rutiene

Updated to current Glyph Choices. I'll update specific math related posts later.

--Affliction (53/0/17 +1)

Life Tap
Haunt
Curse of Agony --- In the case of CoE Duty, use Corruption.

--Hybrid Demonology/Destruction (0/40/31)--

Life Tap
Conflagrate
Incinerate

--Hybrid Demonology/Destruction (0/41/30)--

Life Tap
Felguard
Incinerate

--Destruction (0/13/58 w/ ISL)--

Immolate
Incinerate
Conflagrate

--Destruction (3/13/55 w/o ISL)--

Life Tap
Incinerate
Conflagrate

--Demonology (0/56/15)--[/u]

Life Tap
Felguard
Immolate

Unstable Affliction Glyph DOES NOT reduce the GCD and is not a viable PvE option. STOP BRINGING IT UP.

 

Woggle

While I do have the same issue, I would think that the Corruption glyph is the weakest of the available glyphs for an affliction build. Maybe someone would care to do the math.

 

PSGarak

I don't get your arguement for not including spellpower. While there is no school-specific gear anymore, the spells still scale differently with spellpower, and how the glyphs interact with that is important. Immolate's worth also depends on your spell crit chance. Additionally, the DPCT is not a relevant statistic here. DPCT is only useful in determining spell priority--it's a shortcut to whether you will see a damage increase or decrease by including it in your rotation. After you've determined that both spells are damage impacts, you want to look at the raw DPS increase.
Just to back up my claim, a thought experiment: Consider two DoTs with the same DPS of 100, one with a duration of 30s, one with a duration of 60s. A glyph to increase the damage of the shorter by 20% and to increase the longer by 10% give the same raw DPCT increase of 300. But, the glyph on the shorter one is better, because it's a 20 DPS increase vs 10.

http://sig.gamerdna.com/quizzes/INFL...erks/Garak.png

 

Nicarras

I end up needing to cast CoE (no moonkin/uDK) so no CoA for me. So I went with the glyphs for things I'm casting.

Corr
Immo
SL

Really the glyphs I see varying a little due to raid comp.

Corruption isnt weak for Aff as it is still an additional means to an insta-SB. Yes DPCT is low for SB, but insta SB's are perfect.

 

blarnie

I'm assuming from the lack of discussion about Glyph of Unstable Affliction that it is considered worthless by the majority of users?

 

dcpwns

Yea been wondering this for awhile. Was hoping someone ran the numbers back in 3.0 but I couldn't find anything.Like someone said above me was wondering what three to use.I want the CoA glyph but really debating whether at of the three of corrup, sl,and immo which to would be the best.

 

rutiene

I agree with everything here. I wrote this quickly this morning, and as I was driving to school realized the exact same thing as you said about DPCT. As for spell damage, I noted at the bottom for things that would change the results:

"What about +SD? Immolate has a higher coefficient than SL, but they have the same effective casting time due to talents."

I would need to know how it interacts with the glyph the model it correctly. My argument is invalid if the glyph effects the damage after +SD, which is probably the case. But even if it effects the damage after tacking on +SD, there are multiple ways it could do so. Thus, why I didn't do it yet. Also, I just checked and apparently wowwiki didn't have the updated ranks of Immolate or SL... -facepalm-

A revision of my calculations (assuming +2000 damage and 15% Crit chance):

Immolate
Direct Damage:
before glyph - (460 base + (.2 coefficient * 2000 SD))(1 + .4 crit modifier*.15 crit chance) = 911.6
after glyph - (multiply by .9) = 820.44

DoT Damage:
before glyph -
785 base + (1 coefficient*2000 SD) = 2785
2785 * 1.2 haunt * 1.1 shadow embrace = 3676.2
after glyph - (* 1.2) = 4411.44

Total Damage:
before glyph - 4587.8
after glyph - 5231.88

Difference: 644.08/15s = 42.94DPS increase

Siphon Life
before glyph -
81*10 base + ((1 coefficient + .05 everlasting affliction) * 2000 SD) = 2910
2910 * 1.2 haunt * 1.1 shadow embrace * 1.15 shadow mastery * 1.042 deaths embrace = 4502.91

after glyph - *1.2 = 5523.49

Difference: 1020.58/30s = 34.01DPS increase

So Immolate would come out on top.

I'm going to test the validity of how I included +SD later tonight. I also didn't include any modifiers that would effect all three types of damages equally. (like Malediction) Credit to Clearly and Fallenman for the insight I gained from their damage modeling equations.

About UA: This glyph doesn't shorten the GCD, and since UA is already at 1.5s it only effects your UA cast time as much as your haste does. Which... for aff locks would not be that much since .2 off of your GCD requires quite a bit of haste.

About Corruption: That would be interesting to model, although we will need something like the ISB simulator to do it properly. It will have to interact with your rotation as well as with Nightfall. Is there even a consensus on its proc rate atm?

 

supercres

Are you sure about the UA glyph. My understanding was that it also effected the GCD of UA as well. At least in beta I believe this was true.

 

rutiene

I'll see if I have access to the Glyph and check tonight. I had read that it didn't effect GCD though.

 

Korchak

Re: Unstable Affliction, I have also read that as of one of the later beta builds the glyph of UA was having an effect on the GCD. Whether or not that is still true isn't something that I've bothered to test, mainly due to lack of initiative. I'll also try and get the glyph tonight and see what comes up, I don't currently have a glyph in my third slot yet because I'm too lazy and cheap to buy one for it right now so this is an excellent opportunity to do so.

Re: Corruption, I was under the impression that the glyph of corruption provided a second nightfall. 4% chance on corruption tick for shadow trance, just like the talent. While I have no explicit proof (ie: a log from a couple target dummy tests) this agrees with my experiences with the glyph. It stacks on top of nightfall providing an 8% chance on tick for an instant shadow bolt, even going so far as to have it overwrite itself if it procs again before it can be used up. To that end its contribution can probably be considered just like the talent and compared accordingly, not that I have any numbers for that.

 

rutiene

About Corruption:

My main concern with its interaction with Nightfall is overlap, and being able to squeeze it in on time. With two 4% chances, that doesn't mean you will get 8% chance of Nightfall. It becomes a entirely different ball game with each tick rolling 2 die, and each having a 4% chance at proccing Nightfall. I guess the mathematics behind that probability would be:

1 - (.96*.96) = 0.0784 (or 100% chance - chance of both dice rolling for no Nightfall)

So 7.84% chance for Nightfall to proc at all. But my probability is rusty, can anyone check on this?

Edit Sorry, got my calculations wrong.

 

Korchak

Well your math is off a bit, but not in the way you think. Each tick has a .04 chance to proc, not a .4 chance, the math would then put the overall chance per tick at .04*.04 + 2*(.04*.96) = .0784 or 7.84% chance per tick, which is almost 8. Further we aren't absolutly certain that it adds another roll, it could theoretically also add .04 to the chance on a single roll for corruption, which would make for an 8% chance per tick, while this is unlikely it is possible. Granted the .15% difference between the two would hardly be noticable.

 

rutiene

Lol, yeah I caught that earlier. I have to see things posted before I see that I'm being an idiot sometimes. I edited my calculations.

 

Korchak

So lets get back to the meat of the comparaison. Can we then assume based on the additional 3.84% proc chance per tick of corruption that under nominal conditions with both nightfall and the glyph that the glyph is worth 96% of what nightfall is and then compare it to other glyphs accordingly?

When I say 'nominal conditions' I mean that drain life dosn't need to be used for the duration and that no, or a minimal number of double procs have a chance to occur.

 

rutiene

I think that would be a safe assumption, however we would still need to calculate its real DPS value for a true comparison. I've never seen a calculation for the DPS increase that Nightfall actually gives, only that the DPCT value of an instant SB is higher than most spells. That is where the true trouble lies. I'm inclined to think that a Simulation would be the only possible way of seeing this...

 

Cepha

Your math is wrong, shadow embrace only applies to shadow damage but not to fire damage.

 

rutiene

Shadow Embrace effects all periodic damage.

 

Cepha

Hmm seems like german and english tooltip are not the same here -_-

 

nom

It might be possible to get a napkin estimate like this:

With Nightfall talent OR Nightfall glyph:

Each corruption tick gives a 4% chance at an instant sb. An instant sb is worth 2.5-1.5 = 1 sec of extra filler. So corruption awards 4% * 1 secs of extra filler every 3 sec tick. Extra filler time every sec is like haste. 0.04 secs / 3 secs = 1.33% haste.

I guess for 30/0/41 builds this is close to 1.33% dps, but for affliction, this is only an extra 1.33% * 0.5 = 0.67% dps or so, roughly speaking. The "0.5" relies on assumptions in 1% haste gives 0.5% affliction dps.

If the above is correct then it makes Nightfall one of the crummier affliction talents, at +0.33% dps/talent point.

 

PSGarak

Modelling the 3.84% extra nightfall chance as a 96% improvement to the talent is a highball of the effectiveness. The proc rate is exactly that, a 96% improvement, but unfortunately it raises the likelyhood of wasted procs. If you take enough time refreshing DoTs that nightfall procs again, it's a wasted proc, and the higher proc chance the more of the procs are wasted.

The thing about DPCT is a little screwy in this situation. Even if a nightfalled (nightfell? nightfallen?) shadowbolt has a higher DPCT than a DoT, I think that in general it's still higher DPS to sit on the proc and refresh the DoT. This is because the nightfall proc is a one-time event rather than a spammable spell, and the damage increase it gives doesn't depend (much) on when it's spent, while the DoT does. Basically the fact that you can sit on a nightfall proc means you should.

The tension is, these two intuitions contradict each other. Sitting on nightfall procs gives a chance to randomly waste them, and a single wasted nightfall proc is worth more than 1.5s of DoT time (2/5 of a shadowbolt vs 1/10 to 1/20 of a DoT). But, of course, you don't actually know if the nightfall proc will be wasted or not. You can only talk averages, and on average there's only a ~8% chance it will be lost (I'm assuming that your refreshes will only give corruption one time to tick before you would use nightfall anyways). That tends to lose in expectation value (~1/30 of a shadowbolt vs 1/10 to 1/20 of a DOT). So, while it is possible to double-proc and lose a proc when refreshing multiple DoTs, it's such a rare occurence that it's still advantageous to save nightfall procs than use them.

http://sig.gamerdna.com/quizzes/INFL...erks/Garak.png

 

Kalku

What I would really like to see is a list of glyphs, and the percentage change in DPSC of the spells they effect.
For example (making up numbers) Glyph of Curse of Agony : +20% DPSC.

Once we have that list, it is easy to look at your damage breakdown glyphless, and then work out which glyphs are best based on both those numbers. This will be hard/impossible for some glyphs, ie Corruption, but for most of them it should be easy, I just don't feel like doing it myself at the moment :P.

 

Havik0

This isn't entirely true, from my understanding, they stack so you just have a flat rate of ~8% rather than the regular 4% if this is true then there is no down side (other than maybe having a just plain better glyph) to having the Corr. Instant SB is very nice for Aff spec since you don't want it to take time out of your rotation of keeping your DoTs up.

This is all that I have heard on my server, is that Corr glyph is a must with Aff spec, because it does double the chance from 4% to ~8%. I may be completely wrong, but having a ~8% rather than 4% chance doesn't seem too unreasonable and so it seems realistic for them to just stack.

 

• dragon12

What evidence to you have that they stack to 8%? You're effectively saying "it's definitely 8% and NOT 7.84%", so presumably you have thousands of ticks of Corruption to back it up.

Stop thinking with your nuts and start thinking with the black and bitter ball of hatred buried in your chest

 

Neron

Hi, long time reader, first time poster here...

Being Affliction-specced myself, i am using Agony, Corruption and Shadowbolt as Glyphs...

now reading this, I find myself wondering about the effectiveness of the Shadowbolt Glyph... anybody have an idea of how to calculate it's effectiveness in comparison to the alternatives?

 

rutiene

@Havik0: I'm not saying they don't stack, but that the probabilities are not additive. From the way the glyph is worded, I'm inclined to think that it adds an extra roll to the Nightfall proc, but it is also possible that it just increases the probability of the original roll.

@Nero: The effectiveness of the Shadowbolt Glyph depends entirely on the fight. The entire point of getting it is to save mana and, effectively, GCD from Lifetaps. So the more stationary the fight, the better the glyph is. Since we have so little time to actually SB spam now as Aff locks, I think it is a safe assumption that it would not be worth it. Just a little math to back it up:

Lifetap gives me about 1.9k mana.
SB is 454 mana, so with the glyph it will save you 45.4 mana.

Therefore, 1900/45.4 = 42 SB's to save you one GCD.

Now assuming that you are able to SB twice per DoT refresh on average. That is 14 seconds per 2 SB's. Therefore it would take:

(42/2)*14 = 294s or 4 minutes 54 seconds to save one life tap on perfect rotation and on a completely stationary fight.


--

I can't seem to find a glyph of SL or Immo or UA anywhere.