SimulationCraft for Warlocks (4.0.3a numbers at 80 and 85) - Page 62

Warlocomotif · 03/23/09, 12:25 AM

You seem to be taking the time difference between spell_cast_start and spell_damage- this is wrong.
You should be going for when the spell finished casting.

The most effective way that I can think of is by casting something immediatly after you finish casting, and take that "spell_cast_start" as the "spell_cast_finished" of your spell. You can also take your current time and reduce it by your cast time.

• Zakalwe · 03/23/09, 3:49 AM

Originally Posted by Sumbish

Re: Imp being buffed by DP
As a quick hack I took the imp away from the 40/31 lock and reran the tests and it looks like DP boosts the imp by 200DPS, possibly more. This is not a completely valid test of course but serves as a rough check. Note the shaman was using searing for that test to isolate DP from flametongue.

I'm still not entirely convinced that 140SP should give the boosts it seems to, especially as the 21/50 mage has no pet to scale and is getting 400DPS. So I compared 40/31 (with Imp) to flametongue vs no flametongue (no DP/wrath) and I only get 23 DPS difference - perhaps this is the bug you found?

That was indeed the bug I found - flametongue was being reduced to 15% (or 21.6 SP) by a talent rather than being increased by 15% (to 166 SP). That'll be fixed in the next release.

• Zakalwe · 03/23/09, 4:26 AM

Originally Posted by Ruic

I find it hard to believe that not using Conflagrate below 35% is a DPS increase in 0/40/31; my Conflagrate is critting for about 15k on average. What was the DPS differential between using it and not using it?

It's not. The only thing my profile drops from the rotation below 35% is corruption.

Originally Posted by Ruic

Since Glyph of Life Tap is 20% again, are the iterations considering it used when needed again or overused to keep the spirit buff up?

They still life tap every 20 seconds - it's still a minor DPS increase.

• Zakalwe · 03/23/09, 5:43 AM

Originally Posted by Vux

So the travel time at 30 yards is indeed 1.4-1.5 seconds at 30 yards. However, I still find it odd that speed changes based on how close you are to the target. The closer you are, the slower it goes.

I don't think this is so odd - it can be easily explained by assuming there's a short delay between the cast ending and the spell actually leaving your hands.

Troffel · 03/23/09, 7:02 AM

Originally Posted by Zakalwe

I don't think this is so odd - it can be easily explained by assuming there's a short delay between the cast ending and the spell actually leaving your hands.

Regression analysis of the data gives the following equation:

Time To Hit = 0.0381 * RANGE + 0.2995 with R^2 = 0.9819

That means that the travel time is 1/0,0381 = 26,2 yard/sec and the time it needs to leave your hands is 0.2995 sec.

I think the difference is based on when the cast starts server side AND ends server side PLUS some additional lag:

UNIT_SPELLCAST_SENT-Event

Between this events the spell starts server side.

UNIT_SPELLCAST_START-Event
UNIT_SPELLCAST_SUCCEEDED-Event

Spell travels to the target.

Spell hits the target server side.

DAMAGE-Event occured

You can eliminate the network lag by queue the spell, but you can not get rid of the additional lag, which, I believe, is
some kind of serverlag. Perhaps it is caused by some priorisation of events on the server side.

The same mechanic overestimates the damage done by dots inside simulationcraft, since this additional lag occures with the SPELL_AURA_APPLIED- and SPELL_AURA_REMOVED-events as well. You are able to queue direct damage casts, but can not recast dots earlier, because you may loose the last tick.

faru · 03/23/09, 7:26 AM

Something I've been wondering. Would 0/38/33 be more dps then 0/40/31. Picking up the new pyroclasm instead of 6% crit. Obviously with low crit it wouldnt be better but raid buffed 50% crit is easily attainable and so wouldnt 6% dmg with 50% uptime be better then 56% crit?

Memnarchon · 03/23/09, 8:06 AM

Originally Posted by faru

Something I've been wondering. Would 0/38/33 be more dps then 0/40/31. Picking up the new pyroclasm instead of 6% crit. Obviously with low crit it wouldnt be better but raid buffed 50% crit is easily attainable and so wouldnt 6% dmg with 50% uptime be better then 56% crit?

By using 00/38/33, you are using only 3 talents at tier7 of talent points so it will be 1 conflag and 2 pyroclasm which they mean 4% dmg and not 6% dmg.

Warlocomotif · 03/23/09, 9:06 AM

So the travel time at 30 yards is indeed 1.4-1.5 seconds at 30 yards. However, I still find it odd that speed changes based on how close you are to the target. The closer you are, the slower it goes.

Latency will probably have to be factored in somewhere, I suspect that your combat log might record "start casting" immediatly, and wait for server side confirmation for "damage"- however thats mostly a guess.

Darkstarrz · 03/23/09, 10:23 AM

I have found that conflaging once per immolate at the end of the dots duration is leading to more overall dps. I have compared it to conflagging twice per immolate and I think that for an ideal rotation for maximizing dps that would be the proper thing to do. I am going to recount a couple attempts on the heroic training dummy on the ptr to backup my conclusion here. If anyone has any thoughts to this please dont hesitate.

trangoul · 03/23/09, 10:27 AM

Originally Posted by Darkstarrz

I have found that conflaging once per immolate at the end of the dots duration is leading to more overall dps. I have compared it to conflagging twice per immolate and I think that for an ideal rotation for maximizing dps that would be the proper thing to do. I am going to recount a couple attempts on the heroic training dummy on the ptr to backup my conclusion here. If anyone has any thoughts to this please dont hesitate.

I don't see why this would be the case. The damage per cast time of Conflag is higher than all of your other spells, so not using it on cooldown at least theoretically seems like a dps loss. That is, say conflag averages ~ 9.5k a cast, and incinerate is 7.5k a cast. Add that to the fact that conflag only uses a gcd, whereas incinerate takes 1.8-1.9 sec depending on haste, that means that conflag should be a decent ways ahead of incinerate (your filler) and should be used on every cooldown.

Sergius · 03/23/09, 10:51 AM

Question: After 3.1 will it be best (i.e. the smallest DPS loss) for mages or for warlocks to keep up the 5% crit chance debuff on bosses?

Right now, both from a fun and DPS point view, both of us locks in my guild are leaning towards a demo/destro fire build with decimate. I understand that affliction and the various hybrid builds are reasonably close in DPS however.

I got into a discussion with one of the officers yesterday (a mage) about should they have to keep up scorch in future when warlocks will be able to do it as well. I pointed out that neither of us will likely have the talent, she said that neither do the top DPS mage builds include imp scorch.

If we think about this a bit, affliction will be casting shadow bolts, and probably meta demo builds as well. Destruction and demo/destro hybrid will be heavily fire focused and are unlikely to take the imp shadow bolt talent.

With dual talent specs I guess we could /roll for who has to raid as affliction, but I for one want to do a bit less watching of DoTimer with the patch.

I guess what I am really asking is, "mages hate having to be the scorch bitch, with the new changes is it worth this becoming the warlocks job exclusively as at least two specs will apply the debuff without having to do anything extra".

wow2k6 · 03/23/09, 11:03 AM

Is it confirmed that T8 2-piece bonus works with Conflagrate?
And if so is it a huge DPS boost over 4-piece T7 with Life tap?
Is the Life Tap Glyph worth it for 0/40/31 when not using 4-piece T7 or is Incinerate better?

Troffel · 03/23/09, 11:03 AM

Originally Posted by Sergius

Question: After 3.1 will it be best (i.e. the smallest DPS loss) for mages or for warlocks to keep up the 5% crit chance debuff on bosses?
...

Short answer: The Mages should, if you do not have a Meta specced warlock in the raid.

Long answer:
Only a Meta-spec will use SB during the whole duration of a fight.

Mid-Destro-builds like (0/40/31 oder 0/30/40) and deep-Destro-builds will use Incinerate and no SB at all.

Affliction-specs will use below 25% boss health no SB anymore, because the Drain Soul will be their new filler.

Sergius · 03/23/09, 11:12 AM

I think her point was that both mages and locks will have to spec non-optimally in order to keep the debuff up, so therefore who looses the most DPS by doing it?

I can't just say "**** the rest of you, I'm fire specced now and you're my scorch bitch", it's not ideal for mages to cast scorch, atm they roll for it and one has to respec each raid. The under 25% thing is a good point, it would mean that an affliction lock would still have to cast the odd SB every 10 seconds to keep the debuff up (much like mages do now).

You are clearly of the opinion that this is a mage job (and I would like to agree). However if the maths shows that a perfectly viable lock build applies the debuff by default (affliction), and can keep it up with only minimal changes to rotation (i.e. occasionally SB after 25%); then the mages will have a good point in suggesting that they no longer have to spec for or cast scorch. It's not even a high DPCT spell like shadow bolt.

Eeks · 03/23/09, 11:19 AM

Originally Posted by Sergius

I think her point was that both mages and locks will have to spec non-optimally in order to keep the debuff up, so therefore who looses the most DPS by doing it?

I can't just say "**** the rest of you, I'm fire specced now and you're my scorch bitch", it's not ideal for mages to cast scorch, atm they roll for it and one has to respec each raid. The under 25% thing is a good point, it would mean that an affliction lock would still have to cast the odd SB every 10 seconds to keep the debuff up (much like mages do now).

You are clearly of the opinion that this is a mage job (and I would like to agree). However if the maths shows that a perfectly viable lock build applies the debuff by default (affliction), and can keep it up with only minimal changes to rotation (i.e. occasionally SB after 25%); then the mages will have a good point in suggesting that they no longer have to spec for or cast scorch. It's not even a high DPCT spell like shadow bolt.

The debuff actually lasts 30 seconds