WotLK Talent Preview / Discussion - Page 127

turturin · 09/29/08, 9:08 PM

Originally Posted by Darian_TruBlade

What I'm wondering is if Corruption won't fall prey to scaling (again). I think it's likely that as one gears up its usefulness decreases. I'd wager it'll eventually be consigned to acting as an inferior replacement for when CoA isn't an option.

For a destro build it will almost certainly have scaling problems if only considered in isolation (i.e. total coruption dmg/(1.5+haste)).

I think the question will become:
(60% uptime)*(10% increased fire dmg)*(total avg dpct of ~16.5 sec immolate + incinerate rotation) + (corruption dpct) > (incinerate dpct assuming immolate is up)

If the expression above is true at higher gear levels, corruption will still be worth casting. Said in English: it remains to be seen if the molten core dmg contribution from corruption will be enough "corruption scaling" to make it worth casting for deep destro as gear gets better.

PyroTEK85 · 09/29/08, 9:23 PM

Corruption wouldn't be so bad off if they had stuck with their original dot changes in 2.0, where the coefficient was duration divided by 15. Instead they neutered corruption from 1.2 to .9-something and coa from 1.6 to 1.2 (correct me if I'm wrong). I thought their goal had been to make coefficients more consistant for spells, but it seems like most spells have their own set of rules nowadays.

FallenPie · 09/29/08, 9:26 PM

Ah, right...you had mentioned that concern for Corruption previously to me--for some reason it didn't register. Worst case scenario for that I'd think is that it's only worth keeping on the toolbar for especially movement heavy encounters--barring, naturally, debuff limit issues, in which case it might very well be dead in the water. Speaking here, of course, within a context where CoA is available for you to use.

I must admit also that I've not tried the sbx3 backdraft rotation since the nerf to ISB; I'd kind of (prematurely) assumed that this effectively put the purple hadoken out to pasture as far as destro was concerned--I'll have to give that a day in court later tonight. I had noticed something of a dps increase using the rotation back when it was 15 percent. Won't trouble you guys further with whatever the results are, as it's been given a rundown more than a few times over the past couple weeks here, and any increase I see gained from it isn't likely to be mindblowing. Something new to play with and nitpick at until blizzard definitively gets to tuning up lower destro tree, at least.

Fulgurite · 09/29/08, 9:28 PM

Did some DPS testing on Beta of this spec:

7/13/51
Talent Calculator - World of Warcraft
I had:
1812 SP
4.5% hit
19.89% crit (+5% destro only for 24.89% total)
0% haste
That Crit is WITH a Firestone on my weapon.

I purposely didn't spec "100% raid dps" because I wanted to see what kind of DPS we could do while still getting "helpful" or "fun" talents (Aftermath, Shadowburn, Shadowfury, Nether Protection).

I left out Conflag and Backdraft because I am not going to use them in the rotations I'm testing.

I left out Cataclysm because the Imp doesn't get that +hit, and neither does Soulshatter.

Spells used:
Corruption, Immolate, Chaos Bolt, Incinerate, various curses.

I went through ~5 minutes of solid nuking (note this negatively affects the Imp) on level 80 target dummies with different cases:

Case 1:
control = no curse. Keep up Corr and Immo, use Chaos Bolt every CD, Life Tap and Incinerate fillers
Player DPS: ~1650 (declined over the 5 minutes, I attribute this to the Imp being OOM).
Imp DPS: 250 DPS with mana, ~220 DPS at the end

Case 2:
Curse of Elements. Same rotation, but with CoE up, no Maledict.
Player DPS: ~1950 (slow decline from 2000->1920 over the 5 minutes. Started at 2100, but I attribute that to lucky crits early on.)
Imp DPS: 280 DPS with mana, ~240 DPS at the end

Conclusion: DPS was about 17% higher with CoE. Not sure why.

Repeat case 2:
Ok, seems like the first CoE test was slightly screwed. Was averaging ~1900 for the first 2 minutes, then down to ~1850 with the Imp OOM. That's only about a 14% difference.

Case 3:
Curse of Doom. Same rotation, but CoD every 60 seconds.
Player DPS: ~1850 (slow decline to ~1800)
Imp DPS: 250 DPS with mana, ~230 at end

Case 4:
Curse of Agony. So the total rotation is keep up Immo, Corr, and CoA, use CB on cooldown, and LT/Incinerate fillers.
Player DPS: ~1850 (slow decline to ~1820)
Imp DPS: 250 DPS with mana, ~220 at end

Notes:
In case 4, Molten Core uptime was ridiculous. It was about 80-90% uptime, compared to I'd estimate 50% +/-10% with case 3. Thus I was somewhat surprised at having the same DPS as with CoD. The extra GCD's from CoA come out to be almost exactly one Incinerate per minute. Calculating the DPS with an average tick of 353, comes to ~10500 per 60 seconds, which is within 500 damage of 1 CoD. So it comes down to 1 Incinerate vs. more MC uptime. With about double the uptime, that's an average 5% increase in damage. ~3900 (incinerates were 3000 normal, 6k crit) average damage every 60 seconds is 65 DPS. Which means over 1300 DPS and 5% more should beat it out, but they were about equal.

Possibly it was errors with my CoA uptime or something. It's beyond me, but mathematically CoA seems to win out. You *do* have to balance that if you use CoD, you get 2 more points in Imp. Corr, for +20% instead of +12%.

I'm considering dropping Soul Leech. It's 5 points, and I checked my mana returns for every test. Imp. SL was 7-12% (7% was really bad, 10% was average, with a high of 12%) of my mana every test. I would estimate it at like 6000 mana every 5 minutes, or 100 mp5.) The mana the Imp gets from it is laughable. He gets like 60 mana per proc. The HP return is barely noticeable in a non-solo situation. Dummies return no HP because they're capped at 1 HP, but I was still at like 40% health after 5-6 minutes of nuking. That means a warlock needs like 50 healing per second to stay topped without SL. You could drop the 5 points in that for Fire and Brimstone, or whatever else you need.

I'm estimating here, but this is my breakdown of my percent DPS from various spells:
Incinerate 50%
Chaos Bolt 17%
Curse of Agony 10%
Corruption 9%
Immolate DOT 8%
Immolate DD 6%

Curse of Doom essentially replaced CoA when it was used, though Corruption gained slightly (from the lost MC uptime), and similar numbers with no curse. I think the highest I saw Chaos Bolt was ~19.5%, lowest 15%. Highest I saw CoA was 12%, similar with Corruption, and Immolate barely changed (the DoT and DD were in the exact same positions every time).

Natasmai · 09/29/08, 9:47 PM

Originally Posted by Maels

Just did some pet testing with another warlock. Felhunter's Shadowbite doesn't benefit from other warlock's Affliction debuffs, only your own.

Succubus gets a Transform Succubus buff that does...something to it's dps. Mine doesnt' have it (pre 1.2 patch succy doesn't get it for some reason). His crit WAY more (50%).

2/2 Demonic Power with a succ out is more dps than a felhunter. 1 point in demonic power is also more dps than 1 in Malediction, thus my spec was:

Talent Calculator - World of Warcraft

Ok, I have a few questions for you Maels.

First, I am only on PTR, not Beta, so all my experience revolves around level 70. I'm at approx 1300 sp 21% crit 236 spirit and about 6.5% haste

1/ Dark pact gives me 1853 mana vs. 1647 for life tap. These numbers are completely unbuffed other than fel armor and with felhunter active. Is scaling from level 70 --> level 80 the reason everyone says dark pact sucks or is it the raid buffs that make the difference?

2/ Why did you choose suppression when a) you need the extra 3% for shadow bolts anyway, and b) you could have had points in soul siphon (or isnt Soul Drain as an "Execute" effect a valid option?) and maybe even in fel conc. (I dont know how much aoe pushback might be in level 80 instances, but in T6 instances it would seem helpful)? It seemed to me that mana was not much of an issue with affliction, is the 6% reduct a good enough reason to take the talent?

3/ I've noticed (though I dont know how to measure it) that when using the corruption glyph as destro and as affliction the nightfall effect happened far more often as affliction. I wasnt casting drain life at all, so its only proc is corruption. I am inclined to believe that the glyph and talent are in fact stacking, any certainty on this yet?

4/ Amp curse really worth a talent point? I was completely ignoring it. One talent point to lose .5 sec off a curse only seems useful if you are casting CoA, and even then it seems marginal. Now if it was off all DoT spells it would be super, but...?

5/ Someone above said that shadow embrace isnt worth the points, you maxed it out (as I have been), thoughts on the talent?

6/ Last question and a biggie, if you had to turn your afflic spec into a level 70, what would you drop? I assume that you HAVE to have haunt and 10 in destro for ISB/Bane. You obviously lose DP in destro and ruin of course which is a major hit, but I cant see not having haunt. I was aiming for something like this.... 51/0/10 Level 70

Obviously I would have thought 1/2 fel hunter and 1/1 dark pact a better choice, but answering my above question would determine that. Eradication or Deaths Embrace (or felhunter/dark pact)? I suppose fel conc isnt as good as the -4% mana and 2% hit, though I wouldnt put 3/3 in suppression unless Soul Siphon (for soul drain) is actually bad.

Thanks for any help, my time on the PTR has so far been limited to "not enough for a raid" so I have mostly solo'd stuff with an occasional heroic run tossed in there.

As far as I see it, affliction will be my level 70 choice mostly because the destro rotation and cooldown juggling just irritates me and takes away from the enjoyment I have always had playing a warlock. Even spamming one button endlessly was more fun than this current disaster.

Sebalot · 09/29/08, 10:10 PM

Originally Posted by turturin

For a destro build it will almost certainly have scaling problems if only considered in isolation (i.e. total coruption dmg/(1.5+haste)).

I think the question will become:
(60% uptime)*(10% increased fire dmg)*(total avg dpct of ~16.5 sec immolate + incinerate rotation) + (corruption dpct) > (incinerate dpct assuming immolate is up)

If the expression above is true at higher gear levels, corruption will still be worth casting. Said in English: it remains to be seen if the molten core dmg contribution from corruption will be enough "corruption scaling" to make it worth casting for deep destro as gear gets better.

You can also add whatever extra damage we get from one talent point and the utility of two points in aftermath to the latter side. Aftermath I am sure most raiders would survive without but one extra point deep in the tree could account for some dps.

Splot · 09/29/08, 10:23 PM

Originally Posted by Fulgurite

Case 2:
Curse of Elements. Same rotation, but with CoE up, no Maledict.
Player DPS: ~1950 (slow decline from 2000->1920 over the 5 minutes. Started at 2100, but I attribute that to lucky crits early on.)
Imp DPS: 280 DPS with mana, ~240 DPS at the end

Conclusion: DPS was about 17% higher with CoE. Not sure why.

Repeat case 2:
Ok, seems like the first CoE test was slightly screwed. Was averaging ~1900 for the first 2 minutes, then down to ~1850 with the Imp OOM. That's only about a 14% difference.

This confirms what I've been experiencing on the test dummies. If you go for CoE instead of CoA you wind up with 2 more points in corruption. For level 70 you could aim at 7/3/51 and just level into the remaining points for demonology.

My approach was to refresh corruption when ever it dropped to maintain MC procs, and with the glyph of corruption you could look at throwing a shadowbolt on procs (not sure how this would affect the DPS or mana usage yet). This leaves for a fairly low complexity rotation built on priorities:
Immolate > Corruption > Chaos Bolt > Incinerate > Life tap

I was trying to work out the probability of maintaining molten core (MC) with just corruption and with corruption and CoA. Once MC goes up in the following 12 seconds you have a 50% chance of refreshing MC with just corruption. With both you have 80% chance of a refresh within the following 12 seconds.

edit:
Based on the post below 80% chance for both corruption and CoA to proc MC and 50% chance for corruption to proc. Sentence asking for somebody to model removed. I miscalculated on CoA ticks orriginally as well.

Fulgurite · 09/29/08, 10:31 PM

Originally Posted by Splot

I was trying to work out the probability of maintaining molten core (MC) with just corruption and with corruption and CoA (corr+coa). Once MC goes up (by my bad maths) in the following 12 seconds you have a 50% chance of refreshing MC with just corruption. With both you have (again by my bad maths) 70% chance of a refresh within the following 12 seconds. This is where my maths fails me, I can't model the percentage up time on a rolling basis. Is there somebody who has modeled this?

Over the course of 12 seconds, you have 6 CoA proc chances and 4 Corruption proc chances. 10 chances to hit a 15% proc. That's .85^10 chance NOT to proc, or 0.197. Which means your chance at any given second to have Molten Core DOWN is 19.7%, and conversely an 80.3% chance for it to be up.

Thus, 80% uptime with CoA and Corr.

Corr only is: 1 - (.85^4) = 47.8% uptime.

Sebalot · 09/29/08, 10:33 PM

Originally Posted by Natasmai

6/ Last question and a biggie, if you had to turn your afflic spec into a level 70, what would you drop? I assume that you HAVE to have haunt and 10 in destro for ISB/Bane. You obviously lose DP in destro and ruin of course which is a major hit, but I cant see not having haunt. I was aiming for something like this.... 51/0/10 Level 70

Obviously I would have thought 1/2 fel hunter and 1/1 dark pact a better choice, but answering my above question would determine that. Eradication or Deaths Embrace (or felhunter/dark pact)? I suppose fel conc isnt as good as the -4% mana and 2% hit, though I wouldnt put 3/3 in suppression unless Soul Siphon (for soul drain) is actually bad.

Assuming Dark Pact is better than life tap at level 70 and assuming the affliction lock have to use CoE, I think http://talent.mmo-champion.com/?warl...00000000000000 is a pretty realistic build.

A fun Corruption/Siphon Life/Shadow Bolt spam build I tested was only 3-5% behind in dps and is much easier to maintain. Also since it has Ruin and there are more raid buffs giving crit and haste than buffs that help affliction, it should be even closer in a raid setting. This is assuming Glyph of Corruption is available.: http://talent.mmo-champion.com/?warl...00000000000000

What I have found is that both destruction and affliction are very close on the PTR. The worst performing (realistic) destruction build is no worse than 5% behind the very best dps build I can find (which is the 51/0/10 I pasted above). Demonology seems to lag behind, partly because of it having by far the worst mana management. No improved life tap, no improved soul leech=puny life taps. That might change at level 80, when the top end of the tree becomes viable and we have more gear with spirit.

dedmonwakeen · 09/30/08, 1:45 AM

Version r600 of SimulationCraft is available. Mostly pet updates for Warlocks and Shaman. Demonology moved up a bit, but still trails Aff and Destro significantly.

Sample output can be found here: SampleOutput - simulationcraft - Google Code

Soul · 09/30/08, 3:07 AM

Originally Posted by Fulgurite

Over the course of 12 seconds, you have 6 CoA proc chances and 4 Corruption proc chances. 10 chances to hit a 15% proc. That's .85^10 chance NOT to proc, or 0.197. Which means your chance at any given second to have Molten Core DOWN is 19.7%, and conversely an 80.3% chance for it to be up.

Thus, 80% uptime with CoA and Corr.

Corr only is: 1 - (.85^4) = 47.8% uptime.

[s]That reasoning seems flawed. Given CoA and Corruption are on the target, the probability that MC is not refreshed within 12 seconds is 19.7%. This does not imply that your expected downtime is 19.7% of the time[/s](OK, never mind, I guess it does... where are my strikethrough tags?). An alternative derivation follows...

The number of applications of shadow damage that are expected to elapse before a Molten Core proc at any given time is 1/0.15 = 6.667.

So, given that MC has just expired, the amount of downtime you'd expect is 6.667 ticks of whatever spell you're casting.

Now the duration of MC debuffs is a bit more complicated.

Let $n$ be the number of ticks in the duration of the MC debuff. We assume the MC debuff lasts for 12 seconds. If it lasts for 6, cut the following values of $n$ in half, or, better yet, forget this calculation entirely, because MC sucks. For Corruption, $n=4$ , for CoA, $n=6$ , for Corr+CoA, $n=10$ .

The probability that a given proc is followed by $n$ successive non procs is $0.85^{n}$ , which leads us to the conclusion that a given proc has a $1-0.85^{n}$ chance of NOT being the final proc in a given period of MC uptime.

Given that a proc is not the final proc in a period of MC uptime, we can calculate the expected number of ticks until the next proc. Let $m$ be the number of ticks between procs within a given period of MC uptime.

$E(m)=\frac{\displaystyle\sum_{i=0}^{n-1}i0.85^i}{\displaystyle\sum_{i=0}^{n-1}0.85^i}$

For Corruption alone, $E(m)=1.3$ , for CoA, $E(m)=2.03$ and for Corr + CoA, $E(m)=3.21$ .

Edit: Note a subtlety here: $m$ measures the number of ticks that elapse before a MC proc. The actual time that elapses is $m+1$ ticks.

Let $P_{n}$ be the number of MC procs in a given period of MC uptime. Then $P_{n}-1$ is the number of procs that are not the final proc in that particular period of MC uptime. Now, the expected number of such procs is:

$\displaystyle E(P_n-1) =\frac{\sum_{i=0}^{\infty}i(1-0.85^n)^i}{\sum_{i=0}^{\infty}(1-0.85^n)^i}=0.85^n\frac{1-0.85^n}{0.85^{2n}}=\frac{1-0.85^n}{0.85^{n}}$

For Corruption alone, we get $E(P_n-1)=0.916$ , for CoA alone, we get $E(P_n-1)=1.65$ and for Corr+CoA, we get $E(P_n-1)=4.07$

Let $t$ be the number of applications of shadow damage that are contained in a single period of MC uptime.

$\displaystyle E(t)=n+[E(m)+1]E(P_n-1)$

For Corruption, $E(t)=6.10$ , for CoA, $E(t)=11.01$ , for CoA+Corr, $E(t)=27.20$

So, the percentage uptime is just

$\displaystyle\frac{E(t)}{6.6667+E(t)}$

So, for Corruption alone, expected uptime is 48%, for CoA alone, expected uptime is 62% and for Corr+CoA, expected uptime is roughly 80% (bearing in mind we approximate the interval between ticks as being flat instead of varied like it actually is).

I hope this is comprehensible (not to mention correct).

dedmonwakeen · 09/30/08, 4:45 AM

Using a 12sec duration for Molten Core, the Destruction spec gave the following report. SimulationCraft defines uptime as the frequency with which the buff is present at the time of casting a spell affected by the buff. This is not the same thing as the actual time in seconds that the buff is present...... but perhaps a more useful metric.

Excerpt from the config file:

#
# Destruction Warlock
#
warlock=Warlock_Destruction
level=80
talents=http://talent.mmo-champion.com/?warlock=550332050000310513352303510350020000000000000000000000030000000000000000000000000
actions=flask,type=pure_death/fire_stone/fel_armor/summon_pet,imp/curse_of_doom/chaos_bolt/corruption/immolate/conflagrate/incinerate,molten_core=1/shadow_bolt/life_tap
gear_stamina=708
gear_intellect=777
gear_spirit=411
gear_spell_power=2200
gear_spell_crit_rating=461
gear_spell_hit_rating=250
gear_haste_rating=492
# Glyphs
glyph_corruption=1
glyph_imp=1
glyph_shadow_bolt=1
# Pets
pet=imp
actions=fire_bolt
blessing_of_kings=1
blessing_of_wisdom=1
sanctified_retribution=1
swift_retribution=1
quiet=1
active=owner
# Throw in a pair of Paladins.......  Eventually these static parms will be replaced with real actors.
blessing_of_kings=1
blessing_of_wisdom=1
sanctified_retribution=1
swift_retribution=1

sjogren · 09/30/08, 5:43 AM

Shadow Trance? How? Is there some way to get Shadow Trance other than Nightfall?

Imbattable · 09/30/08, 5:56 AM

Originally Posted by sjogren

Shadow Trance? How? Is there some way to get Shadow Trance other than Nightfall?

Yes, Glyph of Corruption

Maels · 09/30/08, 8:50 AM

Originally Posted by Natasmai

Ok, I have a few questions for you Maels.

First, I am only on PTR, not Beta, so all my experience revolves around level 70. I'm at approx 1300 sp 21% crit 236 spirit and about 6.5% haste

1/ Dark pact gives me 1853 mana vs. 1647 for life tap. These numbers are completely unbuffed other than fel armor and with felhunter active. Is scaling from level 70 --> level 80 the reason everyone says dark pact sucks or is it the raid buffs that make the difference?

2/ Why did you choose suppression when a) you need the extra 3% for shadow bolts anyway, and b) you could have had points in soul siphon (or isnt Soul Drain as an "Execute" effect a valid option?) and maybe even in fel conc. (I dont know how much aoe pushback might be in level 80 instances, but in T6 instances it would seem helpful)? It seemed to me that mana was not much of an issue with affliction, is the 6% reduct a good enough reason to take the talent?

3/ I've noticed (though I dont know how to measure it) that when using the corruption glyph as destro and as affliction the nightfall effect happened far more often as affliction. I wasnt casting drain life at all, so its only proc is corruption. I am inclined to believe that the glyph and talent are in fact stacking, any certainty on this yet?

4/ Amp curse really worth a talent point? I was completely ignoring it. One talent point to lose .5 sec off a curse only seems useful if you are casting CoA, and even then it seems marginal. Now if it was off all DoT spells it would be super, but...?

5/ Someone above said that shadow embrace isnt worth the points, you maxed it out (as I have been), thoughts on the talent?

6/ Last question and a biggie, if you had to turn your afflic spec into a level 70, what would you drop? I assume that you HAVE to have haunt and 10 in destro for ISB/Bane. You obviously lose DP in destro and ruin of course which is a major hit, but I cant see not having haunt. I was aiming for something like this.... 51/0/10 Level 70

Obviously I would have thought 1/2 fel hunter and 1/1 dark pact a better choice, but answering my above question would determine that. Eradication or Deaths Embrace (or felhunter/dark pact)? I suppose fel conc isnt as good as the -4% mana and 2% hit, though I wouldnt put 3/3 in suppression unless Soul Siphon (for soul drain) is actually bad.

Thanks for any help, my time on the PTR has so far been limited to "not enough for a raid" so I have mostly solo'd stuff with an occasional heroic run tossed in there.

As far as I see it, affliction will be my level 70 choice mostly because the destro rotation and cooldown juggling just irritates me and takes away from the enjoyment I have always had playing a warlock. Even spamming one button endlessly was more fun than this current disaster.

1/ It's the mass amount of spirit on gear.

2/ I took suppression for hit. Not hit capped on beta (it's a little hard without gemming for it). A Haunt resist is devastating, and +hit to just affliction is still going to raise your dps, even if your shadowbolts can miss more. I'd take the points out of Suppression as I got more hit though, the mana cost decrease is less useful than other talents.
I did the math on Drain Soul, and as I was getting closer and past 2k spell power, it wasn't worth using since shadowbolt has the benefit of ISB procs. It is a very clunky spell to use as well, since you want to ensure most of oyur dots stay up for dps. With a 15 second channel, you have to break mid cast because Haunt or Corr drop or whatever. Could be wrong about it not being worth casting.

3/ I can't tell you for sure, but it does feel like they're stacking.

4/ Amp. Curse is better than any other point you can spend at this point. You could be a dickhead and argue that it scales inversely with haste since GCD can never go lower than what you have with this talent.
But since we're not dickheads, it's worth about 12 seconds in a 5 minute fight. That's 4% more time (4% more dps.. For one talent point that's decent. I agree it doesn't practically work like that (like most theorycraft), but it's not as bad a talent as you seem to think.

5/ I didn't read that post, but they're wrong. It's around ~7% more damage for 5 talent points. Great investment.

6/ Your spec is very close and will probably be good enough. Death Embrace is more damage, point for point, than malediction. It's 4% of 35% which is 1.4% more damage total compared to malediction's 1%. Works out to 4.2% DE vs 3% Male. Plus it helps for enrages.
I haven't put too much thought into level 70 raiding specs but I guess something like:
Talent Calculator - World of Warcraft

Maybe drop 2 from ISB and slap em in Death's Embrace but remember, the slight dps gain/loss will never help you kill a boss you haven't killed already.

The talents are generally higher quality, and you'd have to try harder to pick a bad spec, so I don't think level 70 specs should be a huge concern.
We have 6 weeks left at 70, 0.7% more dps won't matter too much. Just pick a non-tard spec and go.

Fulgurite · 09/30/08, 9:41 AM

Originally Posted by Soul

[s]That reasoning seems flawed. Given CoA and Corruption are on the target, the probability that MC is not refreshed within 12 seconds is 19.7%. This does not imply that your expected downtime is 19.7% of the time[/s](OK, never mind, I guess it does... where are my strikethrough tags?). An alternative derivation follows...

Heh, don't sweat it. I was against my own logic at first too, when I was trying to figure it out. (I posted on this 5-10 pages back, with a Shadowbolt+Corruption spec.) Basically my concern was, "Ok, I can get this probability for one proc, at one particular moment, but can how can I use that to get the overall probability for a fight? Am I going to to need to do calculations with specific fight durations???"

If someone reading is having trouble swallowing our conclusions, think of it this way:

Let's take a random second of the fight. How can we know if Molten Core will be up at this second? Well if it procced at some point in the previous 12, of course. The time remaining is irrelevant. The probability for our "random second" is as was predicted earlier, (1 - (.85^10)) = 80.3%, where 10 can be changed based on your rotation and is the number of chances you have to proc Molten Core in 12 seconds; and .85 is 1 - proc chance, so these values can be altered for some other spell/duration/rotation/proc.

Anyway:

So, we have one second out of however long the fight is. But beyond the fact that we assumed there were 12 seconds before this second, there are no restrictions on "which second" that has to be. It's safe to say that having MC up in the first 12 seconds of the fight is a lower chance than the rest, but that is such a small subsection of the seconds in a fight that we can safely ignore the effect they have.

A first way to look at it is 'If I have an 80% chance for Molten Core to be on the target at any second I choose, then 80% of the time in the fight, it will be there, and I will have 80% uptime.'

You can also look at it as a sort of sum. Let's take every second barring the first 12 and do a /roll 1 100 for each. Assign that /roll number to each second. We know that any particular second has an 80% chance for Molten Core to be there, so if the /roll for a second is between 21 and 100, evaluate that as a '1' and a "Yes, MC is there" check. If it's between 1 and 20, that's a "Nope, it's down" and assign a '0'. If you sum the 1's and 0's, assuming the rolls were properly randomized, you'll get 80% 1's, and a value equal to (0.8 * fight duration in seconds excluding the first 12). Thus, 80% of the seconds of a fight, Molten Core is there, and by definition, Molten Core uptime is 80%.

Both of these methods are independent of fight duration, except of course that in practice, the longer the fight, the more likely to have a well-distributed uptime. Also, the longer the fight, the less of an effect the first 12 seconds will have on overall uptime.

Fizban · 09/30/08, 4:24 PM

Hello,

Sorry if this has been covered, but I admit I only skimmed most of the 125+ pages before this.

What are everyone's feelings as to the best "levelling" builds right now? I've tried heavy affliction on beta, and I've tried 20/41/0. Felguard is a real beast right now at 70, I can't imagine how much better it will get when I get Siphon life next level. My gear is ok, the best badge/10-man gear available right now. Any thoughts on optimal leveling builds?

Bonestorm · 09/30/08, 5:31 PM

I'm not sure if this already been answered, but do the new spell pushback mechanics affect channeled spells? Can you only get knocked 1s off of drain life for example, or is it like it is on live?

Morwen · 09/30/08, 5:55 PM

Originally Posted by Fizban

What are everyone's feelings as to the best "levelling" builds right now? I've tried heavy affliction on beta, and I've tried 20/41/0. Felguard is a real beast right now at 70, I can't imagine how much better it will get when I get Siphon life next level. My gear is ok, the best badge/10-man gear available right now. Any thoughts on optimal leveling builds?

Killing rate is quite comparable across all specs, but a major advantage of deep demonology is that you will not have to find a group to do the 3/5-man outdoor group quests because there are no elites that can hit through demonic resilience + improved health funnel + AoE avoidance.

Zed · 09/30/08, 6:16 PM

Originally Posted by Bonestorm

I'm not sure if this already been answered, but do the new spell pushback mechanics affect channeled spells? Can you only get knocked 1s off of drain life for example, or is it like it is on live?

As far as I remember it is -25/50% from the channeled duration for the first 2 incoming hits, and then immune to pushback - this is before pushback reduction effects. This would mean -2.5s max for drain life with no talents.

Sebalot · 09/30/08, 9:17 PM

New build with the following change:

* Chaos Bolt now deals Fire damage instead of Chaos damage

That should be the nail in the coffin for all full shadow/fire destruction hybrids, since Improved Shadow Bolt won't help boost Chaps Bolt damage anymore. None of the other changes seems to be of any importance or fix anything important. A huge nerf to Enslave Demon. Guess someone on Beta found a fun demon to enslave.

PSGarak · 09/30/08, 9:17 PM

Incoming changes in build 9014:

Originally Posted by MMO-Champion

Skills
-Affliction
--Curse of Doom now summons a Doomguard only if the target yields experience or honor.
--Haunt now heals you for 100%. (Down from 200%)
--Drain Soul now causes four times the normal damage if the target is at or below 25% health

-Demonology
--Demonic Circle now lasts 6 minutes.
--Inferno now summons an infernal under the command of the caster for 1 minute. (Down from 5 minutes)

-Destruction
--Chaos Bolt now deals Fire damage instead of Chaos damage
--Conflagrate can now be casted on targets afflicted by Shadowflame

Talents
-Affliction
--Everlasting Affliction doesn't affect Shadow Bolt anymore but affects Haunt instead.

-Demonology
--Demonic Pact doesn't work with enslaved demons anymore

Fimotik · 09/30/08, 10:17 PM

The problem with these Molten Core calculations is that they're based on the 12 second duration. Given the tooltip and the fact that the 12 second duration is being investigated as a bug on the PTR forums, we need to look at the tooltip being the correct duration. If that is the case, it's a horrible talent for a deep destruction build (which may be deliberate given a comment that someone made about Blizzard having intended it to be an adjunct to demo, although I haven't seen any source posted cited on this).

Even if it ends up being 12 seconds, it will still most likely become redundant at T8 and T9 gear levels with more crit and haste.

I still think Destro is a disaster in PvE terms and seems to be more structured as a PvP burst tree. Conflag was never a PvE talent/spell, and this whole business of linking it in with Backdraft and Fire & Brimstone just weakens those talents. Conflag has awful mana use and scales poorly due to its coefficient. Again, it will scale poorly with gear.

Fire & Brimstone, as has been pointed out, is a lacklustre talent. Buffing Immolate, which again scales poorly with gear due to having a DoT component and also Conflag, a talented spell with poor coefficient and mana efficiency which you'd only take if you were getting Backdraft (which has its own issues as has been pointed out with benefitting a spell which has hit the GCD).

Soul Leech is a filler (if you've ever seen WWS parses, the overheal on it is normally over 85%) and the return on Imp Soul Leech to too poor to justify spending 5 points.

Emp imp is good for the first 2 minutes of a fight until the imp goes oom (or until it gets killed).

Basically you have to waste 10 points just to get Chaos Bolt after getting Shadow and Flame, which is the last really decent talent in the tree. Is it worth it? Has destro been designed to be a PvP tree and the developers just haven't come out and said it?

My concern is, despite the nice buffs Affliction has received to improve its longevity, it will ultimately start to suffer at higher gear levels simply due to DoT mechanics, same as shadow priests. Destro in TBC was the opposite, it was weak at T4 but mid T5 started to shine. The theorycrafting on destro for higher levels doesn't seem to indicate the same trend, so unless demo shines at the top gear levels (maybe Demonic Pact will provide this boost, at least in terms of raid utility for 1 lock) then all trees suffer scaling issues. Our turn to be the mages of Wrath?

dedmonwakeen · 09/30/08, 11:29 PM

So instead of reigning in Drain Soul, they have come out publicly that it will get a 4x multiplier.

Could we get some testing regarding how the various Affliction damage multipliers are combined?

Right now, the pattern seems to be that static passive talents/glyphs are all combined additively while the more dynamic buffs (and all debuffs) are combined multiplicatively.

Check out this DPS timeline..... Drain Soul is kicking in at 20% of target health. (It will get worse once I change code to 25%)

That spike seems so crazy OP that I feel I must be modeling it wrong. If more of the Affliction talents were stacked additively (especially Soul Siphon) it might reign in the wild horses a bit.....

PsyBomb · 09/30/08, 11:34 PM

Hmmm, let's see point-by-point on the changes I can find...

-Affliction skill changes are a wash. All were either known already or pointless. Exception is Drain Soul, which got its extra-damage range boosted to 25% (from 20). Talent changes, though... I rather like that Haunt replaced Shadowbolt in EA, since it was always a much higher priority cast for Affliction spec warlocks.

-Demonology: Increased duration on the Circle is nice, though not too much of a buff. The Enslave changes (along with Infernal/Doomguard) will, for the most part, be unnoticeable (gimmick fights aside). Pact not working with Enslaved demons is another wash.

-Destruction: Conflag off of Shadowflame is nice, makes an interesting series of events for PvP. Chaos bolt dealing plain old Fire damage is something of a nerf, but it retains it's un-resistable and immunity-piercing mechanisms. Chaos damage was just scaling way too well, I'd assume.

EDIT: At this point, I'm just hoping they up and decide to either remove the cooldown on Haunt or make Drain Soul into a viable damage spell (IE it ticks for slightly more than Shadowbolt hits for, once all effects are up). If either were to happen, then we could say goodbye forever to the 5 most-used points in all of Warlock history (Bane), and go for something like 56/13/0 +2. It would make me happy