The Mage Equivalent: The Destruction Warlock - Page 40

Tinava · 01/30/09, 10:25 AM

Originally Posted by Bosmonster

It kind of depends on your rotation as well. When do you cast Conflag? Is it whenever it is up, or for example only on a certain point in your rotation?

Saphiron isnt really a good fight to test it either btw, because you can barely use your imp (probably why its dps is missing from the WWS repots). And the difference being 20% imp damage, it is kind of hard to have any conclusion from your reports.

Personally I think F&B is still slightly better even if it were just for the +15% spellpower to immo damage. The bit of extra conflag crit just is a bonus. Adding to this that pets arent always usable and can die, I prefer the non-pet buffs as a more steady increase.

So it depends on your playstyle and personal preference mostly.

I use conflag whenever its up.

I chose sapph because it was the only way to compare with the raids I have at this time...meaning, missed out on patch this week, and WWS for KT was bugged, not showing my pet, therefore it was impossible to have a fair comparison.

Spose we could try it with maly, I realize that's not an ideal situation either.

* Maly 1/18 WWS
* Maly 1/29 WWS

EDIT: And again, the damn WWS bugged out with my pet, making comparisons bad. I assure you, he was alive the entire time up to p3.

Warlocomotif · 01/30/09, 11:44 AM

Originally Posted by Tinava

I use conflag whenever its up.

I chose sapph because it was the only way to compare with the raids I have at this time...meaning, missed out on patch this week, and WWS for KT was bugged, not showing my pet, therefore it was impossible to have a fair comparison.

Spose we could try it with maly, I realize that's not an ideal situation either.

* Maly 1/18 WWS
* Maly 1/29 WWS

EDIT: And again, the damn WWS bugged out with my pet, making comparisons bad. I assure you, he was alive the entire time up to p3.

Sorry but this is just not a good WWS to go by for the simple reason that you appear to have had some massive spellpower buffs in the 1-29 log that you don't in the other. Just look at the difference of your incinerate/chaos bolt average non-crit damage. Getting or not getting Fire and Brimstone shouldn't have any effect on that, and yet they do a ton more damage on average.

You probably were out of range of the spellpower totem in the other raid.

Dirich · 01/30/09, 1:21 PM

Originally Posted by Tinava

Actually, I was hoping for folks with math skills to use the info above for some kind of analysis on the value of F&B over not using F&B, especially as it compares to using ISL as a tradeoff in the different builds.

Anyone with said math skills around who are up to it?

Math for F&B, concerning the crit part can be found in an old post of mine http://elitistjerks.com/1053473-post303.html

And it turns out to be < 1% crit chance due to all the random factors, this is a "per talent point" value. 5% being the ideal case value for a fully specced F&B.

For immo it's easier to do the math. If I'm not reading the talent wrong, it augments by 15% Immo's spell power coefficient (20% for direct damage part and 100% for dot, before F&B).
F&B is a great talent to improve immo, since, in percentages, it gives

+0.15*X/(460+0.2*X)*100 % (direct damage part)
+0.15*X/(785+0.2*X)*100 % (dot part)
EDIT: there's an error in the previous one, but due to a wrong assumption on how F&B works, I've redone the counts in a following post.

at 2k spell power it's +34.884% for the direct damage part and +25.316% for the dot one.
The higher your spell power, the higher the benefits. With a limit fixed at 0.15/0.2 = 0.75 for both parts (for the sake of the info, impossible to reach it of course).

When I analyzed your WWS back in page 11 (http://elitistjerks.com/1050093-post257.html) the result of ISL+SL was that it awarded you 0.1568%dps.

You now have 2k spell power, unbuffed. I'll use 2k spellpower than. I know in raid you have more, but I don't know how much you had back than, but I'm pretty sure you were higher than 2k when raid buffed. If my point stands for 2k, than stand for higher values too, since I proved that the higher your spellpower, the better F&B performs on immo, which means an higher percentage of immo dps was due to F&B.

At that time it was 3.0.3, so:

1) You had no immo glyph
2) You had no conflag glyph (aka you lost part of your immo damage)

I don't know if you were specced F&B, I guess so and I'll use this hypotesys. If you weren't, all the better, since the actual percentage I'm using is lower than the one I should be using since your immo would have done more damage and thus the percentage for incinerate (used to evaluate ISL+SL contribution) should be lower than the one I used.

From my counts we know that F&B, for 2k spellpower, multiply the value of immo by 1.34884 (dd) and 1.25316 (dot).
Let's keep math easy. In both case it awards more than +25%. If it were to provide +25% than the total damage from immo you would have done, would have been 7/1.25 => 5.6% of the damage you did in the fight, which means that 1.4% = 7% - 5.6% is thanks to F&B.

Now, if we compare the values: 0.1568% dps vs 1.4% dps you see the difference is about one order of magnitudo.
Also note that:

1) You had more spellpower than what I used in my example
2) The percentage increse for immo contribution is underestimated
3) You lacked the 5th immo tick
4) You had no immo glyph (which is a good dps increase)
5) The dps increase for conflagration given by the crit increase has not been taken into account
6) If you ended the fight with more than 0 mana, than some of the mana you got from ISL was useless (aka ISL provided less dps)
7) Regardless that points 1-5 mean you got more than 1.4% out of 5/5 F&B, and that point 6 means you got less than 0.1568% from ISL, F&B still was (or would have been) on the order of 10 times more dps than 3/3 SL + 2/2 ISL.

Also note that petchwork is THE fight where ISL+SL can shine the most. ISL's value is calculated by estimating the number of Life Taps you don't need to do thanks to the mana gain ISL provides over the course of a fight, which means that in movement fights, if you can tap without losing dps (no instant spell to cast and the reason you are moving isn't going to kill you for the tap you are doing), than you cam tap without losing dps (which lowers ISL contribute due to point 6).

I don't want to get back to the discussion about the utility of ISL+SL, I think everyone understand when it's useful. We are discussing "how" useful it is, and more precisely, how useful is in comparison to F&B.
Of course a WWS isn't a proof, there's a billion things that could be argued about it, but if by underestimating EVERY single possible variable that would help F&B we still get a whole order of magnitudo of difference, I think it's highly unlikely that ISL+SL is better than F&B.

P.S.
Back at the time we were discussing about how we felt F&B to be a weak talent (and how poorly ISL was performing for us). As you can see in http://elitistjerks.com/1050664-post270.html it was exactly those thoughts that caused MD/Backdraft to be born. Point is: you need either ISL+SL or F&B to reach CB, and by just looking at our WWS we know that ISL+SL works poorly dps wise in our 25 man raids. So we are stuck with F&B or we need to go for an hybrid build (without backdraft I consider the hybrid to be primarly demonology). Which brings us back to simulations (altough I haven't seen a MD/Backdraft simulation up to now, nor a WWS report for that matter, which would be interesting for me).

Akj · 01/30/09, 1:56 PM

Originally Posted by Dirich

...

For immo it's easier to do the math. If I'm not reading the talent wrong, it augments by 15% Immo's spell power coefficient (20% for direct damage part and 100% for dot, before F&B).
F&B is a great talent to improve immo, since, in percentages, it gives

+0.15*X/(460+0.2*X)*100 % (direct damage part)
+0.15*X/(785+0.2*X)*100 % (dot part)

at 2k spell power it's +34.884% for the direct damage part and +25.316% for the dot one.
The higher your spell power, the higher the benefits. With a limit fixed at 0.15/0.2 = 0.75 for both parts (for the sake of the info, impossible to reach it of course).

...

Doesnt F&B just change the immolate coefficients from 20% & 100% for the direct dmg, dot portions to 25% & 110% respectively?

Trickykid · 01/30/09, 2:20 PM

Originally Posted by Bosmonster

MC procs off shadow spells, that is why we use CoA in the first place. Considering that ticks every 2 seconds, CoA alone theoratically cannot give 100% uptime. It will proc once every 6.67 ticks, which take 13.34 seconds. Adding Corruption would get you to 100% uptime (of course, still theoratically). With the extra T7-2 bonus and the bit of damage that corr does itself, it is good to add Corruption to the cycle (already proven in several other posts). Although corruption can be of a low priority.

Some more math: CoA + Corr gives 0.83 ticks per second, which equals a 12.45% change per second to proc MC. This means MC should proc once every 8 seconds. With 10 second duration, that should be 100% uptime. Completely ignoring RNG or actual tick-times that is

If you take the once per 3 second tick of Corruption into account, you should theoratically get a proc every 9 seconds (which equals 7 ticks)

Not quite. The uptime is equal to 100% minus the chance that all the ticks for the buff's duration don't re-proc it. In other words = 1-(.85^ticks_per_10_seconds). If you just use CoA, you get 5 chances per 10 seconds to refresh. Uptime(coa) = 1-.85^5 = 55.6%. Corruption by itself gives you 3.3 chances per 10 seconds. Uptime(corr) = 1-.85^3.33 = 41.8% With both you get 8.33 chances per 10 seconds. Uptime = 1-.85^8.33 = 74%.

Dirich · 01/30/09, 3:14 PM

Originally Posted by Akj

Doesnt F&B just change the immolate coefficients from 20% & 100% for the direct dmg, dot portions to 25% & 110% respectively?

I assume you are rethorically speaking. Since I am no confident at all in my sources (which is only the tooltip), I'll redo the counts here using your +5/+10 scheme. I spotted an error in my counts anyway, you just stopped me from editing my other post.

+0.05*X/(460+0.2*X)*100 % (direct damage part)
+0.10*X/(785+X)*100 % (dot part)

For 2k spell power numbers are:

11.627% (dd part) - limit is 25% (impossible to reach)
7.181% (dot part) - limit is 10% (impossible to reach)

If we assume 7% on both the result is that F&B dps is 2.8 times ISL+SL, which is a bit of a small margin to be "almost sure" that this result would hold regardless of the fact that I examined a specific WWS (altough there are a lot of approssimations in favor of ISL+SL).
I'll try to be a bit more precise. In a F&B immo (check the other post for details on this) dot damage is 2985, while dd is 960. Total damage is 3945, which means dot damage is 75.665% and dd is 24.335%.
Note that I'm going to multiply those percentages by the coefficients I got before.
I assume only the dd part of immo can crit, in which case a crit means the dd deals an higher percentage than what it does in a normal immo. And since the dd coefficient is better, assuming only normal hits works against F&B. If I'm wrong than nothing changes.

The 11.627% of 24.335% is 0.24335/1.11627 = 0.21800. In other words without F&B the dd would have provided 21.8% of the damage instead of the 24.335%.
7.181% of 75.665% means that without F&B the dot part would have provided 70.596% of the damage.

70.595% + 21.8% = 92.395%

Which means F&B accounted for 100% - 92.395% = 7.605% of immo's dps.
Since immo dps was 7% of the total dps, F&B provided 0.53235% dps. Which is nearly 3.4 times ISL+SL dps.

Well, if I hadn't done that error I wouldn't have posted my counts, since a 3.4 multiplier isn't what I'd consider worth of what a 10 multiplier is. Even with all the approssimations against F&B, and the fact that I didn't account for the 1% crit (or less) to Conflagrate in this comparison.

Ippocrate · 01/30/09, 3:35 PM

Originally Posted by Bosmonster

MC procs off shadow spells, that is why we use CoA in the first place. Considering that ticks every 2 seconds, CoA alone theoratically cannot give 100% uptime. It will proc once every 6.67 ticks, which take 13.34 seconds. Adding Corruption would get you to 100% uptime (of course, still theoratically). With the extra T7-2 bonus and the bit of damage that corr does itself, it is good to add Corruption to the cycle (already proven in several other posts). Although corruption can be of a low priority.

Some more math: CoA + Corr gives 0.83 ticks per second, which equals a 12.45% change per second to proc MC. This means MC should proc once every 8 seconds. With 10 second duration, that should be 100% uptime. Completely ignoring RNG or actual tick-times that is

If you take the once per 3 second tick of Corruption into account, you should theoratically get a proc every 9 seconds (which equals 7 ticks)

This is wrong. What Trickykid said is accurate.

Remember that a 100% up-time chance is not possible to begin with, since the chance to proc is lower than 100%.

It may make more sense to ask "what is the chance of having 100% up-time during a certain amount of time" (which is a different question) or "what is the chance of not having 100% uptime during a certain fight" or even (and this is probably the most popular number people seek) "what is the average up-time over a period of time", and so on.

These are all interesting statistics (and all different numbers) so it is in general useful to first decide on what is the statistic which you are interested in and then calculate it.

Not trying to make the "smart guy", it's just that people tend to report results from EJ elsewehere and if possible wrong notions should not make it out.

Snappie · 02/01/09, 5:07 PM

I was wondering about something in the openings post:

Spell Power = 1.0
Hit = 1.6-2.2 Spell Power (depending on how far away from hit starting at BC hit cap)
Haste = 0.65-0.75 (depending on gear)
Critical rating and Spirit are almost equal when specced into 3/3 Demonic Aegis from 0.4-0.5 else w/o DA it's around 0.35-0.45 for Spirit

This means Haste should be chosen over Crit. I have to admit the text is pretty vague, as you can see it depends on what level your gear is. Currently I am sitting at 298 Hasterating (9.09%) and 716 Critrating (over 25%), meaning I got plenty of both. My question is at which point a person would favor Haste over Crit (per rating), and visa verse. This all of course when one is Hitcapped and with high spellpower (think ~2000 unbuffed), and what would be the math behind it (perhaps something to add to the openings post as well?)

supplicium · 02/01/09, 5:39 PM

Originally Posted by Snappie

I was wondering about something in the openings post:

This means Haste should be chosen over Crit. I have to admit the text is pretty vague, as you can see it depends on what level your gear is. Currently I am sitting at 298 Hasterating (9.09%) and 716 Critrating (over 25%), meaning I got plenty of both. My question is at which point a person would favor Haste over Crit (per rating), and visa verse. This all of course when one is Hitcapped and with high spellpower (think ~2000 unbuffed), and what would be the math behind it (perhaps something to add to the openings post as well?)

To be fare you don't have plenty of haste, plenty of haste is 450+ You have soo much crit, it's silly. 25% unbuffed/talented crit is a silly amount of crit, that means self buffed as destruction you sit at over 34% crit (with a fire stone) with raid buffs, you have another, 53% crit or so You simply don;'t need that mich, you would benifit a lot more from haste. You can see the stat numbers on previous pages, or right on the front page.

Basicly 1 haste is rough .7 spelldmg while 1 crit is roughly .45 spelldmg. You can see where stacking haste vs crit is much better. also 1% of haste is far less stat points than 1% of crit, and often times they are the same stat worth on gear.

Atippingofthescales · 02/01/09, 5:56 PM

When factoring in the Imp's DPS, all % values are multiplicative correct?

So in assuming glyph of imp and all available destro talents for the imp (or a 20/51 spec, getting Soul Leech over FnB):

1.10(glyph of imp)*1.15(empowered imp)*1.20(unholy power)*1.30(Improved Imp) = 1.9734 additional damage from the imp. In my current spec, imp has a 1.6445 modifier. Given that in most cases in WWS logs I have, my imp contributed dps, on average, equal to my immolate (both application and DoT). Note that I am using CoA, Conflag, and Imp as glyphs.

My big question here is when does the scaling on Immolate result in a greater DPS gain from buffing your imp as much as possible? With a little micro management, the imp can stay alive in nearly all situations (baring Sapphiron and Thaddius due to how pet mechanics work for the jump). I've managed to keep him alive with little effort in most encounters that require both movement and management of where your imp is (Grob and Heigan come to mind). So he hasn't been a DPS loss in those situations.

I can provide the WWS log, but my gear during the run was absolutely horrible (started with 1347 unbuffed spellpower, ended with over 1750) and I'd prefer getting answers from a more constant environment.

Orgath · 02/01/09, 6:58 PM

Originally Posted by supplicium

To be fare you don't have plenty of haste, plenty of haste is 450+ You have soo much crit, it's silly. 25% unbuffed/talented crit is a silly amount of crit, that means self buffed as destruction you sit at over 34% crit (with a fire stone) ....

Looking at your Warlock_2_13_56_corr chart on page 14, wouldn't it make more sense to use spellstone instead?
Estimating 1/5 of damage from dots (CoA/Corr/Dot part of IM( vs 4/5 of damage beeing direct damage from your graphics my napkin math results in:
spellstone:
1/5*1% + 5/5*1.83% = 2.03% increase
firestone:
4/5*1% + 4/5*1.07% = 1.66% increase
Master Conjurer would increase the spellstone advantage even more.

supplicium · 02/01/09, 7:31 PM

Originally Posted by Orgath

Looking at your Warlock_2_13_56_corr chart on page 14, wouldn't it make more sense to use spellstone instead?
Estimating 1/5 of damage from dots (CoA/Corr/Dot part of IM( vs 4/5 of damage beeing direct damage from your graphics my napkin math results in:
spellstone:
1/5*1% + 5/5*1.83% = 2.03% increase
firestone:
4/5*1% + 4/5*1.07% = 1.66% increase
Master Conjurer would increase the spellstone advantage even more.

Don't misunderstand what I was saying as one stone is better than the other I was just going off the base that many people use firestone as destruction, as the difference is very very minimal, but the point was Just how much crit the guy was running. In theory spellstone is the greater of the 2 stones.

Azraelanna · 02/02/09, 9:39 AM

I've read the thread and don't see this, so if I've missed a post forgive me.

Is there any math on what the haste cap is taking Backdraft into account?

For example with 20.21% haste and backdraft up my CB is 1.13s.

Asmodaeus · 02/02/09, 10:57 AM

Originally Posted by Azraelanna

I've read the thread and don't see this, so if I've missed a post forgive me.

Is there any math on what the haste cap is taking Backdraft into account?

For example with 20.21% haste and backdraft up my CB is 1.13s.

Hmm, are you sure these numbers are correct? According to my calculations those CBs should be 1.16s, or you should have around 23.89% haste. With a Wrath of Air Totem and 20.21% haste you should see 1.12s CBs. I don't get it.

There are several haste caps, depending on spell and effects like Backdraft and Bloodlust/Heroism.

I would assume they would look like this (CAVEAT: the numbers you provide are not consistent with these):

Spell	No Buff	Heroism	Backdraft	Backdraft & Heroism
Immolate and Instant spells	50%	15.38%	5%	0%
Chaos Bolt	100%	53.84%	40%	7.69%
Incinerate	125%	73.08%	57.5%	21.15%

This assumes cast times are multiplied by 0.7 under Backdraft, and divided by 1.3 under Heroism.

The annoying ones are Immolate during Heroism and Incinerate during Backdraft and Heroism, since these haste values are achievable with gear alone and you can't really avoid them the way you can avoid e.g. the haste cap for CB under Heroism and Backdraft (by not casting Chaos Bolt under Backdraft when you have Heroism).
Also, Wrath of Air Totem gives 5% haste.

EDIT: retri paladin with you?

deathmancer · 02/02/09, 12:51 PM

Originally Posted by supplicium

Don't misunderstand what I was saying as one stone is better than the other I was just going off the base that many people use firestone as destruction, as the difference is very very minimal, but the point was Just how much crit the guy was running. In theory spellstone is the greater of the 2 stones.

I was curious as well and I downloaded a fresh copy of simulationcraft and duplicated the 2_13_56 spec and made one version use the Spellstone and Another use the Firestone and one version that didn't use either and these are the averages after 5000 iterations(instead of 1000):

5091 - No Spellstone or Firestone
5151 - Firestone
5140 - Spellstone

Fire stone appears to win out by a little bit. The rationalization I have for this result is that Firestone may be winning because of the mana consumption differences between haste and crit.