Kalku

I don't understand this point about crit having diminishing returns. It was true when factoring ISB uptime was important, but I don't see how the 40th % crit you gain is any worse than the first. Each % is the same absolute DPS benefit.
Unless what you really mean is that the more crit you have, the more valuable the next point of spell power is?

 

Darian_TruBlade

What they may mean is something more like this.

If you do 100 DPS but never Crit then adding 10% Crit adds 10 DPS, a straight up 10% DPS increase. However, adding another 10% crit adds 10 DPS, but as you were doing 110 DPS before it is only a .91% DPS increase.

It's basically an issue for those who look at 1% Crit and presume that means a 1% DPS increase. The absolute worth is identical.

"A man's IQ, yearly income, sexual prowess, ingenuity, physical appearance and generally every other aspect of his character can be condensed down to four digits: his Arena rating." - Zechsy [70 Rogue - Skullcrusher (EU) - 10/23/2007]

 

Fulgurite

As a Destro Lock before and after 3.0, I've long used this philosophy as a reason to avoid crit on all gear that doesn't have multiple stats on it (Hood of Hexing, Belt of Blasting, etc). However, something struck me tonight as I was discussing the worth of crit to various classes (with my Warrior and Mage friends). It's true that 1% haste is 1% faster casting on all our spells, and is generally close to 1% DPS increase (although now that we have DOTs we have to cast, those only gain 1% filler DPS). However, it's the old argument that people made against Spell Haste when it was introduced, that means 1% more mana used.

Is there any consensus on what the DPS loss of 1% more Life Taps over the course of a fight does to the worth of 1% haste? It's obviously still *some* DPS increase, just as 1% more crit is still a DPS increase when you have 99% crit, just a very small one.

Considering it, two big factors are how much you life tap for and how much mana per second you spend. It actually gets simplified when you consider DOTs, because they result in 1% more filler damage, so 1% haste is 1% more incinerates, and 1% more mana cost of incinerate per second. The base mana cost is mitigated by points in Imp. Soul Leech, which as a static proc chance should be easy enough to calculate, given a ballpark mana pool.

I assume a spreadsheet could calculate this exactly given your gear stats. Do you guys figure this into your general weighting of stats though? With both crit and haste having drawbacks that make 1% more of the stat worth less than 1% DPS, how much better is Spell Power? Could it be advantageous for us to simply stack Spell Power gems in Wrath, even as Destro?

 

Soul

With a 25% crit rate (reasonable for an Affliction Lock in a raid?), Imp. Shadowbolt improves Shadowbolt and Haunt dps by between 1.15 and 1.35% per point. MC boosts Immolate by 6.5% for the first point and 2.4% for the second point. I'm rather dubious as to the value of Imp. Shadowbolt because the gains from Imp. Shadowbolt are very paltry, boosting the lowest DPCT spells in the Affliction Warlock's arsenal.

 

Troffel

I beg your pardon, but this is not true:

Assume we have a direct damage spell, which is able to crit, and doing non-crit 100 Damage. If it crits, it does the double damage (like the full skilled talent Ruin). The spell has a 10% crit chance.

If you cast 10 spells in a row. The counts of crits a distributes as an binominal distribution, so the average you do
100 * 10 + (10%*10*100) = 1100 damage to the target. Increase it with 10% you will do 1100 * ( 1 + 10% ) = 1210 damage to the target.

Now compare it with 10% more crit. You do on average 100 * 10 + ((10% + 10%) * 10 *100) = 1200 damage to the target.

Conclusion: 1% dps gain is more worth than 1% crit gain.

Another important difference is usually neglected, the calculation is based ON AVERAGE. In reality, you have to consider dominance-effects. The crits are binomial distributed and on a limited length of a fight, you may end with quite very different result in damage, if the randomness is on your side or not.

I think everyone experienced the following, on a bossfight you played very well, do a perfect chaincasting, but in the result, only far behind in the damage statistics.

Why? If you look at the reports, you see that your count of critical hits are very low compared to your displayed crit chance.

This phenomena is enlarged, since Blizzard uses a clustered random source. Even is equally distributed, high values from the random source gives a better chance of a high next random number, a low random number gives you a better chance of getting a next random number, which is low. (I am not sure, if the clustered random source is still used by Blizzard.)

 

Woggle

I don't want to interrupt the discussion, but I thought it might be worth mentioning this. I disagree that casting Haunt whenever the cooldown is up is the way to go as affliction. Rather, Haunt must be dealt with in a more "sophisticated" manner. Haunt by itself certainly is not the highest DPCT spell, it gains its importance by boosting the damage of other dots. That means, its value increases greatly in the timeframe of 'cooldown over' and 'falling off the mob', being lower than ALL other dots at the beginning and rising to the top over this very small gap of about 2.5 to 3 seconds depending on haste. Posting an example might help on this argumentation:
Let's assume the cooldown of haunt will be over at time 0 (t=0s). Depending on travel time, it will fall off the mob 4-5 seconds afterwards, but you will want to have finished the cast at exactly 4 seconds after the cooldown is up (which equals t=4s), to account for the travel time (on a sidenote, casting it up to a second late MAY NOT be a dps-decrease depending on when your dots tick). On the other side, UA is about to fall off at t=1.5s. Let's say you finish casting, a shadowbolt at t=0.5s. Haunts cooldown is over, but you have 3.5s left to refresh it. UA is about to fall off and it has a higher DPCT in this timewindow, so you want to prioritize UA over Haunt. On the other side, Let's assume UA is falling off at t=3.5s and you finish casting your last cast at t=2s. Now, you will want to prioritize Haunt.

My point is, even something simple as 'cast Haunt whenever it is up' does not hold true for any given case. Affliction's spellcasting queue is rather influenced by a constant flow of data and information, which I myself find hard to process. I tried a 56/0/5 spec in sunwell last night and despite being able to top the meters on both M'uru and Kil'Jaeden (was tank on Twins and only got in for Felmyst without having set up the interface for the new spec.. Had Incinerate binded where Shadowbolt was supposed to be and Siphon Life not even in the bars at the start of the fight, wws: Wow Web Stats , and my gear is not at all optimized for affliction and worse than that of all other damage dealers in the raid except for maybe the moonkin and the elemental shaman, who had both respecced from heal), I found it VERY awkward to maintain the dots, especially on KJ, where even more input comes from your environment and your flow of casting constantly gets messed up by the big bombs, (our dps was very low due to massive amounts of disconnects over the fight) meteors, and so on. Additionally, I am raiding with a latency of over 300ms, which caused me to sometimes cast immolate and UA twice during bloodlust.

Despite all the hassle, I think that nothing has to be massively changed in the affliction tree. I personally think that it is ok to put a large personal effort into a spec which then allows you to pull a high amount of dps. While I do see the problem that other classes (say, mages) do not have this possibility, I am saying this from a Warlock's point of view. I guess no-one wants the 0/21/40 afk-shadowboltspamspec back and as Blizzard has stated, player skill (though one might discuss if there is anything like 'skill' in wow, I would say yes, and even if it is only reaction time) should somewhat influence dps, so that a highly complex (I guess people will disagree with me here) spec as affliction, which requires a great amount of attention and is hard to perfect, should be able to stand on or close to the top of the meters if performed correctly.

Reducing the complexity might have to include drawbacks on the dps, which I personally would regret. After all, wow itself does not pose too many challenges to a player.

 

Pentamorfi

No, with Sunwell gear expect at least a 40% raid buffed crit chance for an affliction lock, possibly more.

Shadowbolt and Haunt are 50% of your DPS. Immolate is 10% (and that's a rather generous estimate).

 

dedmonwakeen

Perhaps part of the reason Aff seems to be performing better in simulation than in live is because of issues like this.

SimulationCraft provides an option to Haunt that will prevent it from being cast unless the debuff will wear off in 1.5*haste() seconds.

http://code.google.com/p/simulationcraft

 

tetracycloide

This is the wrong way to evaluate talents. DPCT is the metric that an affliction lock is looking at when establishing a casting priority and is the metric used to eliminate DoTs from the rotation. When evaluating talents the DPS boost, not the DPCT boost, of the talents is what's important. Not taking imp SB because SB is the lowest DPCT spell for an affliction lock is extreamly poor reasoning.

My vanity is justified.

 

Troffel

This condition I use in my addon RippedAdvancedWarlockStrategy as well. It seems that a possible dot-tick without haunt-debuff is less fatal than having a haunt early and skipa shadow bolt as the filler spell at this slot in time.

 

Woggle

Unless I'm misunderstanding you, I have to disagree here too. Haunt should provide higher DPCT than shadow bolt and therefore should at all times be prioritized above shadow bolt.

edit: well, scaling may change this issue. I don't know at which point shadow bolt will outdamage haunt, maybe someone else might care to elaborate.

 

Darian_TruBlade

It is true, you just misunderstood me.

What I was saying was that some people see +1% Crit and think, "Aha! 1% more DPS!" which is incorrect so long as you have non-zero Crit. However, the absolute worth of 1% Crit is always the same no matter how much Crit you have (ignoring ISB-like mechanics).

"A man's IQ, yearly income, sexual prowess, ingenuity, physical appearance and generally every other aspect of his character can be condensed down to four digits: his Arena rating." - Zechsy [70 Rogue - Skullcrusher (EU) - 10/23/2007]

 

Medu

At 80 SB will easily out scale haunt due to ruin. The amount of crit casters get in a raid envirnoment is a little silly. SB already has the advantage of ISB and bane at 70.

 

Woggle

I can't follow this argumentation, Haunt benefits from ISB and Ruin just as shadow bolt does. I do see that shadow bolt's coefficient is higher proportionally due to Bane. Once shadow bolt has superior scaling, ISB and Ruin of course push it further towards shadow bolt. But Haunt has higher base damage, so this point is not reached at 0 spelldamage, that is what i meant. Now, watching the output of SimCraft, Haunt's dpct is higher than that of shadow bolt in Naxx-gear.

edit: never mind the last sentence. I'm wrong.

I might also add that I confused the threads for WOTLK und live, so I was stating from a live-server point of view. At level 80, what was posted above certainly is true.

 

Natasmai

Haunt does not benefit from Ruin, it is not a destruction spell. I still have a hard time believing that Haunt is not strictly better at L70 than casting a shadowbolt. Shadowbolts at 70 as 56/0/5 are just terrible, the best filler we have for sure, but still terrible. At L80 with enough points for Ruin and ISB maybe its an arguement.

 

kaib

So, what are you going to level with? As I see it there are two options:
1. Deep Affliction and taking some demon talents as you advance.
2. 21 Affl to Siphon Life and then rest in demon and use Felguard from 71 onwards.

Mobs in the first areas have 10.000 hpish which rules out destro imo, three nukes are required to kill a mob way too often, which makes it very inefficient.
On beta I used the 2nd spec and was quite happy with it. Between Dem Resilience and imp Health Funnel I could solo every group quest including five-men tagged elite NPCs. If you have good gear, any mob will die from Siphon Life/CoA/Corr while the Felguard just tags along on defensive and helps out.

Now when I leveled on beta, the Haunt healing effect was crap, which made Affliction a lot worse then it is now. Between using Haunt on CD and an occasional SL/Drain Life, Affliction can pull a similar amount of mobs but has higher dps. Felguard also makes up for a decent bit of the lost damage, especially the cleave is quite powerful.

So in the end it comes down to higher dps output & cast time spells vs more survivability along with more solo-group-quest-capabilities. Plus affl/demon also offers more protection against getting ganked, although in general the guy jumping usually wins anyway, no matter if you are affliction or demon and you should beat when jumping someone else with either spec as well.

I'm pretty much torn between the two specs, would be nice to hear what others are going with and why.

 

Afterglow

As somebody levelling on a PvE server, it's definitely going to be deep affliction for me, at least for the first few levels. Just about zero downtime (thanks to Dark Pact, Haunt, and Siphon life).

And you can't get much faster than most mobs dying in around 3 GCD's - just put up the 3 instant-cast DoTs, hit tab, and start dotting up the next target. Felhunter on defensive, and throw in an occasional haunt/tap/pact as neccessary, and you can keep going almost indefinitely.

I only levelled to 71 on the beta - but it certainly worked nicely in the starting areas. If it gets tougher later, as my lv70 gear becomes less effective, I'll probably reconsider speccing for a Felguard.

 

Woggle

D'oh. How could I forget that. Well, at level 70 I will stick to Haunt over shadow bolt. At level 80, it will be even harder to decide what to cast between Haunt's cooldown and it falling off..

On the topic of leveling: I guess for 'pure' power-leveling, meaning only caring about xp, the full-affliction spec seems to be the better choice, something between 51/10/0 and 56/5/0 to start out with and adding more points to demo while leveling, as ISB/bane are not necessary when tab-dotting. Elite quests are not needed while leveling (besides that level 2-3 player quests are easily solo-able, especially when fear works) and can be done for gold at 80. That builds lacks some survavibilty, but I guess master summoner can be picked up for some sort of flexibility.

 

Juised

Hi, I'm a long time reader, but this is my first actual post.

Responding to the debate about the best time to cast Haunt, I believe that waiting till 12-1.5*haste() to cast nor immediatly after the CD is up is the best use of the spell. Rather one must estimate about 8 to 12-1.5*haste() seconds into the future to determine if you will be able to recast it then. Granted, this is difficult for a simulation, such as SimulationCraft to model. As a suggestion, when I created a quick simulation to try and get a feel for the flow of Affliction, (just a quick little thing that would evaluate spells in a DPCT order and display which to cast) I included 2 different conditions for casting Haunt. Highest priority was Duration-1.5*Haunt(), to recast if the buff was going to fall off. But then I included another condition that would cast Haunt any time the CD was up, and prioritized that just above Shadow Bolt at the bottom of the priority queue. While far from a perfect model (still might cause issues with Haunt expiring at the same time as other DoTs) I'm pretty sure that it is a better method of modeling when to cast Haunt rather than just when it comes off CD or when it is about to expire.

P.S. Hi Guys!

 

Fulgurite

Depending on how things work out, I may stick with destruction, but AOE level. Basically the strategy is to pull with a combination of my pet and Seeds, then Shadowfury/Rain of Fire to finish them off when everything is grouped.

That was my general strategy for clearing Scourge Invasion mobs by myself, and I could easily take 6, possibly 7-8 at a time. Those mobs all have between 9,000 and 10,000 health, so it's fairly similar to Wrath mobs.

The thing I never liked about Affliction is even though you had the mobs dotted up to die, you couldn't move too much or they would leash back. I did level 1-70 as Aff though.

I've had a lot more fun soloing things as destro, though it's different now with Chaos Bolt because you can't sac a Fel Hunter. If I go for a 0/30/40 spec, maybe I'll sac one again, see how the downtime is with both Fel Armor and a sacced Fel Hunter ticking. Destro's advantage in killing the mobs in a few nukes is that they don't hit you for as long when you're aff, so you have less incoming damage to heal back up.

 

Fimotik

Whilst the maths in these posts has been correct, the in-game dynamics does add some extra flavour. For example, with any metagem or talent that boosts critical damage (e.g. Chaotic Skyfire Diamond), you're comparing a 100% hit v. 203% crit, so it tightens the margin slightly. The other thing someone eluded to but didn't expressly state in a previous post is that crit is better DPM (damage per mana), because you're getting twice as much damage for the same amount of mana spent. The downside of crit is the RNG factor (or an upside depending on the roll).

 

Sinthia

It seems to me that most everyone's guess on when to recast haunt is wrong. One of the problems that Simulationcraft doesnt take into account is that haunt has a travel time. What this means is that at max range of 30 yards, it will take 1-1.5 seconds for haunt to actually hit the target. Therefore you actually want to recast haunt when it has 3-2.5 seconds duration left, and not actually 1.5*haste before it falls off.

It also seems like the debuff counts the travel time in its duration, ie when it hits it only has 10 seconds left! I could be wrong here, if anyone can confirm this it would be appreciated. When haunt had a 10 second cd it seemed to fall off with no way for me to prevent it on Brutallus. I haven't tried affliction again since the patch, I will probably respec affliction and try again this Tuesday to see if 8 second cd have solved this uptime issue.

 

Razumikhin

Shadow Embrace doesn't appear in my timers unless I hit the mob with a Shadow Bolt or Haunt before another warlock with this talent. I believe that 3.03 was supposed to fix problems with Shadow Embrace, but has anyone tested this yet? Thanks.

 

Issa

Can only say sofar that I noticed the same thing, with another affli in the raid, my shadow embrace won't show on my bars. Sofar I haven't been able to find anything on this on the wow-europe-forums.

 

Zephro

I've been noticing strange behaviour with Haunt and Corruption. I respecced Affliction for the first time in a year last night and I was futzing around with the level 70 test dummy in Undercity. Haunt didn't seem to boost Corruption's damage by 20% as stated, it was considerably less. And when Haunt fell off, the Corruption ticks were for less than they were before it was applied.

Interestingly, the Haunted Corruption ticks were 20% more than the post-Haunt ticks.

Has anyone else had this problem? I got the impression it was some problem with the component of Haunt that resets Corruption perhaps ignoring one or several of the Corruption boosting talents, so your "new" Corruption spell ends up being less potent than the one originally cast.