[top]Haste and the Soft Cap

Torrential · 12/14/10, 6:00 PM

Originally Posted by Kaang

To the best of my knowledge buffs that increase stats by 5% (Kings & Mark) Spirit would be raised by the same portion. In that effect would (as one of these buffs would be available in all/most raid comps) it be wise to gem spirit in lieu of hit and aiming for 5% less the hit cap to allow for more Mastery and Haste be a wise decision? The 5% is a vague number for the sake of posing the question as having all hit come from Spirit is unlikely, but if say half of Hit is via Spirit one could free up 43.55 hit rating (2.5% of 1742 spell hit cap).

Also, regarding UE, outside of the direct damage being less than LB for the GcD, the 30% increase to the next fire spell's damage (assuming FT imbue) would make a DPS wash, if not a gain over LB if LVB is cast immediately after UE. The caveat being that casting UE on it's own when LvB is not or may not be available immediately aftarward would still be a loss, in which case LB takes priority.

Blessing of Kings - Spell - World of Warcraft

They've seen through your plan. Those buffs do not include Spirit.

As for UE do you have math to support that the low damage from UE + extra LvB damage comes out 'a wash'? I ask because pretty much everybody agrees that's not how it is.

Obeast · 12/14/10, 6:21 PM

No, arcane is not harder to model than any other specs. Any rotation in this game can always be modeled by an appropriate priority list with conditionals. Now maybe the sim doesn't have enough freedom to approach it, but this is not the case at the moment.

You're neglecting the fact that arcane does not have a single rotation. It actually is harder to model via simulation than other specs because of the nonlinearities introduced by mana adept. For other specs, it is trivial to list abilities in order of damage / casting time, and thus create a priority list to feed to simcraft. For arcane, your chioces in the present affect your damage in the future, so it is not at all trivial to generate an "optimal" priority list. You have to make assumptions about what is optimal; for opther specs, you just divide damage by casting time and solve the problem trivially.

But if you wish, knock yourself out -- here's what I'd like to see arcane's simcraft do:

Arcane has N optimal rotations (by which I mean rotations that cannot be improved upon by a linear combination of two other rotations), which it will use in ascending order of MPS. First, find the optimal rotations for a chosen set of gear.

Then, find the two optimal rotations to mix to create a rotation with 0 mana / sec. At present, my math (and other people's) says that a mix of this 0 MPS rotation with a "burn" period of AB spam + arcane power is at least close to optimal DPS. Mix the two optimal rotations to maintain mana near 100% in between burn periods.

This is not what simcraft does currently (I did actually read the priority list before posting). But even if you code that in, you still need to account for moments where RNG gives you different MPS values than the mean -- let's say you're trying to maintain MPS=0, but you hit a long stretch without AM procs and consume more mana than you expect. A "real" player compensates for this by switching to a rotation with MPS>>0 (although which one is not necessarily obvious) to make up the gap; until simcraft can do this (admittedly, probably a small DPS contribution) it won't give optimal results.

Glycell · 12/14/10, 7:03 PM

Originally Posted by Exemplar

[top]Haste and the Soft Cap

The ultimate nirvana of Crusader Strike is to reduce the GCD to 3 seconds. This permits you to use CS, a single filler, then another CS with no wasted time.

In a perfect world at L80 it takes around 1015 Haste (with Wrath of Air and JotP) to reduce CS to a 3 second CD. However, with latency a factor, on live the best you can do is cast CS, press an ability during GCD, new GCD begins as soon as first completed. This means the minimum time between two CS is your CS cooldown minus your latency. You would need to push the cooldown below 3 seconds so the cooldown would end in sufficient time for you to send a new CS command to the server and it be accepted precisely 3 seconds after the first occurred.

This value is your Soft Cap. Beyond this point haste is unable to accelerate your CS usage (but still increases Autoattacks).

Since you cannot queue during non-GCD CD periods, you are normally stuck with this single-times-latency issue in a CS+filler sequence. When the filler completes you must wait for CS to end its CD, then press the key and wait for latency. If you push CS CD below 3 seconds then you will be able to queue during the second ability's GCD.

Note: Spells have a special rule with the GCD. The spell GCD is reduced by haste (to a minimum of 1 second). If you manage to acquire 1015 Haste your spell GCD becomes 1 second. Thus even with zero latency you would have: CS, Exo, 0.5 seconds empty time, CS. Remember that Exo, HW, Cons, and DP are spells. Judge, TV, DS, and HoW are not and use the full 1.5 GCD every time.

Second Note: Latency displayed in-game is an average. You could be lower than this value the majority of the time with intermittent spikes producing a higher average. Thus your CS would not be soft capped the majority of the time, then suddenly beyond your soft cap when that odd lag spike occurs.

It is unlikely that even in end-tier Cataclysm gear that it will be possible to fully cap your CS CD. Unless several mechanics severely change, do not attempt to meet the cap via Gemming. Strength is more than twice as advantageous as Haste. Any gems which are Haste rather than Str are overall detrimental to your total DPS.

I can't test myself, but we should work on providing the Info to help the OP get this up to date. At what haste rating at level 85 will produce the same results of 1015 at level 80. From my observations from spread sheets and such it's around 3500, I could be really wrong. Not having cata and being locked at 80 is hindering me from exploring further.

Edit: Ok I found from just spread sheets that it's 3950 Haste Rating from Gear/Enchants to reach the pure 3.0s CD CS at 85. Once again I doubt this number it's seems really high to me.

Renshaw · 12/14/10, 7:39 PM

I know this is probably beating a dead horse by now, but with all the varying ideas, I haven't been able to find a clear answer as to what is the better stat to gem/stack. I'm just getting myself geared for heroics (currently in my pre-heroic stage), and was thinking of stacking stamina. Is it a better idea to go with that, or should I focus more on the likes of agi/exp/mast/dodge?

Pussifire · 12/14/10, 9:53 PM

Sorry guys this is my first time on this forum. Just to inquiry the arcane mage rotation what is the updated one now? Currently I'm having a hard time coping with the rotation to deliver sustained damage. Is anyone having the same problem as me?

slandry · 12/14/10, 9:58 PM

nor sure this is the right place to ask this. i just started replaying my shammy after a few months off playing my main getting leveled up and geared. now i wanted to get my shammy to lvl 85 a resto but i just dont know how or what rotation to use as i will be doing lots of instances and raid. oh i'm kind of new to the whole healing thing i leveled up as a elemental shammy

Reik · 12/14/10, 10:00 PM

I know they say Dual Wield is not as good for threat, but would the increased creation of Blood Worms be enough free healing to warrant it?

TheExodu5 · 12/14/10, 10:13 PM

I can't believe how bad the melee dps enchant is from engineering. That and the increased chance of rocket boots going off makes me a sad panda.

Any word if we're better off just using Hyperspeed Accelerators?

hangloose9316 · 12/15/10, 12:16 AM

1. you are missing the concrete number for the hit cap at lvl 85 (i think) ;-)
2. i'm a little confused by the fact that blizz didnt make my spirit convert to hit on the stat pane ---> i just have to assume from the amount of spirit i have that the percent chance i have to miss a boss mob (blizzard provided on character pane) is incorrect? or am i missing something?

thanks

RavenJet · 12/15/10, 3:06 AM

Okay, reading through the affliction threads already posted the advice is to use either Succubus (with Lash of Pain glyph) or Imp (with glyph of Imp) as your pet for affliction. This despite the advantages the old fell puppy would otherwise have - because that is the one pet that doesn't have a damage boosting glyph.

But it seems that all of these were written from a not-yet-level-85 position ? Level 85 brings demon soul (Demon Soul - Spell - World of Warcraft) as a spell for all specs, and looking over the spell - it's clear that the best demon soul effect for affliction is from the fell puppy (which is to be expected since it's meant to be our signature demon after all). As it gives a major boost to shadow DoTs.

The catch is that it's a cooldown and it only lasts 20s. I am not at all sure that the effect won't be more than outdone by the constant effects of glyphs. More-over the Succubus's effect is actually not bad for affliction locks either - as it's a boost to shadow bolt.
Personally I went for the drain-life-filler rotation over the shadow-bolt-filler - I found damage to be comparable but in cataclysm survivability is at least as important as DPS and healer mana is much more limited so those major self-heals are a huge boon - I can actually tell my healers "never heal me unless I am low on health *and* out of mana".

So that makes the succubus's shadow bolt less than useful to me (and I suspect in time - most afflocks, everything I'm seeing suggests that draintanking is due for a major comeback in cataclysm). The imp effect however is entirely useless to affliction as it's literally a boost to destro spells only (and that means Bane of Agony and Fel Flame is the only spells we'll see an effect on and fel flame itself is mostly only useful to get an instant-cast renew on UA). In short - the imp has no real value with demon soul.

So the real question is - how much difference can we reasonably expect demon soul to make in a fight ? Enough to alter the existing considerations that lead to the suggestion of imp for drain-life and succubus for shadow bolt ? Is the change enough to let the puppy make a comeback ?
There is one other major advantage to the puppy - it gives warlocks an interrupting silence which has in the past proven incredibly useful (if situational) as a way to disrupt caster mobs. Not a primary consideration but a worthwhile one at least on fights where interrupts are known to be particularly valuable.

Do we have any data on the real effect of demon soul then ? Should I rethink my pet choices with it in mind ? Should I stick to the imp for drain life and just think of DS as a boost to BoA ?

randa · 12/15/10, 3:31 AM

I don't get why simulationcraft gives so much emphasis on haste rating for my character when I am above 579 threshold. Here are my results.

Zalfirion · 12/15/10, 4:07 AM

Kintoun wrote:
Also tried to favor items that can be purchased with Reputation or Justice Points.

I guess thats why.

csarichmond · 12/15/10, 5:07 AM

Ok, thanks, I hadn't factored that in.

Jahamakarahl · 12/15/10, 6:19 AM

Originally Posted by Torrential

This is not how things work. When you cast a Lava Burst the damage it's going to do is determined the instant your cast is finished. Any flame shock application, UE use, trinket procs etc. while the spell is in flight to the target have no effect.

When you say this I feel dumb that I forgot that completly. With that in mind then yes UE is worth using while having a slight movement like Masanbol mentioned.

One thing I've been wondering and want to know if you have done any research on it.
Which of the potions are better. The 1200 SP for 25 seconds (Volcanic Potion) or the 600 Int for 20 seconds (Potion of Pure Genius).
Since 1 Int = 1 SP I would assume that 1200 SP is better, but then for what class would 600 Int be better. Any thoughts?

diospadre · 12/15/10, 6:52 AM

Originally Posted by hangloose9316

1. you are missing the concrete number for the hit cap at lvl 85 (i think) ;-)
2. i'm a little confused by the fact that blizz didnt make my spirit convert to hit on the stat pane ---> i just have to assume from the amount of spirit i have that the percent chance i have to miss a boss mob (blizzard provided on character pane) is incorrect? or am i missing something?

thanks

It does, you're looking at the melee hit number.