn't show much difference, other than the Quag's Eye moves around depending on your haste value (long story short, the Quag's Eye buff can be pseudo-extended in its duration from a 6 second proc, to nearly a 7.5 second proc depending on how much haste, your latency and your rotation due to the mechanics of haste being calculated only at the beginning of the cast, if you start a cast at 5.8, 5.9 on the Quag buff, you get that plus whatever your cast time of "buff time," which magnifies the value of the trinket substantially.)

innocent · 04/11/08, 4:14 PM

Using Daidalos`s (hight lvl) stats value:
1 crit = 0.65 spellpower
1 haste = 1.22 spellpower

Skull (ignoring ToHit*) = 55 + (29.17)x1.22 = 90.59
Hex = 53 + 35.17 = 88.17
Crusade = 80
Icon of the Silver Crescent = 43 + 25.9 = 68.9
Battlemaster's Alacrity = 40x1.22 = 48.8

If someone had enough knowledge to give an equivalence for Quag, TLC and Sextant it would be amazing.
I think its quite interesting to mention the 63spellpower trinket from alchemy that great for potolholics.

the Quag's Eye buff can be pseudo-extended in its duration

The same way Skycaller's proc can be extended. That might be the reason why it gives very good results on testing versus Ancestral.

*Actually you might not have to ignore it wearing some very particular outset gear

draconis0101 · 04/11/08, 4:18 PM

Has anyone noticed that the shaman badge gear in 2.4 is leaning towards haste with no spell crit while the cloth is high spell crit and no haste? Are they trying to lean shamans off of crit and go more for a machine gun approach?

Daidalos · 04/11/08, 4:45 PM

Originally Posted by draconis0101

Has anyone noticed that the shaman badge gear in 2.4 is leaning towards haste with no spell crit while the cloth is high spell crit and no haste? Are they trying to lean shamans off of crit and go more for a machine gun approach?

Maybe they read my stat weights and figured out haste is better than crit!!

Originally Posted by innocent

If someone had enough knowledge to give an equivalence for Quag, TLC and Sextant it would be amazing.
I think its quite interesting to mention the 63spellpower trinket from alchemy that great for potolholics.

I am still unsatisfied with all of the quags eye modeling. Everytime we think we have some hard numbers on it people post something different about it.

TCL and sextant have been covered to death the math on it is very solid. Just search for it.

Putting all of this info in a single place would be a good idea though. I think my next little project will be adding in better modeling for trinkets since this seems to be a major problem. Bink's sham stats does a pretty good job at this already though.

Aramund · 04/11/08, 4:55 PM

In order to push the duration of the Quag's Eye buff to its maximum duration you'd need 139 passive haste rating + the Skycall Totem proc + the Quag's Eye proc. That would push your cast time to around 1.475 seconds, making it possible to push the "duration" of the proc out to the 7.4 second range, though obviously it would leave little room for latency or slow casting. Obviously the Quag's Eye benefits from having the Skycall Totem and visa versa, though its quite unpredictable.

draconis0101 · 04/11/08, 4:57 PM

Well, I know this isn't mathematical, or hard numbers, but last night, now that I'm running some haste, we did za. I was averaging 1350 dps over my normal of 1138-1200 due to the eye. Thats in conjunction with my haste chest as well but it definately made a difference now that it is proc'ing more often. I've been touchy bout TLC but since I see so many of you suggesting it and backing it up, I'm going to give it a go next time it drops.

Moshne · 04/11/08, 5:56 PM

I'm pretty satisfied with my Quag's eye numbers, the only thing I'm not solid on is the actual hidden CD on it. I've been estimating high to avoid overselling the trinket. I need to readjust my Skull numbers as well, as I haven't updated that for the "extended duration" problem. I haven't worried about the Skull numbers as it will only go up with the adjustment, and have been more concerned with the "middle" trinkets.

I've been trying to help Bink get ShamStats up and running, and the models should all be included in it once it is finished. But when that is is up to Bink.

Daidalos · 04/11/08, 6:15 PM

Originally Posted by Moshne

I'm pretty satisfied with my Quag's eye numbers, the only thing I'm not solid on is the actual hidden CD on it. I've been estimating high to avoid overselling the trinket. I need to readjust my Skull numbers as well, as I haven't updated that for the "extended duration" problem. I haven't worried about the Skull numbers as it will only go up with the adjustment, and have been more concerned with the "middle" trinkets.

I've been trying to help Bink get ShamStats up and running, and the models should all be included in it once it is finished. But when that is is up to Bink.

The issue with quags seems to be the a possible undervalueing due to already hasted effects aplifiing its usefullness by giving more hasted casts than a average of droc duration / average proc time would seem to indicate.

Moshne · 04/11/08, 6:21 PM

Originally Posted by Daidalos

The issue with quags seems to be the a possible undervalueing due to already hasted effects aplifiing its usefullness by giving most hasted casts than a average of droc duration / average proc time would seem to indicate.

I agree wholeheartedly, despite that, I think it is easy to overvalue it.

(ROUNDUP(6/Quaghastedcast))*Quaghastedcast) / (45+(1/((1/avgcast)*0.10)))

This gets you an average haste value for the trinket, using your average cast for the proc rate, and the QuagHaste for the proc duration. This does not include any "extra" haste like Skycall, as its impossible to model, and the numbers seems to be showing the other totem as better anyhow.

In my gear, this gives about 83 spell damage equivalent when converted from haste to damage equivalent. I use:

Spell Damage = 1
1 Crit rating = 0.857318064
1 Haste Rating = 1.270297572