DamnDirtyApe

Idk, your simulation's priorities will inevitably lead to immolate durations of 18 seconds. At some point during the simulation, you will cast UA then corruption then immolate. Assmuing all of these hit, they will all expire at the same time. When this happens your simulated warlock will cast UA again, then corruption again, refreshing these exactly on time but delaying the immolate refresh by 3 seconds. This will continue to happen as long as all of the dots hit and siphon life does not interfere. I have seen this happen in my affliction raiding experience and my solution has been to refresh immolate early, before refreshing UA and corruption.

Of course this is moot if you find that your shadow bolt DPCT is higher than your immolate DPCT.

 

Idk

Refreshing early is not a good idea unless the DPCT of your shorter immolate is still more than the DPCT of your shadowbolt. You could potentially start UA, Co, SL, SB, SB, Immo, SB... which would make your first immolate's gap something on the order of .5s:

0: begin ua
1.5: cast ua, cast cor, set timer for 19.5, begin gcd
3: gcd ends, cast sl, begin gcd
4.5 gcd ends, begin sb
7: cast sb, begin immo
8.5: cast immo, set timer for 23.5, begin sb
11: cast sb, begin sb
13.5: cast sb, begin sb
16: cast sb, begin sb
18.5: cast sb, begin ua
19.5: ua runs out, co runs out
20: cast ua, cast co, begin gcd (.5s downtime of ua and co)
21.5: gcd up, lifetap, begin gcd
23: gcd ends, begin immo
23.5: immo ends
24.5: cast immo (1s downtime of immo)

Your point about UA, Co, Immo makes sense.. I've even brought up that very same point in one of these warlock threads. However, I think I remember that "typical" cycle getting mucked up by lag. Let me see what it looks like on the model. I'll post results here.

 

Gaborn

The best solution is to start with Immolate. That alone will desyncronize your dots.

 

Idk

This first chunk is the first 24s (plus the full summary of the 600s simulation) of a standard run of the casting model with no dot resists... priority: ua => co => sl => im. Note that Immolate is down for 2.8s after its first full run, this is even with the casting gcd of sl in between ua/co and immo:

0.0: begin ua
1.6: cast ua
1.6: cast co
3.2: cast sl
4.8: begin im
6.4: cast im
6.4: begin sb
9.0: cast sb
9.0: begin sb
11.6: cast sb
11.6: begin sb
14.2: cast sb
14.2: begin sb
16.8: cast sb
16.8: begin sb
19.4: cast sb
19.4: begin ua
19.6: end ua
19.6: end co
21.0: cast ua
21.0: cast co
21.4: end im
22.6: begin im
24.2: cast im

...

ua: 32 0.944998 0.0
co: 32 0.942332 0.0
sl: 20 0.950998 0.0
im: 34 0.842665 0.0
sb: 129 0.031
sb: 129 4
lt: 48 preemptive: 0.479167 needed: 0.520833

---

I tried a quick little hack of the cycle I mentioned in the previous post.. for the first set of casts, do a shadowbolt between the first SL and the first Immo. I ran the simulation until I got an output with no dot resists just like the previous. Note that the first immo dot refresh is almost perfect (only .2s downtime!) But by the 2nd full run of immolate (around 39s into combat) 3.4s goes by between the end of immo and when it's finally recast:

0.0: begin ua
1.6: cast ua
1.6: cast co
3.2: cast sl
4.8: begin cycle-setting sb
7.4: cast sb
7.4: begin im
9.0: cast im
9.0: begin sb
11.6: cast sb
11.6: begin sb
14.2: cast sb
14.2: begin sb
16.8: cast sb
16.8: begin sb
19.4: cast sb
19.4: begin ua
19.6: end ua
19.6: end co
21.0: cast ua
21.0: cast co
22.6: begin im
24.0: end im
24.2: cast im
24.2: begin sb
26.8: cast sb
26.8: begin sb
29.4: cast sb
29.4: begin sb
32.0: cast sb
32.0: cast tap: 1800
33.2: end sl
33.6: cast sl
35.2: begin sb
37.8: cast sb
37.8: begin ua
39.0: end ua
39.0: end co
39.2: end im
39.4: cast ua
39.4: cast co
41.0: begin im
42.6: cast im
...
ua: 32 0.944998 0.0
co: 32 0.942332 0.0
sl: 20 0.950998 0.0
im: 34 0.842665 0.0
sb: 129 0.016
sb: 129 2
lt: 48 preemptive: 0.479167 needed: 0.520833

---

Uptime of the 4 dots from the first example:
94.5%, 94.2%, 95.1%, 84.3%

Uptime of the 4 dots from the second example (the one with the cycle-setting sb):
94.5%, 94.2%, 95.1%, 84.2%

The difference is marginal assuming similar lag conditions and resist conditions (we're already talking about a resist-free situation since that's when the UA => Co => Immo cycle isn't broken). In fact, immolate's uptime is even lower in the second cycle because it starts later.

The problem is that immolate is the shortest dot of the cycle and the lowest priority.. it needs to be refreshed more often and other refreshes will always take precedence over it. You can adjust when the immolate timeline starts at the beginning of the fight but it's only a matter of time before immolate ends when another dot needs to be refreshed. You should always be making the decision to cast the highest DPCT spell that you can at any given moment (unless you can use that moment to tap favorably) and that often means letting immolate stay down while you refresh another, more important dot.

There are only a few of ways that you can actually improve immolate uptime in the middle of a fight:

1. Refresh it before it expires. This hurts the DPCT and DPM of the spell and I'm not convinced that it's ever worth it. I'd need to see the math.
2. Make it a higher priority to refresh. This is a bad idea since its DPCT is typically lower than the other 3 dots.
3. Use lifetaps/darkpacts to adjust the casting timeline so that you never stop casting but the gaps between dot end and dot refresh are smaller.

Of these, I believe that #3 is the only one that you can do without negatively affecting your overall dps.. this is because lifetaps must happen at some point in the casting cycle, but their timing is highly adjustable.

 

Idk

This is NOT true, for the same reasons as I stated above. Eventually your immolate dot will run into your ua/co dots and they'll take precedence on the refresh.

The following is a casting cycle that starts with an immolate and then follows the traditional priority system.. to keep things simple I've set it up to have no lag and no resists:

0.0: begin initial im
1.5: cast im
1.5: begin ua
3.0: cast ua
3.0: cast co
4.5: cast sl
6.0: begin sb
8.5: cast sb
8.5: begin sb
11.0: cast sb
11.0: begin sb
13.5: cast sb
13.5: begin sb
16.0: cast sb
16.0: begin im
16.5: end im
17.5: cast im
17.5: begin sb
20.0: cast sb
20.0: begin ua
21.0: end ua
21.0: end co
21.5: cast ua
21.5: cast co
23.0: begin sb
25.5: cast sb
25.5: begin sb
28.0: cast sb
28.0: begin sb
30.5: cast sb
30.5: cast tap: 1800
32.0: begin im
32.5: end im
33.5: cast im
33.5: cast tap: 3355
34.5: end sl
35.0: cast sl
36.5: begin sb
39.0: cast sb
39.0: begin ua
39.5: end ua
39.5: end co
40.5: cast ua
40.5: cast co
42.0: begin sb
44.5: cast sb
44.5: begin sb
47.0: cast sb
47.0: begin sb
48.5: end im
49.5: cast sb
49.5: begin im
51.0: cast im
51.0: begin sb
53.5: cast sb
53.5: begin sb
56.0: cast sb
56.0: begin sb
58.5: end ua
58.5: end co
58.5: cast sb
58.5: cast tap: 790
60.0: begin ua
61.5: cast ua
61.5: cast co
63.0: begin sb
65.0: end sl
65.5: cast sb
65.5: cast sl
66.0: end im
67.0: begin im
68.5: cast im

By 51s we're having 2.5s go between the end of immo and the time that it starts again.. it only gets worse as the fight goes on.

By the end, uptime of the 4 dots:
94.8%, 94.6%, 96.1%, 84.6%

Compare that to not using the initial immo.. same deal with no resists, no lag:
95.1%, 95.1%, 96.8%, 86.1%

Of course, this is a little unrealistic because it's assuming 0 lag.. just the summary results comparing the same two cycles, this time with .1s lag, no resists:

start with an immolate then do standard priorities:
93%, 93%, 95%, 85%

standard cycle:
94.5%, 94.2%, 95.1%, 84.3%

---

Of course, all of this assumes that you trust my simulator.

 

Idk

The numbers were an approximation since I didn't have the real numbers available to me. The spreadsheet tells me (and I find these to be accurate):

UA: 4168
Co: 4358
SL: 3097
Im: 2118
SB: 2929

Re Immolate: The problem is that my typical raid has no CoE nor imp scorch and always has a shadow priest. On top of this, my affliction gear has close to 200 more shadow damage than fire damage. The DPCT of immolate is 1310 vs. 1129 of shadowbolt. Under these conditions it's a DPS loss for me to use immolate.. I presume because of the increased mana cost and reduced ISB uptime?

Plugging the above damage numbers back into my calculator and simplified the format a little bit. This time we have haste rating, dps (just total damage over 600s), and change from 0% haste. For a baseline, the spreadsheet tells me that my total damage with .1s lag and 1.25s avg dot gap is 1281.60 with 0% haste.

Haste: 0.0
dps: 1303.33
change: 0.0

Haste: 15.7
dps: 1306.99
change: 1.0028

Haste: 31.4
dps: 1315.47
change: 1.0093

Haste: 47.1
dps: 1323.16
change: 1.0152

Haste: 62.8
dps: 1326.17
change: 1.0175

Haste: 78.5
dps: 1325.59
change: 1.0171

Haste: 94.2
dps: 1325.77
change: 1.0172

Haste: 109.9
dps: 1338.75
change: 1.0272

Haste: 125.6
dps: 1341.43
change: 1.0292

Haste: 141.3
dps: 1349.69
change: 1.0356

Haste: 157.0
dps: 1353.56
change: 1.0385

For grins, the spreadsheet tells me that with 10% haste my dps is 1336.17

So the spreadsheet shows 10% haste as 4.25% damage increase.. my model shows it as 3.85%. So haste may in fact be _over_valued on the spreadsheet. Still don't think I know enough to say anything conclusive about haste => damage for affliction locks.

 

Gaborn

I missed the part where you have to prioritize Immo over UA/Corr. Of course there will be a moment where all 3 dots will finish at the same time and if you cast UA/Corr before Immo you will end up with those 3 dots syncronized again.

I thought it was clear.

 

Idk

Perhaps it's confusion with my use of "they".. let me rephrase my statement. Where I originally said:

"This is NOT true, for the same reasons as I stated above. Eventually your immolate dot will run into your ua/co dots and they'll take precedence on the refresh."

It should read as:

"This is NOT true, for the same reasons as I stated above. Eventually your immolate dot will expire at the same time as your UA and corruption dots. To maximize damage when this happens, the right thing to do is to refresh UA and Corruption and then refresh immolate.. even if this leaves a several second gap of immolate downtime."

It doesn't matter when you launch your first immolate cast, over a relatively short period of time (less than a minute) your immolate will expire within a second UA and corruption.

 

Arelenda

About immolate:

makes sense to drop it, then. Weird, I figured affliction would always cast SB. I'll add something about it to the compendium.

There are a few really good items that do only shadow damage: [Ritssyn's Lost Pendant], [Orb of the Soul-Eater], [Nethervoid Cloak], [Boots of the Shifting Nightmare] and the FSW, so it makes sense that warlocks run around doing with more +shadow damage over general spellpower.


About haste for affliction:

I think you've proven that haste should be valued at less than 50% of +damage for sustained damage, for affliction locks.

Maybe it's worth using that simulator of yours to test a few other things? For example, it should be very easy for you now to figure out what dps gain you'd get with +157 spellpower, or what happens if you lose 5% hit.

 

Shai

How about adding Dr.Damage (http://www.wowace.com/wiki/DrDamage) to the tools/mods section of the first post?

It's useful for evaluating the effects of stats on spells directly in-game. Allows to manually add crit/hit etc. to the equipped gear, also scans for buffs and debuffs on mobs and applies the modifiers accordingly. It's only missing manual haste modification at the moment, something the author will hopefully add in the near future.

 

Crepe

You can enchant soulfrost over 40 damage to weapon, too, which would further increase the gap.

 

Idk

It's easier to test the spell hit than the spellpower.. mostly because the spellpower numbers are taken straight from the spreadsheet where spell hit is something that I control directly in the simulator. That said, I can still take 157 damage off the spreadsheet, get the spell damage numbers, and run the simulation.

 

Nikzor

Just a little side note a nice macro for soulstoneing its pretty similar to the Vash macro and works good


/use Master Soulstone
/script local coremsg = nil if (IsItemInRange("Master Soulstone","target") == 1) then coremsg = "SS on YOU" else coremsg = "Get closer for SS" end SendChatMessage(coremsg,"WHISPER","Orcish",UnitName("target"))

(for horde change the orcish if you are alliance)

 

Bias

Hi, I'm loving this PVE Compedium and I'll just be adding to some of the discussion. I've had this damn uncertainty that's been bugging me for so long.

It's to do with the relative effectiveness of Spell Haste, especially to it's effectiveness vs Spell Crit. This is for 0/21/40 DS/SNF LOCKS ONLY. And for experiments sake, assume casting shadowbolts infinitely. The original poster says as a rough guide under destro : Spell hit > Spell crit / spell haste / spell dmg. I'm looking for an even more accurate heirachy.

Putting item budgets aside, I just want to know what 1 spell haste rating is worth compared to 1 spell crit rating. I've looked into shadowpriest.com (SimulationCraft/Scaling/Warlock - Shadowpriest.com Wiki) and it says that with 1100+ dmg, 200 hit, 400 spell crit:
1 Spell Damage = 0.795 dps increase
1 Spell Haste Rating = 0.663 dps increase
1 Spell Crit Rating = 0.396 dps increase

Even at lower spell crits, the order remains the same - Spell damage > Spell haste > Spell crit.

Now, I've been going off on that working out which pieces of gear are an upgrade for me by simply doing multiplication of the differences of stats and seeing which item has a higher dps increase. I'd just like to know if anyone else apart from this shadowpriest.com guy has done any solid experiments to test spell haste out.

I've been reading alot of opinionated theories like spell haste makes u go oom faster = more life tap = less dps, or spell haste is crap unless u stack it (which makes a little more sense) and so forth. What I'm requesting is if anyone out there knows the true relationship between spell haste/spell dmg/spell crit.

Thanks in advance =).

 

Arelenda

I'm 0/21/40 and I base my data on ShadowSeer, which converts gained casting time into Shadow Bolt spam with Life Taps.

The haste > crit is based on the stats you have when you obtain haste gear (which used to be MH/BT). 1400 spell power, 29% crit, max hit. That gave me a pretty convincing "haste beats crit" by about 10-20%. This is including ISB.

Haste diminishes quite fast, however.

It's all relative, though. As mentioned in the guide.

 

Tbonahydraxis

First time poster and many page reader here...I'm very sorry if this topic has been covered before and if it has, please refer me to the page as I cannot find this discussion in depth.

Long story short, Our guild is still running Kara for a few people but most all are geared for the next level. My character is Tbona on hydraxis and I'm bored stiff with affliction...have been it too long. I only respec to aff when I'm needed to help with Illhoof or some no show for a dps slot.

Currently I'm 0/41/20 and love it. With my 'raiding' gear on I have 168 hit, and 24% crit, and 1216 spell damage (I'm at work, could be a smidge off. Items in possetion that may not be on armory are trinkets, spellstrike set, gladiators war staff and shifting probabilities cloak.) For Gruul's I do have a 202 hit set..just FYI

How viable is 0/41/20 for the next level (SSC, etc.)? For pet non-friendly encounters, send him in and sacrifice at appropiate time, adding 10% spell damage and 2% mana, or saccy sac.

By the spreadsheet I constantly show 0/41/20 in the running with 0/21/40 at my gear level. Anyone experienced with this? Or should I just suck it up and go with the old stand by 41/0/20>.<??? BTW we have 3 other warlocks(all affliction blends) and great healers.

Any and all advice much appreciated.

Thanks,

T

 

Arelenda

You might have demon your demon dying on pet-unsafe bosses. However, no one can deny the power of demonology on fights where it doesn't die. 0/40/21 is a very powerful alternative to classic 0/21/40, as well.

From the top of my head:

SSC:
good: Hydross, Lurker, Leotheras, Karatress (can tank Leo!)
good if 2p T5: Morogrim
bad: Vashj

TK:
good: Al'ar, VR
good if 2p T5: Kael'thas
not sure: Solarian

Hyjal: Most fights should be ok with demonology.
BT: some fights will require 2p T5, other than that it should be fine.

This is all from memory, there are good guides out there, there's one linked in the compendium and there's others on this forum. Do some research. I'm pretty sure a well played, well specced demonologist can be competitive on the meter.

 

quickquestion

Quick question (as I don't have time to browse through 29 pages atm);

As 0/21/40, is it better to consistantly spam Shadowbolt even with Curse of Elements and full Scorch debuffs in a raid environment?

(1201dmg/202hit/19.87%crit)

Or would it be better to Immo -> SB and refresh Immolate during fades?

 

Antoine

I just got my Skull the other day, and it raised a question about trinket usage. I've been using them together when doom is about to come off cooldown, but Skull's on-use effect doesn't affect the GCD for Doom. Would the ideal thing be to pop Icon, hit doom, then pop Skull and start spamming SB?

 

• frmorrison

No amount of haste or Heroism will affect the GCD.

So your question is has the right scenario in it.

 

ninielin

Hi there.

Just a quick question. for a shadow specced destruction warlock. Is it better to get illidan staff ( zhar'doom I think its spelled? ) or tempest of chaos + rage offhand? Assuming hit is capped ( well really easy to hit the cap with T6 items).

 

Afflictor

i played around with the new version of leulier last night and i was surprised by the results.

my stats were 1600 +shadow ( this fully raid buffed with a shaman totem)
27% crit, 0%haste, hit capped. also 4pt6 and CSD enabled in the advance options.

the results it gave me were that my next point of crit would be of the equivilent of 1.01 +dmg and the next point of haste 1.15 +dmg.

is this information correct? i have always been led to believe that +dmg was always the best, at this level of gear do crit and haste start to pull ahead?

Edit : 0/21/40

 

Arelenda

I suggest you read the compendium?


For your spec with your gear, I can confirm the statistics you're getting are correct. Damage about equal to crit, with haste being a tad better. I'm a bit below you and get around the same.

 

Krazen

I'm going for zhardoom, personally.

Zhardoom + Cowl of the Illidari Highlord + 4 pc t6 is the optimal gearset. Vashj robes can be substituted in, but few people will farm SSC for those.

 

Afflictor

yes zhardoom icnhes ahead of tempest +chronicle but im 99% sure 4piece then leggings of channeled elements are the best combo. when i ran my spreadsheet last night i found that illidari helm is about 9dmg equivilent above t6, channeled elements are around 20 and vashj robes with very high end gear actually pull ahead with around 22 dmg above t6 robes. pretty much a toss up between leggings and vashj robe.